Add files using upload-large-folder tool

checkpoint-120516/config.json

@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "/workspace/GOT_weights",
+  "architectures": [
+    "GOTQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "freeze_vision_tower": false,
+  "hidden_act": "silu",
+  "hidden_size": 1024,
+  "im_end_token": 151858,
+  "im_patch_token": 151859,
+  "im_start_token": 151857,
+  "image_token_len": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 2816,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "GOT",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 16,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.37.2",
+  "use_cache": true,
+  "use_im_start_end": true,
+  "use_sliding_window": false,
+  "vision_select_layer": -1,
+  "vision_tower": "~/.cache/huggingface/hub/models--openai--clip-vit-large-patch14/snapshots/8d052a0f05efbaefbc9e8786ba291cfdf93e5bff/",
+  "vocab_size": 151860
+}

checkpoint-120516/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.37.2"
+}

checkpoint-120516/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27d5193f8f2922e374a369bcd49677ec9db2ad58e7b8ada37392bd4a6ebe2502
+size 1121111672

checkpoint-120516/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f9ec76214be21c462a1ace41e4425b3b581d670c4beb0782cfeecf4505b134b
+size 1860083898

checkpoint-120516/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee3528bf0ace792176d57cac1ea8e325db1e81a8856e3e8a6e53688b51f9516e
+size 14244

checkpoint-120516/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:752cd3601d7c8d7a568cca5f38f7a484e677f4f877f33284a16880a71db6148a
+size 1064

checkpoint-120516/special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-120516/tokenization_qwen.py

@@ -0,0 +1,264 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Tokenization classes for QWen."""
+
+import base64
+import logging
+import os
+import unicodedata
+from typing import Collection, Dict, List, Set, Tuple, Union
+
+import tiktoken
+from transformers import PreTrainedTokenizer, AddedToken
+
+logger = logging.getLogger(__name__)
+
+
+VOCAB_FILES_NAMES = {"vocab_file": "qwen.tiktoken"}
+
+PAT_STR = r"""(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+"""
+ENDOFTEXT = "<|endoftext|>"
+IMSTART = "<|im_start|>"
+IMEND = "<|im_end|>"
+# as the default behavior is changed to allow special tokens in
+# regular texts, the surface forms of special tokens need to be
+# as different as possible to minimize the impact
+EXTRAS = tuple((f"<|extra_{i}|>" for i in range(205)))
+SPECIAL_TOKENS = (
+    ENDOFTEXT,
+    IMSTART,
+    IMEND,
+) + EXTRAS
+
+
+def _load_tiktoken_bpe(tiktoken_bpe_file: str) -> Dict[bytes, int]:
+    with open(tiktoken_bpe_file, "rb") as f:
+        contents = f.read()
+    return {
+        base64.b64decode(token): int(rank)
+        for token, rank in (line.split() for line in contents.splitlines() if line)
+    }
+
+class QWenTokenizer(PreTrainedTokenizer):
+    """QWen tokenizer."""
+
+    vocab_files_names = VOCAB_FILES_NAMES
+
+    def __init__(
+        self,
+        vocab_file,
+        errors="replace",
+        image_start_tag='<img>',
+        image_end_tag='</img>',
+        image_pad_tag='<imgpad>',
+        ref_start_tag='<ref>',
+        ref_end_tag='</ref>',
+        box_start_tag='<box>',
+        box_end_tag='</box>',
+        quad_start_tag='<quad>',
+        quad_end_tag='</quad>',
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        
+        self.image_start_tag = image_start_tag
+        self.image_end_tag = image_end_tag
+        self.image_pad_tag = image_pad_tag
+        self.ref_start_tag = ref_start_tag
+        self.ref_end_tag = ref_end_tag
+        self.box_start_tag = box_start_tag
+        self.box_end_tag = box_end_tag
+        self.quad_start_tag = quad_start_tag
+        self.quad_end_tag = quad_end_tag
+        self.IMAGE_ST = (
+            ref_start_tag, ref_end_tag,
+            box_start_tag, box_end_tag,
+            quad_start_tag, quad_end_tag,
+            image_start_tag, image_end_tag,
+            image_pad_tag
+        )
+
+        self.errors = errors  # how to handle errors in decoding
+
+        self.mergeable_ranks = _load_tiktoken_bpe(vocab_file)  # type: dict[bytes, int]
+        self.special_tokens = {
+            token: index
+            for index, token in enumerate(
+                SPECIAL_TOKENS + self.IMAGE_ST, start=len(self.mergeable_ranks)
+            )
+        }
+        
+        self.img_start_id = self.special_tokens[self.image_start_tag]
+        self.img_end_id = self.special_tokens[self.image_end_tag]
+        self.img_pad_id = self.special_tokens[self.image_pad_tag]
+        self.ref_start_id = self.special_tokens[self.ref_start_tag]
+        self.ref_end_id = self.special_tokens[self.ref_end_tag]
+        self.box_start_id = self.special_tokens[self.box_start_tag]
+        self.box_end_id = self.special_tokens[self.box_end_tag]
+        self.quad_start_id = self.special_tokens[self.quad_start_tag]
+        self.quad_end_id = self.special_tokens[self.quad_end_tag]
+
+        enc = tiktoken.Encoding(
+            "Qwen",
+            pat_str=PAT_STR,
+            mergeable_ranks=self.mergeable_ranks,
+            special_tokens=self.special_tokens,
+        )
+        assert (
+            len(self.mergeable_ranks) + len(self.special_tokens) == enc.n_vocab
+        ), f"{len(self.mergeable_ranks) + len(self.special_tokens)} != {enc.n_vocab} in encoding"
+
+        self.decoder = {
+            v: k for k, v in self.mergeable_ranks.items()
+        }  # type: dict[int, bytes|str]
+        self.decoder.update({v: k for k, v in self.special_tokens.items()})
+
+        self.tokenizer = enc  # type: tiktoken.Encoding
+
+        self.eod_id = self.tokenizer.eot_token
+        self.im_start_id = self.special_tokens[IMSTART]
+        self.im_end_id = self.special_tokens[IMEND]
+
+    def __len__(self) -> int:
+        return self.tokenizer.n_vocab
+
+    def get_vocab(self) -> Dict[bytes, int]:
+        return self.mergeable_ranks
+
+    def convert_tokens_to_ids(
+        self, tokens: Union[bytes, str, List[Union[bytes, str]]]
+    ) -> List[int]:
+        ids = []
+        if isinstance(tokens, (str, bytes)):
+            if tokens in self.special_tokens:
+                return self.special_tokens[tokens]
+            else:
+                return self.mergeable_ranks.get(tokens)
+        for token in tokens:
+            if token in self.special_tokens:
+                ids.append(self.special_tokens[token])
+            else:
+                ids.append(self.mergeable_ranks.get(token))
+        return ids
+
+    def _add_tokens(self, new_tokens: Union[List[str], List[AddedToken]], special_tokens: bool = False) -> int:
+        if not special_tokens and new_tokens:
+            raise ValueError('Adding regular tokens is not supported')
+        for token in new_tokens:
+            surface_form = token.content if isinstance(token, AddedToken) else token
+            if surface_form not in SPECIAL_TOKENS:
+                raise ValueError('Adding unknown special tokens is not supported')
+        return 0
+
+    def save_vocabulary(self, save_directory: str, **kwargs) -> Tuple[str]:
+        """
+        Save only the vocabulary of the tokenizer (vocabulary).
+
+        Returns:
+            `Tuple(str)`: Paths to the files saved.
+        """
+        file_path = os.path.join(save_directory, "qwen.tiktoken")
+        with open(file_path, "w", encoding="utf8") as w:
+            for k, v in self.mergeable_ranks.items():
+                line = base64.b64encode(k).decode("utf8") + " " + str(v) + "\n"
+                w.write(line)
+        return (file_path,)
+
+    def tokenize(
+        self,
+        text: str,
+        allowed_special: Union[Set, str] = "all",
+        disallowed_special: Union[Collection, str] = (),
+        **kwargs,
+    ) -> List[Union[bytes, str]]:
+        """
+        Converts a string in a sequence of tokens.
+
+        Args:
+            text (`str`):
+                The sequence to be encoded.
+            allowed_special (`Literal["all"]` or `set`):
+                The surface forms of the tokens to be encoded as special tokens in regular texts.
+                Default to "all".
+            disallowed_special (`Literal["all"]` or `Collection`):
+                The surface forms of the tokens that should not be in regular texts and trigger errors.
+                Default to an empty tuple.
+
+            kwargs (additional keyword arguments, *optional*):
+                Will be passed to the underlying model specific encode method.
+
+        Returns:
+            `List[bytes|str]`: The list of tokens.
+        """
+        tokens = []
+        text = unicodedata.normalize("NFC", text)
+
+        # this implementation takes a detour: text -> token id -> token surface forms
+        for t in self.tokenizer.encode(
+            text, allowed_special=allowed_special, disallowed_special=disallowed_special
+        ):
+            tokens.append(self.decoder[t])
+        return tokens
+
+    def convert_tokens_to_string(self, tokens: List[Union[bytes, str]]) -> str:
+        """
+        Converts a sequence of tokens in a single string.
+        """
+        text = ""
+        temp = b""
+        for t in tokens:
+            if isinstance(t, str):
+                if temp:
+                    text += temp.decode("utf-8", errors=self.errors)
+                    temp = b""
+                text += t
+            elif isinstance(t, bytes):
+                temp += t
+            else:
+                raise TypeError("token should only be of type types or str")
+        if temp:
+            text += temp.decode("utf-8", errors=self.errors)
+        return text
+
+    @property
+    def vocab_size(self):
+        return self.tokenizer.n_vocab
+
+    def _convert_id_to_token(self, index: int) -> Union[bytes, str]:
+        """Converts an id to a token, special tokens included"""
+        if index in self.decoder:
+            return self.decoder[index]
+        raise ValueError("unknown ids")
+
+    def _convert_token_to_id(self, token: Union[bytes, str]) -> int:
+        """Converts a token to an id using the vocab, special tokens included"""
+        if token in self.special_tokens:
+            return self.special_tokens[token]
+        if token in self.mergeable_ranks:
+            return self.mergeable_ranks[token]
+        raise ValueError("unknown token")
+
+    def _tokenize(self, text: str, **kwargs):
+        """
+        Converts a string in a sequence of tokens (string), using the tokenizer. Split in words for word-based
+        vocabulary or sub-words for sub-word-based vocabularies (BPE/SentencePieces/WordPieces).
+
+        Do NOT take care of added tokens.
+        """
+        raise NotImplementedError
+
+    def _decode(
+        self,
+        token_ids: Union[int, List[int]],
+        skip_special_tokens: bool = False,
+        errors: str = None,
+        **kwargs,
+    ) -> str:
+        if isinstance(token_ids, int):
+            token_ids = [token_ids]
+        if skip_special_tokens:
+            token_ids = [i for i in token_ids if i < self.eod_id]
+        return self.tokenizer.decode(token_ids, errors=errors or self.errors)

checkpoint-120516/tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "added_tokens_decoder": {},
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_qwen.QWenTokenizer",
+      null
+    ]
+  },
+  "clean_up_tokenization_spaces": true,
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "tokenizer_class": "QWenTokenizer"
+}

checkpoint-120516/trainer_state.json

@@ -0,0 +1,1467 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 120516,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9944682310509254e-05,
+      "loss": 1.5408,
+      "step": 500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.988936462101851e-05,
+      "loss": 1.2376,
+      "step": 1000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9834046931527766e-05,
+      "loss": 1.1307,
+      "step": 1500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.977872924203702e-05,
+      "loss": 1.103,
+      "step": 2000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9723411552546275e-05,
+      "loss": 1.0708,
+      "step": 2500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9668093863055527e-05,
+      "loss": 1.0288,
+      "step": 3000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9612776173564783e-05,
+      "loss": 1.0089,
+      "step": 3500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.955745848407404e-05,
+      "loss": 1.0096,
+      "step": 4000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9502140794583292e-05,
+      "loss": 1.0213,
+      "step": 4500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9446823105092548e-05,
+      "loss": 0.9962,
+      "step": 5000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9391505415601804e-05,
+      "loss": 0.9709,
+      "step": 5500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9336187726111056e-05,
+      "loss": 0.9898,
+      "step": 6000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9280870036620312e-05,
+      "loss": 0.9536,
+      "step": 6500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9225552347129565e-05,
+      "loss": 0.9591,
+      "step": 7000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.917023465763882e-05,
+      "loss": 0.9851,
+      "step": 7500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9114916968148077e-05,
+      "loss": 0.9548,
+      "step": 8000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.905959927865733e-05,
+      "loss": 1.0179,
+      "step": 8500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9004281589166585e-05,
+      "loss": 0.9737,
+      "step": 9000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.894896389967584e-05,
+      "loss": 0.9592,
+      "step": 9500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8893646210185094e-05,
+      "loss": 0.9488,
+      "step": 10000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.883832852069435e-05,
+      "loss": 0.9598,
+      "step": 10500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8783010831203603e-05,
+      "loss": 0.9572,
+      "step": 11000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.872769314171286e-05,
+      "loss": 0.9789,
+      "step": 11500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8672375452222115e-05,
+      "loss": 0.9503,
+      "step": 12000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8617057762731367e-05,
+      "loss": 0.9296,
+      "step": 12500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8561740073240623e-05,
+      "loss": 0.933,
+      "step": 13000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.850642238374988e-05,
+      "loss": 0.9438,
+      "step": 13500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.845110469425913e-05,
+      "loss": 0.919,
+      "step": 14000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.8395787004768388e-05,
+      "loss": 0.9704,
+      "step": 14500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.834046931527764e-05,
+      "loss": 0.9099,
+      "step": 15000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.8285151625786896e-05,
+      "loss": 0.9517,
+      "step": 15500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8229833936296152e-05,
+      "loss": 0.9463,
+      "step": 16000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8174516246805405e-05,
+      "loss": 0.9535,
+      "step": 16500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.8119198557314657e-05,
+      "loss": 0.9372,
+      "step": 17000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.8063880867823917e-05,
+      "loss": 0.9515,
+      "step": 17500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.800856317833317e-05,
+      "loss": 0.9675,
+      "step": 18000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.7953245488842422e-05,
+      "loss": 0.9579,
+      "step": 18500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.7897927799351678e-05,
+      "loss": 0.961,
+      "step": 19000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.7842610109860934e-05,
+      "loss": 0.9397,
+      "step": 19500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.7787292420370186e-05,
+      "loss": 0.9518,
+      "step": 20000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.7731974730879442e-05,
+      "loss": 0.9745,
+      "step": 20500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.7676657041388695e-05,
+      "loss": 0.8979,
+      "step": 21000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.762133935189795e-05,
+      "loss": 0.9224,
+      "step": 21500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.7566021662407207e-05,
+      "loss": 0.9556,
+      "step": 22000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.751070397291646e-05,
+      "loss": 0.978,
+      "step": 22500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.7455386283425715e-05,
+      "loss": 0.9102,
+      "step": 23000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.7400068593934968e-05,
+      "loss": 0.9124,
+      "step": 23500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.7344750904444224e-05,
+      "loss": 0.9089,
+      "step": 24000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.728943321495348e-05,
+      "loss": 0.9296,
+      "step": 24500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.7234115525462733e-05,
+      "loss": 0.9406,
+      "step": 25000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.717879783597199e-05,
+      "loss": 0.9625,
+      "step": 25500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7123480146481245e-05,
+      "loss": 0.9196,
+      "step": 26000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7068162456990497e-05,
+      "loss": 0.9402,
+      "step": 26500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7012844767499753e-05,
+      "loss": 0.977,
+      "step": 27000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6957527078009006e-05,
+      "loss": 0.9082,
+      "step": 27500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.690220938851826e-05,
+      "loss": 0.937,
+      "step": 28000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.6846891699027518e-05,
+      "loss": 0.8986,
+      "step": 28500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.679157400953677e-05,
+      "loss": 0.9493,
+      "step": 29000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6736256320046026e-05,
+      "loss": 0.9245,
+      "step": 29500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.6680938630555282e-05,
+      "loss": 0.8917,
+      "step": 30000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6625620941064535e-05,
+      "loss": 0.9436,
+      "step": 30500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.657030325157379e-05,
+      "loss": 0.9,
+      "step": 31000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6514985562083043e-05,
+      "loss": 0.9481,
+      "step": 31500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.64596678725923e-05,
+      "loss": 0.9146,
+      "step": 32000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6404350183101555e-05,
+      "loss": 0.9303,
+      "step": 32500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6349032493610808e-05,
+      "loss": 0.9305,
+      "step": 33000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.629371480412006e-05,
+      "loss": 0.9222,
+      "step": 33500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.623839711462932e-05,
+      "loss": 0.9495,
+      "step": 34000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6183079425138572e-05,
+      "loss": 0.8726,
+      "step": 34500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6127761735647825e-05,
+      "loss": 0.9479,
+      "step": 35000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.607244404615708e-05,
+      "loss": 0.9395,
+      "step": 35500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6017126356666337e-05,
+      "loss": 0.9261,
+      "step": 36000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.596180866717559e-05,
+      "loss": 0.8737,
+      "step": 36500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5906490977684845e-05,
+      "loss": 0.91,
+      "step": 37000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5851173288194098e-05,
+      "loss": 0.9309,
+      "step": 37500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5795855598703354e-05,
+      "loss": 0.9278,
+      "step": 38000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.574053790921261e-05,
+      "loss": 0.9506,
+      "step": 38500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5685220219721863e-05,
+      "loss": 0.8787,
+      "step": 39000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.562990253023112e-05,
+      "loss": 0.9097,
+      "step": 39500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.557458484074037e-05,
+      "loss": 0.8812,
+      "step": 40000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5519267151249627e-05,
+      "loss": 0.8798,
+      "step": 40500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5463949461758883e-05,
+      "loss": 0.9398,
+      "step": 41000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5408631772268136e-05,
+      "loss": 0.9148,
+      "step": 41500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.535331408277739e-05,
+      "loss": 0.9398,
+      "step": 42000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5297996393286648e-05,
+      "loss": 0.8984,
+      "step": 42500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5242678703795902e-05,
+      "loss": 0.8947,
+      "step": 43000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.5187361014305155e-05,
+      "loss": 0.8767,
+      "step": 43500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.5132043324814409e-05,
+      "loss": 0.906,
+      "step": 44000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.5076725635323666e-05,
+      "loss": 0.9188,
+      "step": 44500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.5021407945832919e-05,
+      "loss": 0.9534,
+      "step": 45000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4966090256342173e-05,
+      "loss": 0.913,
+      "step": 45500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4910772566851428e-05,
+      "loss": 0.8824,
+      "step": 46000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4855454877360684e-05,
+      "loss": 0.9155,
+      "step": 46500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.4800137187869938e-05,
+      "loss": 0.9142,
+      "step": 47000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.4744819498379192e-05,
+      "loss": 0.9195,
+      "step": 47500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.4689501808888446e-05,
+      "loss": 0.9207,
+      "step": 48000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.4634184119397702e-05,
+      "loss": 0.9057,
+      "step": 48500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.4578866429906957e-05,
+      "loss": 0.9273,
+      "step": 49000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.4523548740416211e-05,
+      "loss": 0.9333,
+      "step": 49500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4468231050925465e-05,
+      "loss": 0.8996,
+      "step": 50000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.4412913361434721e-05,
+      "loss": 0.9298,
+      "step": 50500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.4357595671943975e-05,
+      "loss": 0.933,
+      "step": 51000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.430227798245323e-05,
+      "loss": 0.8895,
+      "step": 51500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4246960292962484e-05,
+      "loss": 0.9297,
+      "step": 52000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.419164260347174e-05,
+      "loss": 0.9256,
+      "step": 52500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.4136324913980994e-05,
+      "loss": 0.9088,
+      "step": 53000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.4081007224490249e-05,
+      "loss": 0.9162,
+      "step": 53500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.4025689534999503e-05,
+      "loss": 0.9249,
+      "step": 54000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.3970371845508759e-05,
+      "loss": 0.9561,
+      "step": 54500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3915054156018013e-05,
+      "loss": 0.9376,
+      "step": 55000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.3859736466527267e-05,
+      "loss": 0.9404,
+      "step": 55500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.3804418777036522e-05,
+      "loss": 0.9654,
+      "step": 56000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.3749101087545778e-05,
+      "loss": 0.8642,
+      "step": 56500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3693783398055032e-05,
+      "loss": 0.9651,
+      "step": 57000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3638465708564286e-05,
+      "loss": 0.928,
+      "step": 57500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.358314801907354e-05,
+      "loss": 0.9215,
+      "step": 58000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.3527830329582796e-05,
+      "loss": 0.8917,
+      "step": 58500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.347251264009205e-05,
+      "loss": 0.8971,
+      "step": 59000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3417194950601305e-05,
+      "loss": 0.9151,
+      "step": 59500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.336187726111056e-05,
+      "loss": 0.8999,
+      "step": 60000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.3306559571619812e-05,
+      "loss": 0.9411,
+      "step": 60500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.325124188212907e-05,
+      "loss": 0.9059,
+      "step": 61000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.3195924192638324e-05,
+      "loss": 0.8741,
+      "step": 61500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.3140606503147576e-05,
+      "loss": 0.93,
+      "step": 62000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.308528881365683e-05,
+      "loss": 0.9157,
+      "step": 62500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.3029971124166088e-05,
+      "loss": 0.8834,
+      "step": 63000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.2974653434675341e-05,
+      "loss": 0.9047,
+      "step": 63500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.2919335745184595e-05,
+      "loss": 0.9396,
+      "step": 64000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.286401805569385e-05,
+      "loss": 0.9448,
+      "step": 64500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.2808700366203106e-05,
+      "loss": 0.9332,
+      "step": 65000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.275338267671236e-05,
+      "loss": 0.9394,
+      "step": 65500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.2698064987221614e-05,
+      "loss": 0.9217,
+      "step": 66000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.2642747297730868e-05,
+      "loss": 0.893,
+      "step": 66500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.2587429608240124e-05,
+      "loss": 0.8987,
+      "step": 67000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.2532111918749379e-05,
+      "loss": 0.8831,
+      "step": 67500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.2476794229258633e-05,
+      "loss": 0.9192,
+      "step": 68000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.2421476539767887e-05,
+      "loss": 0.9175,
+      "step": 68500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.2366158850277143e-05,
+      "loss": 0.9207,
+      "step": 69000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.2310841160786397e-05,
+      "loss": 0.8792,
+      "step": 69500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.2255523471295652e-05,
+      "loss": 0.9052,
+      "step": 70000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.2200205781804906e-05,
+      "loss": 0.9362,
+      "step": 70500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.2144888092314162e-05,
+      "loss": 0.8689,
+      "step": 71000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.2089570402823416e-05,
+      "loss": 0.8964,
+      "step": 71500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.203425271333267e-05,
+      "loss": 0.9057,
+      "step": 72000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.1978935023841925e-05,
+      "loss": 0.9527,
+      "step": 72500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.192361733435118e-05,
+      "loss": 0.9061,
+      "step": 73000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.1868299644860435e-05,
+      "loss": 0.9662,
+      "step": 73500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.181298195536969e-05,
+      "loss": 0.9286,
+      "step": 74000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.1757664265878944e-05,
+      "loss": 0.953,
+      "step": 74500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.17023465763882e-05,
+      "loss": 0.9109,
+      "step": 75000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.1647028886897454e-05,
+      "loss": 0.9167,
+      "step": 75500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.1591711197406708e-05,
+      "loss": 0.9382,
+      "step": 76000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.1536393507915962e-05,
+      "loss": 0.9144,
+      "step": 76500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.1481075818425218e-05,
+      "loss": 0.8935,
+      "step": 77000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.1425758128934473e-05,
+      "loss": 0.9181,
+      "step": 77500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.1370440439443727e-05,
+      "loss": 0.9167,
+      "step": 78000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.131512274995298e-05,
+      "loss": 0.8912,
+      "step": 78500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.1259805060462234e-05,
+      "loss": 0.9078,
+      "step": 79000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.1204487370971491e-05,
+      "loss": 0.9164,
+      "step": 79500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.1149169681480744e-05,
+      "loss": 0.9178,
+      "step": 80000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.1093851991989998e-05,
+      "loss": 0.8802,
+      "step": 80500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.1038534302499253e-05,
+      "loss": 0.8861,
+      "step": 81000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.0983216613008509e-05,
+      "loss": 0.9514,
+      "step": 81500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.0927898923517763e-05,
+      "loss": 0.9349,
+      "step": 82000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.0872581234027017e-05,
+      "loss": 0.8864,
+      "step": 82500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.0817263544536271e-05,
+      "loss": 0.9031,
+      "step": 83000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.0761945855045527e-05,
+      "loss": 0.9202,
+      "step": 83500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.0706628165554782e-05,
+      "loss": 0.964,
+      "step": 84000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.0651310476064036e-05,
+      "loss": 0.8934,
+      "step": 84500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.059599278657329e-05,
+      "loss": 0.8834,
+      "step": 85000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.0540675097082546e-05,
+      "loss": 0.9234,
+      "step": 85500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.04853574075918e-05,
+      "loss": 0.9243,
+      "step": 86000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.0430039718101055e-05,
+      "loss": 0.9032,
+      "step": 86500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.0374722028610309e-05,
+      "loss": 0.8856,
+      "step": 87000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.0319404339119565e-05,
+      "loss": 0.918,
+      "step": 87500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.026408664962882e-05,
+      "loss": 0.9737,
+      "step": 88000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.0208768960138074e-05,
+      "loss": 0.8813,
+      "step": 88500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.0153451270647328e-05,
+      "loss": 0.8957,
+      "step": 89000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.0098133581156584e-05,
+      "loss": 0.9095,
+      "step": 89500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.0042815891665838e-05,
+      "loss": 0.9336,
+      "step": 90000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 9.987498202175092e-06,
+      "loss": 0.8921,
+      "step": 90500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 9.932180512684347e-06,
+      "loss": 0.8593,
+      "step": 91000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 9.876862823193601e-06,
+      "loss": 0.9258,
+      "step": 91500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 9.821545133702857e-06,
+      "loss": 0.927,
+      "step": 92000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 9.766227444212111e-06,
+      "loss": 0.9453,
+      "step": 92500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 9.710909754721366e-06,
+      "loss": 0.9271,
+      "step": 93000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 9.65559206523062e-06,
+      "loss": 0.9393,
+      "step": 93500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 9.600274375739876e-06,
+      "loss": 0.9376,
+      "step": 94000
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 9.54495668624913e-06,
+      "loss": 0.929,
+      "step": 94500
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.489638996758384e-06,
+      "loss": 0.9438,
+      "step": 95000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.434321307267639e-06,
+      "loss": 0.9369,
+      "step": 95500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 9.379003617776895e-06,
+      "loss": 0.9364,
+      "step": 96000
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 9.323685928286149e-06,
+      "loss": 0.9375,
+      "step": 96500
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 9.268368238795403e-06,
+      "loss": 0.8816,
+      "step": 97000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 9.213050549304657e-06,
+      "loss": 0.916,
+      "step": 97500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.157732859813913e-06,
+      "loss": 0.994,
+      "step": 98000
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.102415170323166e-06,
+      "loss": 0.9202,
+      "step": 98500
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 9.04709748083242e-06,
+      "loss": 0.9044,
+      "step": 99000
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 8.991779791341676e-06,
+      "loss": 0.8644,
+      "step": 99500
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 8.93646210185093e-06,
+      "loss": 0.8926,
+      "step": 100000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.881144412360185e-06,
+      "loss": 0.9647,
+      "step": 100500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.825826722869439e-06,
+      "loss": 0.9277,
+      "step": 101000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.770509033378695e-06,
+      "loss": 0.8776,
+      "step": 101500
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.71519134388795e-06,
+      "loss": 0.8863,
+      "step": 102000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.659873654397204e-06,
+      "loss": 0.9286,
+      "step": 102500
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.604555964906458e-06,
+      "loss": 0.9329,
+      "step": 103000
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.549238275415714e-06,
+      "loss": 0.9091,
+      "step": 103500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.493920585924968e-06,
+      "loss": 0.9215,
+      "step": 104000
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.438602896434222e-06,
+      "loss": 0.9013,
+      "step": 104500
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.383285206943477e-06,
+      "loss": 0.9064,
+      "step": 105000
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.327967517452733e-06,
+      "loss": 0.8798,
+      "step": 105500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 8.272649827961985e-06,
+      "loss": 0.9142,
+      "step": 106000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 8.217332138471241e-06,
+      "loss": 0.9415,
+      "step": 106500
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 8.162014448980496e-06,
+      "loss": 0.8729,
+      "step": 107000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 8.10669675948975e-06,
+      "loss": 0.9351,
+      "step": 107500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 8.051379069999004e-06,
+      "loss": 0.9186,
+      "step": 108000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.99606138050826e-06,
+      "loss": 0.9713,
+      "step": 108500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.940743691017514e-06,
+      "loss": 0.9179,
+      "step": 109000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.885426001526769e-06,
+      "loss": 0.9631,
+      "step": 109500
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.830108312036023e-06,
+      "loss": 0.891,
+      "step": 110000
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.774790622545279e-06,
+      "loss": 0.9189,
+      "step": 110500
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 7.719472933054533e-06,
+      "loss": 0.8631,
+      "step": 111000
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 7.664155243563787e-06,
+      "loss": 0.9599,
+      "step": 111500
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 7.608837554073042e-06,
+      "loss": 0.9375,
+      "step": 112000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 7.553519864582297e-06,
+      "loss": 0.921,
+      "step": 112500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 7.498202175091551e-06,
+      "loss": 0.8958,
+      "step": 113000
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 7.442884485600806e-06,
+      "loss": 0.9505,
+      "step": 113500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 7.3875667961100605e-06,
+      "loss": 0.9405,
+      "step": 114000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 7.332249106619316e-06,
+      "loss": 0.8996,
+      "step": 114500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 7.27693141712857e-06,
+      "loss": 0.8697,
+      "step": 115000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.221613727637825e-06,
+      "loss": 0.9324,
+      "step": 115500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 7.166296038147079e-06,
+      "loss": 0.8865,
+      "step": 116000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 7.1109783486563345e-06,
+      "loss": 0.9143,
+      "step": 116500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 7.055660659165589e-06,
+      "loss": 0.8974,
+      "step": 117000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 7.000342969674844e-06,
+      "loss": 0.926,
+      "step": 117500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 6.945025280184098e-06,
+      "loss": 0.9321,
+      "step": 118000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 6.889707590693352e-06,
+      "loss": 0.9175,
+      "step": 118500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 6.834389901202607e-06,
+      "loss": 0.9066,
+      "step": 119000
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 6.779072211711861e-06,
+      "loss": 0.9149,
+      "step": 119500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 6.723754522221116e-06,
+      "loss": 0.9249,
+      "step": 120000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 6.66843683273037e-06,
+      "loss": 0.8616,
+      "step": 120500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 180774,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "total_flos": 1.9790822990016e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-120516/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5ddd77616b4f164cac0cdef0d02b6f4236309ff39a32650cbcc8037f91ac321
+size 4920

checkpoint-180774/config.json

@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "/workspace/GOT_weights",
+  "architectures": [
+    "GOTQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "freeze_vision_tower": false,
+  "hidden_act": "silu",
+  "hidden_size": 1024,
+  "im_end_token": 151858,
+  "im_patch_token": 151859,
+  "im_start_token": 151857,
+  "image_token_len": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 2816,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "GOT",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 16,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.37.2",
+  "use_cache": true,
+  "use_im_start_end": true,
+  "use_sliding_window": false,
+  "vision_select_layer": -1,
+  "vision_tower": "~/.cache/huggingface/hub/models--openai--clip-vit-large-patch14/snapshots/8d052a0f05efbaefbc9e8786ba291cfdf93e5bff/",
+  "vocab_size": 151860
+}

checkpoint-180774/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.37.2"
+}

checkpoint-180774/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6e637814a4af8b7bb2c9e5970aa6d0446fac977fa90d4ab100fdb36f3670242
+size 1121111672

checkpoint-180774/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70fcabda2179e380856fe9a85afe434b5c974eadc88735379106df3f30f85ef5
+size 1860083898

checkpoint-180774/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c92a95a97d689d636b085d406167a1d143dce26fb83ee64d21cf4b37a120302
+size 14244

checkpoint-180774/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10730cb78cefecd13571c16a900d439750dc91437db2ac5298d27e60f99d4f0d
+size 1064

checkpoint-180774/special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-180774/tokenization_qwen.py

@@ -0,0 +1,264 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Tokenization classes for QWen."""
+
+import base64
+import logging
+import os
+import unicodedata
+from typing import Collection, Dict, List, Set, Tuple, Union
+
+import tiktoken
+from transformers import PreTrainedTokenizer, AddedToken
+
+logger = logging.getLogger(__name__)
+
+
+VOCAB_FILES_NAMES = {"vocab_file": "qwen.tiktoken"}
+
+PAT_STR = r"""(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+"""
+ENDOFTEXT = "<|endoftext|>"
+IMSTART = "<|im_start|>"
+IMEND = "<|im_end|>"
+# as the default behavior is changed to allow special tokens in
+# regular texts, the surface forms of special tokens need to be
+# as different as possible to minimize the impact
+EXTRAS = tuple((f"<|extra_{i}|>" for i in range(205)))
+SPECIAL_TOKENS = (
+    ENDOFTEXT,
+    IMSTART,
+    IMEND,
+) + EXTRAS
+
+
+def _load_tiktoken_bpe(tiktoken_bpe_file: str) -> Dict[bytes, int]:
+    with open(tiktoken_bpe_file, "rb") as f:
+        contents = f.read()
+    return {
+        base64.b64decode(token): int(rank)
+        for token, rank in (line.split() for line in contents.splitlines() if line)
+    }
+
+class QWenTokenizer(PreTrainedTokenizer):
+    """QWen tokenizer."""
+
+    vocab_files_names = VOCAB_FILES_NAMES
+
+    def __init__(
+        self,
+        vocab_file,
+        errors="replace",
+        image_start_tag='<img>',
+        image_end_tag='</img>',
+        image_pad_tag='<imgpad>',
+        ref_start_tag='<ref>',
+        ref_end_tag='</ref>',
+        box_start_tag='<box>',
+        box_end_tag='</box>',
+        quad_start_tag='<quad>',
+        quad_end_tag='</quad>',
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        
+        self.image_start_tag = image_start_tag
+        self.image_end_tag = image_end_tag
+        self.image_pad_tag = image_pad_tag
+        self.ref_start_tag = ref_start_tag
+        self.ref_end_tag = ref_end_tag
+        self.box_start_tag = box_start_tag
+        self.box_end_tag = box_end_tag
+        self.quad_start_tag = quad_start_tag
+        self.quad_end_tag = quad_end_tag
+        self.IMAGE_ST = (
+            ref_start_tag, ref_end_tag,
+            box_start_tag, box_end_tag,
+            quad_start_tag, quad_end_tag,
+            image_start_tag, image_end_tag,
+            image_pad_tag
+        )
+
+        self.errors = errors  # how to handle errors in decoding
+
+        self.mergeable_ranks = _load_tiktoken_bpe(vocab_file)  # type: dict[bytes, int]
+        self.special_tokens = {
+            token: index
+            for index, token in enumerate(
+                SPECIAL_TOKENS + self.IMAGE_ST, start=len(self.mergeable_ranks)
+            )
+        }
+        
+        self.img_start_id = self.special_tokens[self.image_start_tag]
+        self.img_end_id = self.special_tokens[self.image_end_tag]
+        self.img_pad_id = self.special_tokens[self.image_pad_tag]
+        self.ref_start_id = self.special_tokens[self.ref_start_tag]
+        self.ref_end_id = self.special_tokens[self.ref_end_tag]
+        self.box_start_id = self.special_tokens[self.box_start_tag]
+        self.box_end_id = self.special_tokens[self.box_end_tag]
+        self.quad_start_id = self.special_tokens[self.quad_start_tag]
+        self.quad_end_id = self.special_tokens[self.quad_end_tag]
+
+        enc = tiktoken.Encoding(
+            "Qwen",
+            pat_str=PAT_STR,
+            mergeable_ranks=self.mergeable_ranks,
+            special_tokens=self.special_tokens,
+        )
+        assert (
+            len(self.mergeable_ranks) + len(self.special_tokens) == enc.n_vocab
+        ), f"{len(self.mergeable_ranks) + len(self.special_tokens)} != {enc.n_vocab} in encoding"
+
+        self.decoder = {
+            v: k for k, v in self.mergeable_ranks.items()
+        }  # type: dict[int, bytes|str]
+        self.decoder.update({v: k for k, v in self.special_tokens.items()})
+
+        self.tokenizer = enc  # type: tiktoken.Encoding
+
+        self.eod_id = self.tokenizer.eot_token
+        self.im_start_id = self.special_tokens[IMSTART]
+        self.im_end_id = self.special_tokens[IMEND]
+
+    def __len__(self) -> int:
+        return self.tokenizer.n_vocab
+
+    def get_vocab(self) -> Dict[bytes, int]:
+        return self.mergeable_ranks
+
+    def convert_tokens_to_ids(
+        self, tokens: Union[bytes, str, List[Union[bytes, str]]]
+    ) -> List[int]:
+        ids = []
+        if isinstance(tokens, (str, bytes)):
+            if tokens in self.special_tokens:
+                return self.special_tokens[tokens]
+            else:
+                return self.mergeable_ranks.get(tokens)
+        for token in tokens:
+            if token in self.special_tokens:
+                ids.append(self.special_tokens[token])
+            else:
+                ids.append(self.mergeable_ranks.get(token))
+        return ids
+
+    def _add_tokens(self, new_tokens: Union[List[str], List[AddedToken]], special_tokens: bool = False) -> int:
+        if not special_tokens and new_tokens:
+            raise ValueError('Adding regular tokens is not supported')
+        for token in new_tokens:
+            surface_form = token.content if isinstance(token, AddedToken) else token
+            if surface_form not in SPECIAL_TOKENS:
+                raise ValueError('Adding unknown special tokens is not supported')
+        return 0
+
+    def save_vocabulary(self, save_directory: str, **kwargs) -> Tuple[str]:
+        """
+        Save only the vocabulary of the tokenizer (vocabulary).
+
+        Returns:
+            `Tuple(str)`: Paths to the files saved.
+        """
+        file_path = os.path.join(save_directory, "qwen.tiktoken")
+        with open(file_path, "w", encoding="utf8") as w:
+            for k, v in self.mergeable_ranks.items():
+                line = base64.b64encode(k).decode("utf8") + " " + str(v) + "\n"
+                w.write(line)
+        return (file_path,)
+
+    def tokenize(
+        self,
+        text: str,
+        allowed_special: Union[Set, str] = "all",
+        disallowed_special: Union[Collection, str] = (),
+        **kwargs,
+    ) -> List[Union[bytes, str]]:
+        """
+        Converts a string in a sequence of tokens.
+
+        Args:
+            text (`str`):
+                The sequence to be encoded.
+            allowed_special (`Literal["all"]` or `set`):
+                The surface forms of the tokens to be encoded as special tokens in regular texts.
+                Default to "all".
+            disallowed_special (`Literal["all"]` or `Collection`):
+                The surface forms of the tokens that should not be in regular texts and trigger errors.
+                Default to an empty tuple.
+
+            kwargs (additional keyword arguments, *optional*):
+                Will be passed to the underlying model specific encode method.
+
+        Returns:
+            `List[bytes|str]`: The list of tokens.
+        """
+        tokens = []
+        text = unicodedata.normalize("NFC", text)
+
+        # this implementation takes a detour: text -> token id -> token surface forms
+        for t in self.tokenizer.encode(
+            text, allowed_special=allowed_special, disallowed_special=disallowed_special
+        ):
+            tokens.append(self.decoder[t])
+        return tokens
+
+    def convert_tokens_to_string(self, tokens: List[Union[bytes, str]]) -> str:
+        """
+        Converts a sequence of tokens in a single string.
+        """
+        text = ""
+        temp = b""
+        for t in tokens:
+            if isinstance(t, str):
+                if temp:
+                    text += temp.decode("utf-8", errors=self.errors)
+                    temp = b""
+                text += t
+            elif isinstance(t, bytes):
+                temp += t
+            else:
+                raise TypeError("token should only be of type types or str")
+        if temp:
+            text += temp.decode("utf-8", errors=self.errors)
+        return text
+
+    @property
+    def vocab_size(self):
+        return self.tokenizer.n_vocab
+
+    def _convert_id_to_token(self, index: int) -> Union[bytes, str]:
+        """Converts an id to a token, special tokens included"""
+        if index in self.decoder:
+            return self.decoder[index]
+        raise ValueError("unknown ids")
+
+    def _convert_token_to_id(self, token: Union[bytes, str]) -> int:
+        """Converts a token to an id using the vocab, special tokens included"""
+        if token in self.special_tokens:
+            return self.special_tokens[token]
+        if token in self.mergeable_ranks:
+            return self.mergeable_ranks[token]
+        raise ValueError("unknown token")
+
+    def _tokenize(self, text: str, **kwargs):
+        """
+        Converts a string in a sequence of tokens (string), using the tokenizer. Split in words for word-based
+        vocabulary or sub-words for sub-word-based vocabularies (BPE/SentencePieces/WordPieces).
+
+        Do NOT take care of added tokens.
+        """
+        raise NotImplementedError
+
+    def _decode(
+        self,
+        token_ids: Union[int, List[int]],
+        skip_special_tokens: bool = False,
+        errors: str = None,
+        **kwargs,
+    ) -> str:
+        if isinstance(token_ids, int):
+            token_ids = [token_ids]
+        if skip_special_tokens:
+            token_ids = [i for i in token_ids if i < self.eod_id]
+        return self.tokenizer.decode(token_ids, errors=errors or self.errors)

checkpoint-180774/tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "added_tokens_decoder": {},
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_qwen.QWenTokenizer",
+      null
+    ]
+  },
+  "clean_up_tokenization_spaces": true,
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "tokenizer_class": "QWenTokenizer"
+}

checkpoint-180774/trainer_state.json

@@ -0,0 +1,2187 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 180774,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9944682310509254e-05,
+      "loss": 1.5408,
+      "step": 500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.988936462101851e-05,
+      "loss": 1.2376,
+      "step": 1000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9834046931527766e-05,
+      "loss": 1.1307,
+      "step": 1500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.977872924203702e-05,
+      "loss": 1.103,
+      "step": 2000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9723411552546275e-05,
+      "loss": 1.0708,
+      "step": 2500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9668093863055527e-05,
+      "loss": 1.0288,
+      "step": 3000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9612776173564783e-05,
+      "loss": 1.0089,
+      "step": 3500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.955745848407404e-05,
+      "loss": 1.0096,
+      "step": 4000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9502140794583292e-05,
+      "loss": 1.0213,
+      "step": 4500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9446823105092548e-05,
+      "loss": 0.9962,
+      "step": 5000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9391505415601804e-05,
+      "loss": 0.9709,
+      "step": 5500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9336187726111056e-05,
+      "loss": 0.9898,
+      "step": 6000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9280870036620312e-05,
+      "loss": 0.9536,
+      "step": 6500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9225552347129565e-05,
+      "loss": 0.9591,
+      "step": 7000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.917023465763882e-05,
+      "loss": 0.9851,
+      "step": 7500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9114916968148077e-05,
+      "loss": 0.9548,
+      "step": 8000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.905959927865733e-05,
+      "loss": 1.0179,
+      "step": 8500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9004281589166585e-05,
+      "loss": 0.9737,
+      "step": 9000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.894896389967584e-05,
+      "loss": 0.9592,
+      "step": 9500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8893646210185094e-05,
+      "loss": 0.9488,
+      "step": 10000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.883832852069435e-05,
+      "loss": 0.9598,
+      "step": 10500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8783010831203603e-05,
+      "loss": 0.9572,
+      "step": 11000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.872769314171286e-05,
+      "loss": 0.9789,
+      "step": 11500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8672375452222115e-05,
+      "loss": 0.9503,
+      "step": 12000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8617057762731367e-05,
+      "loss": 0.9296,
+      "step": 12500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8561740073240623e-05,
+      "loss": 0.933,
+      "step": 13000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.850642238374988e-05,
+      "loss": 0.9438,
+      "step": 13500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.845110469425913e-05,
+      "loss": 0.919,
+      "step": 14000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.8395787004768388e-05,
+      "loss": 0.9704,
+      "step": 14500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.834046931527764e-05,
+      "loss": 0.9099,
+      "step": 15000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.8285151625786896e-05,
+      "loss": 0.9517,
+      "step": 15500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8229833936296152e-05,
+      "loss": 0.9463,
+      "step": 16000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8174516246805405e-05,
+      "loss": 0.9535,
+      "step": 16500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.8119198557314657e-05,
+      "loss": 0.9372,
+      "step": 17000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.8063880867823917e-05,
+      "loss": 0.9515,
+      "step": 17500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.800856317833317e-05,
+      "loss": 0.9675,
+      "step": 18000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.7953245488842422e-05,
+      "loss": 0.9579,
+      "step": 18500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.7897927799351678e-05,
+      "loss": 0.961,
+      "step": 19000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.7842610109860934e-05,
+      "loss": 0.9397,
+      "step": 19500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.7787292420370186e-05,
+      "loss": 0.9518,
+      "step": 20000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.7731974730879442e-05,
+      "loss": 0.9745,
+      "step": 20500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.7676657041388695e-05,
+      "loss": 0.8979,
+      "step": 21000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.762133935189795e-05,
+      "loss": 0.9224,
+      "step": 21500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.7566021662407207e-05,
+      "loss": 0.9556,
+      "step": 22000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.751070397291646e-05,
+      "loss": 0.978,
+      "step": 22500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.7455386283425715e-05,
+      "loss": 0.9102,
+      "step": 23000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.7400068593934968e-05,
+      "loss": 0.9124,
+      "step": 23500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.7344750904444224e-05,
+      "loss": 0.9089,
+      "step": 24000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.728943321495348e-05,
+      "loss": 0.9296,
+      "step": 24500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.7234115525462733e-05,
+      "loss": 0.9406,
+      "step": 25000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.717879783597199e-05,
+      "loss": 0.9625,
+      "step": 25500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7123480146481245e-05,
+      "loss": 0.9196,
+      "step": 26000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7068162456990497e-05,
+      "loss": 0.9402,
+      "step": 26500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7012844767499753e-05,
+      "loss": 0.977,
+      "step": 27000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6957527078009006e-05,
+      "loss": 0.9082,
+      "step": 27500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.690220938851826e-05,
+      "loss": 0.937,
+      "step": 28000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.6846891699027518e-05,
+      "loss": 0.8986,
+      "step": 28500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.679157400953677e-05,
+      "loss": 0.9493,
+      "step": 29000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6736256320046026e-05,
+      "loss": 0.9245,
+      "step": 29500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.6680938630555282e-05,
+      "loss": 0.8917,
+      "step": 30000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6625620941064535e-05,
+      "loss": 0.9436,
+      "step": 30500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.657030325157379e-05,
+      "loss": 0.9,
+      "step": 31000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6514985562083043e-05,
+      "loss": 0.9481,
+      "step": 31500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.64596678725923e-05,
+      "loss": 0.9146,
+      "step": 32000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6404350183101555e-05,
+      "loss": 0.9303,
+      "step": 32500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6349032493610808e-05,
+      "loss": 0.9305,
+      "step": 33000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.629371480412006e-05,
+      "loss": 0.9222,
+      "step": 33500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.623839711462932e-05,
+      "loss": 0.9495,
+      "step": 34000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6183079425138572e-05,
+      "loss": 0.8726,
+      "step": 34500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6127761735647825e-05,
+      "loss": 0.9479,
+      "step": 35000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.607244404615708e-05,
+      "loss": 0.9395,
+      "step": 35500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6017126356666337e-05,
+      "loss": 0.9261,
+      "step": 36000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.596180866717559e-05,
+      "loss": 0.8737,
+      "step": 36500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5906490977684845e-05,
+      "loss": 0.91,
+      "step": 37000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5851173288194098e-05,
+      "loss": 0.9309,
+      "step": 37500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5795855598703354e-05,
+      "loss": 0.9278,
+      "step": 38000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.574053790921261e-05,
+      "loss": 0.9506,
+      "step": 38500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5685220219721863e-05,
+      "loss": 0.8787,
+      "step": 39000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.562990253023112e-05,
+      "loss": 0.9097,
+      "step": 39500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.557458484074037e-05,
+      "loss": 0.8812,
+      "step": 40000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5519267151249627e-05,
+      "loss": 0.8798,
+      "step": 40500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5463949461758883e-05,
+      "loss": 0.9398,
+      "step": 41000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5408631772268136e-05,
+      "loss": 0.9148,
+      "step": 41500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.535331408277739e-05,
+      "loss": 0.9398,
+      "step": 42000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5297996393286648e-05,
+      "loss": 0.8984,
+      "step": 42500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5242678703795902e-05,
+      "loss": 0.8947,
+      "step": 43000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.5187361014305155e-05,
+      "loss": 0.8767,
+      "step": 43500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.5132043324814409e-05,
+      "loss": 0.906,
+      "step": 44000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.5076725635323666e-05,
+      "loss": 0.9188,
+      "step": 44500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.5021407945832919e-05,
+      "loss": 0.9534,
+      "step": 45000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4966090256342173e-05,
+      "loss": 0.913,
+      "step": 45500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4910772566851428e-05,
+      "loss": 0.8824,
+      "step": 46000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4855454877360684e-05,
+      "loss": 0.9155,
+      "step": 46500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.4800137187869938e-05,
+      "loss": 0.9142,
+      "step": 47000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.4744819498379192e-05,
+      "loss": 0.9195,
+      "step": 47500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.4689501808888446e-05,
+      "loss": 0.9207,
+      "step": 48000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.4634184119397702e-05,
+      "loss": 0.9057,
+      "step": 48500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.4578866429906957e-05,
+      "loss": 0.9273,
+      "step": 49000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.4523548740416211e-05,
+      "loss": 0.9333,
+      "step": 49500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4468231050925465e-05,
+      "loss": 0.8996,
+      "step": 50000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.4412913361434721e-05,
+      "loss": 0.9298,
+      "step": 50500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.4357595671943975e-05,
+      "loss": 0.933,
+      "step": 51000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.430227798245323e-05,
+      "loss": 0.8895,
+      "step": 51500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4246960292962484e-05,
+      "loss": 0.9297,
+      "step": 52000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.419164260347174e-05,
+      "loss": 0.9256,
+      "step": 52500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.4136324913980994e-05,
+      "loss": 0.9088,
+      "step": 53000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.4081007224490249e-05,
+      "loss": 0.9162,
+      "step": 53500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.4025689534999503e-05,
+      "loss": 0.9249,
+      "step": 54000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.3970371845508759e-05,
+      "loss": 0.9561,
+      "step": 54500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3915054156018013e-05,
+      "loss": 0.9376,
+      "step": 55000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.3859736466527267e-05,
+      "loss": 0.9404,
+      "step": 55500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.3804418777036522e-05,
+      "loss": 0.9654,
+      "step": 56000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.3749101087545778e-05,
+      "loss": 0.8642,
+      "step": 56500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3693783398055032e-05,
+      "loss": 0.9651,
+      "step": 57000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3638465708564286e-05,
+      "loss": 0.928,
+      "step": 57500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.358314801907354e-05,
+      "loss": 0.9215,
+      "step": 58000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.3527830329582796e-05,
+      "loss": 0.8917,
+      "step": 58500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.347251264009205e-05,
+      "loss": 0.8971,
+      "step": 59000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3417194950601305e-05,
+      "loss": 0.9151,
+      "step": 59500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.336187726111056e-05,
+      "loss": 0.8999,
+      "step": 60000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.3306559571619812e-05,
+      "loss": 0.9411,
+      "step": 60500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.325124188212907e-05,
+      "loss": 0.9059,
+      "step": 61000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.3195924192638324e-05,
+      "loss": 0.8741,
+      "step": 61500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.3140606503147576e-05,
+      "loss": 0.93,
+      "step": 62000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.308528881365683e-05,
+      "loss": 0.9157,
+      "step": 62500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.3029971124166088e-05,
+      "loss": 0.8834,
+      "step": 63000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.2974653434675341e-05,
+      "loss": 0.9047,
+      "step": 63500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.2919335745184595e-05,
+      "loss": 0.9396,
+      "step": 64000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.286401805569385e-05,
+      "loss": 0.9448,
+      "step": 64500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.2808700366203106e-05,
+      "loss": 0.9332,
+      "step": 65000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.275338267671236e-05,
+      "loss": 0.9394,
+      "step": 65500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.2698064987221614e-05,
+      "loss": 0.9217,
+      "step": 66000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.2642747297730868e-05,
+      "loss": 0.893,
+      "step": 66500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.2587429608240124e-05,
+      "loss": 0.8987,
+      "step": 67000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.2532111918749379e-05,
+      "loss": 0.8831,
+      "step": 67500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.2476794229258633e-05,
+      "loss": 0.9192,
+      "step": 68000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.2421476539767887e-05,
+      "loss": 0.9175,
+      "step": 68500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.2366158850277143e-05,
+      "loss": 0.9207,
+      "step": 69000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.2310841160786397e-05,
+      "loss": 0.8792,
+      "step": 69500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.2255523471295652e-05,
+      "loss": 0.9052,
+      "step": 70000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.2200205781804906e-05,
+      "loss": 0.9362,
+      "step": 70500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.2144888092314162e-05,
+      "loss": 0.8689,
+      "step": 71000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.2089570402823416e-05,
+      "loss": 0.8964,
+      "step": 71500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.203425271333267e-05,
+      "loss": 0.9057,
+      "step": 72000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.1978935023841925e-05,
+      "loss": 0.9527,
+      "step": 72500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.192361733435118e-05,
+      "loss": 0.9061,
+      "step": 73000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.1868299644860435e-05,
+      "loss": 0.9662,
+      "step": 73500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.181298195536969e-05,
+      "loss": 0.9286,
+      "step": 74000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.1757664265878944e-05,
+      "loss": 0.953,
+      "step": 74500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.17023465763882e-05,
+      "loss": 0.9109,
+      "step": 75000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.1647028886897454e-05,
+      "loss": 0.9167,
+      "step": 75500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.1591711197406708e-05,
+      "loss": 0.9382,
+      "step": 76000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.1536393507915962e-05,
+      "loss": 0.9144,
+      "step": 76500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.1481075818425218e-05,
+      "loss": 0.8935,
+      "step": 77000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.1425758128934473e-05,
+      "loss": 0.9181,
+      "step": 77500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.1370440439443727e-05,
+      "loss": 0.9167,
+      "step": 78000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.131512274995298e-05,
+      "loss": 0.8912,
+      "step": 78500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.1259805060462234e-05,
+      "loss": 0.9078,
+      "step": 79000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.1204487370971491e-05,
+      "loss": 0.9164,
+      "step": 79500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.1149169681480744e-05,
+      "loss": 0.9178,
+      "step": 80000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.1093851991989998e-05,
+      "loss": 0.8802,
+      "step": 80500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.1038534302499253e-05,
+      "loss": 0.8861,
+      "step": 81000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.0983216613008509e-05,
+      "loss": 0.9514,
+      "step": 81500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.0927898923517763e-05,
+      "loss": 0.9349,
+      "step": 82000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.0872581234027017e-05,
+      "loss": 0.8864,
+      "step": 82500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.0817263544536271e-05,
+      "loss": 0.9031,
+      "step": 83000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.0761945855045527e-05,
+      "loss": 0.9202,
+      "step": 83500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.0706628165554782e-05,
+      "loss": 0.964,
+      "step": 84000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.0651310476064036e-05,
+      "loss": 0.8934,
+      "step": 84500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.059599278657329e-05,
+      "loss": 0.8834,
+      "step": 85000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.0540675097082546e-05,
+      "loss": 0.9234,
+      "step": 85500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.04853574075918e-05,
+      "loss": 0.9243,
+      "step": 86000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.0430039718101055e-05,
+      "loss": 0.9032,
+      "step": 86500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.0374722028610309e-05,
+      "loss": 0.8856,
+      "step": 87000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.0319404339119565e-05,
+      "loss": 0.918,
+      "step": 87500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.026408664962882e-05,
+      "loss": 0.9737,
+      "step": 88000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.0208768960138074e-05,
+      "loss": 0.8813,
+      "step": 88500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.0153451270647328e-05,
+      "loss": 0.8957,
+      "step": 89000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.0098133581156584e-05,
+      "loss": 0.9095,
+      "step": 89500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.0042815891665838e-05,
+      "loss": 0.9336,
+      "step": 90000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 9.987498202175092e-06,
+      "loss": 0.8921,
+      "step": 90500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 9.932180512684347e-06,
+      "loss": 0.8593,
+      "step": 91000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 9.876862823193601e-06,
+      "loss": 0.9258,
+      "step": 91500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 9.821545133702857e-06,
+      "loss": 0.927,
+      "step": 92000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 9.766227444212111e-06,
+      "loss": 0.9453,
+      "step": 92500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 9.710909754721366e-06,
+      "loss": 0.9271,
+      "step": 93000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 9.65559206523062e-06,
+      "loss": 0.9393,
+      "step": 93500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 9.600274375739876e-06,
+      "loss": 0.9376,
+      "step": 94000
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 9.54495668624913e-06,
+      "loss": 0.929,
+      "step": 94500
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.489638996758384e-06,
+      "loss": 0.9438,
+      "step": 95000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.434321307267639e-06,
+      "loss": 0.9369,
+      "step": 95500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 9.379003617776895e-06,
+      "loss": 0.9364,
+      "step": 96000
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 9.323685928286149e-06,
+      "loss": 0.9375,
+      "step": 96500
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 9.268368238795403e-06,
+      "loss": 0.8816,
+      "step": 97000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 9.213050549304657e-06,
+      "loss": 0.916,
+      "step": 97500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.157732859813913e-06,
+      "loss": 0.994,
+      "step": 98000
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.102415170323166e-06,
+      "loss": 0.9202,
+      "step": 98500
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 9.04709748083242e-06,
+      "loss": 0.9044,
+      "step": 99000
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 8.991779791341676e-06,
+      "loss": 0.8644,
+      "step": 99500
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 8.93646210185093e-06,
+      "loss": 0.8926,
+      "step": 100000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.881144412360185e-06,
+      "loss": 0.9647,
+      "step": 100500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.825826722869439e-06,
+      "loss": 0.9277,
+      "step": 101000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.770509033378695e-06,
+      "loss": 0.8776,
+      "step": 101500
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.71519134388795e-06,
+      "loss": 0.8863,
+      "step": 102000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.659873654397204e-06,
+      "loss": 0.9286,
+      "step": 102500
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.604555964906458e-06,
+      "loss": 0.9329,
+      "step": 103000
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.549238275415714e-06,
+      "loss": 0.9091,
+      "step": 103500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.493920585924968e-06,
+      "loss": 0.9215,
+      "step": 104000
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.438602896434222e-06,
+      "loss": 0.9013,
+      "step": 104500
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.383285206943477e-06,
+      "loss": 0.9064,
+      "step": 105000
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.327967517452733e-06,
+      "loss": 0.8798,
+      "step": 105500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 8.272649827961985e-06,
+      "loss": 0.9142,
+      "step": 106000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 8.217332138471241e-06,
+      "loss": 0.9415,
+      "step": 106500
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 8.162014448980496e-06,
+      "loss": 0.8729,
+      "step": 107000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 8.10669675948975e-06,
+      "loss": 0.9351,
+      "step": 107500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 8.051379069999004e-06,
+      "loss": 0.9186,
+      "step": 108000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.99606138050826e-06,
+      "loss": 0.9713,
+      "step": 108500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.940743691017514e-06,
+      "loss": 0.9179,
+      "step": 109000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.885426001526769e-06,
+      "loss": 0.9631,
+      "step": 109500
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.830108312036023e-06,
+      "loss": 0.891,
+      "step": 110000
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.774790622545279e-06,
+      "loss": 0.9189,
+      "step": 110500
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 7.719472933054533e-06,
+      "loss": 0.8631,
+      "step": 111000
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 7.664155243563787e-06,
+      "loss": 0.9599,
+      "step": 111500
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 7.608837554073042e-06,
+      "loss": 0.9375,
+      "step": 112000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 7.553519864582297e-06,
+      "loss": 0.921,
+      "step": 112500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 7.498202175091551e-06,
+      "loss": 0.8958,
+      "step": 113000
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 7.442884485600806e-06,
+      "loss": 0.9505,
+      "step": 113500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 7.3875667961100605e-06,
+      "loss": 0.9405,
+      "step": 114000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 7.332249106619316e-06,
+      "loss": 0.8996,
+      "step": 114500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 7.27693141712857e-06,
+      "loss": 0.8697,
+      "step": 115000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.221613727637825e-06,
+      "loss": 0.9324,
+      "step": 115500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 7.166296038147079e-06,
+      "loss": 0.8865,
+      "step": 116000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 7.1109783486563345e-06,
+      "loss": 0.9143,
+      "step": 116500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 7.055660659165589e-06,
+      "loss": 0.8974,
+      "step": 117000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 7.000342969674844e-06,
+      "loss": 0.926,
+      "step": 117500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 6.945025280184098e-06,
+      "loss": 0.9321,
+      "step": 118000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 6.889707590693352e-06,
+      "loss": 0.9175,
+      "step": 118500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 6.834389901202607e-06,
+      "loss": 0.9066,
+      "step": 119000
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 6.779072211711861e-06,
+      "loss": 0.9149,
+      "step": 119500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 6.723754522221116e-06,
+      "loss": 0.9249,
+      "step": 120000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 6.66843683273037e-06,
+      "loss": 0.8616,
+      "step": 120500
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 6.6131191432396255e-06,
+      "loss": 0.933,
+      "step": 121000
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 6.55780145374888e-06,
+      "loss": 0.9157,
+      "step": 121500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 6.502483764258135e-06,
+      "loss": 0.924,
+      "step": 122000
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 6.447166074767389e-06,
+      "loss": 0.8637,
+      "step": 122500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 6.391848385276644e-06,
+      "loss": 0.9215,
+      "step": 123000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 6.336530695785899e-06,
+      "loss": 0.9098,
+      "step": 123500
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 6.281213006295154e-06,
+      "loss": 0.921,
+      "step": 124000
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 6.225895316804408e-06,
+      "loss": 0.8698,
+      "step": 124500
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 6.170577627313663e-06,
+      "loss": 0.9047,
+      "step": 125000
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 6.1152599378229175e-06,
+      "loss": 0.9108,
+      "step": 125500
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 6.059942248332173e-06,
+      "loss": 0.8666,
+      "step": 126000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 6.004624558841427e-06,
+      "loss": 0.909,
+      "step": 126500
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 5.949306869350682e-06,
+      "loss": 0.9238,
+      "step": 127000
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 5.8939891798599354e-06,
+      "loss": 0.8946,
+      "step": 127500
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 5.838671490369191e-06,
+      "loss": 0.9025,
+      "step": 128000
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 5.783353800878445e-06,
+      "loss": 0.9355,
+      "step": 128500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 5.7280361113877e-06,
+      "loss": 0.8881,
+      "step": 129000
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 5.672718421896954e-06,
+      "loss": 0.9549,
+      "step": 129500
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 5.617400732406209e-06,
+      "loss": 0.9646,
+      "step": 130000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 5.562083042915464e-06,
+      "loss": 0.9239,
+      "step": 130500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 5.506765353424719e-06,
+      "loss": 0.9045,
+      "step": 131000
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 5.451447663933973e-06,
+      "loss": 0.9572,
+      "step": 131500
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 5.396129974443228e-06,
+      "loss": 0.8585,
+      "step": 132000
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 5.3408122849524825e-06,
+      "loss": 0.9026,
+      "step": 132500
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 5.285494595461738e-06,
+      "loss": 0.9043,
+      "step": 133000
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 5.230176905970992e-06,
+      "loss": 0.9234,
+      "step": 133500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 5.174859216480247e-06,
+      "loss": 0.8722,
+      "step": 134000
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 5.119541526989501e-06,
+      "loss": 0.9212,
+      "step": 134500
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 5.064223837498756e-06,
+      "loss": 0.8981,
+      "step": 135000
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 5.008906148008011e-06,
+      "loss": 0.9532,
+      "step": 135500
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.953588458517265e-06,
+      "loss": 0.9129,
+      "step": 136000
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.898270769026519e-06,
+      "loss": 0.9075,
+      "step": 136500
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.842953079535774e-06,
+      "loss": 0.9206,
+      "step": 137000
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.787635390045029e-06,
+      "loss": 0.8512,
+      "step": 137500
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.732317700554284e-06,
+      "loss": 0.9086,
+      "step": 138000
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.677000011063538e-06,
+      "loss": 0.8828,
+      "step": 138500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.621682321572793e-06,
+      "loss": 0.9071,
+      "step": 139000
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.5663646320820475e-06,
+      "loss": 0.8957,
+      "step": 139500
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.511046942591303e-06,
+      "loss": 0.9506,
+      "step": 140000
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.455729253100557e-06,
+      "loss": 0.9241,
+      "step": 140500
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.400411563609812e-06,
+      "loss": 0.9283,
+      "step": 141000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.345093874119066e-06,
+      "loss": 0.9078,
+      "step": 141500
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.2897761846283206e-06,
+      "loss": 0.8944,
+      "step": 142000
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.234458495137576e-06,
+      "loss": 0.9283,
+      "step": 142500
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.17914080564683e-06,
+      "loss": 0.9263,
+      "step": 143000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.123823116156085e-06,
+      "loss": 0.9381,
+      "step": 143500
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.068505426665339e-06,
+      "loss": 0.9135,
+      "step": 144000
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.0131877371745945e-06,
+      "loss": 0.8499,
+      "step": 144500
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.957870047683849e-06,
+      "loss": 0.9364,
+      "step": 145000
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.902552358193104e-06,
+      "loss": 0.8836,
+      "step": 145500
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3.847234668702358e-06,
+      "loss": 0.8774,
+      "step": 146000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.791916979211613e-06,
+      "loss": 0.9224,
+      "step": 146500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.7365992897208676e-06,
+      "loss": 0.8753,
+      "step": 147000
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 3.6812816002301223e-06,
+      "loss": 0.9015,
+      "step": 147500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.625963910739376e-06,
+      "loss": 0.9001,
+      "step": 148000
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.570646221248631e-06,
+      "loss": 0.9434,
+      "step": 148500
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.5153285317578856e-06,
+      "loss": 0.9276,
+      "step": 149000
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.4600108422671403e-06,
+      "loss": 0.9056,
+      "step": 149500
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.404693152776395e-06,
+      "loss": 0.884,
+      "step": 150000
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.3493754632856497e-06,
+      "loss": 0.9658,
+      "step": 150500
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.2940577737949044e-06,
+      "loss": 0.9596,
+      "step": 151000
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.2387400843041587e-06,
+      "loss": 0.9951,
+      "step": 151500
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.1834223948134134e-06,
+      "loss": 0.8763,
+      "step": 152000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.128104705322668e-06,
+      "loss": 0.8953,
+      "step": 152500
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.0727870158319228e-06,
+      "loss": 0.9155,
+      "step": 153000
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 3.0174693263411775e-06,
+      "loss": 0.89,
+      "step": 153500
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.962151636850432e-06,
+      "loss": 0.8809,
+      "step": 154000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.906833947359687e-06,
+      "loss": 0.8995,
+      "step": 154500
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.8515162578689416e-06,
+      "loss": 0.9092,
+      "step": 155000
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.7961985683781963e-06,
+      "loss": 0.9475,
+      "step": 155500
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.740880878887451e-06,
+      "loss": 0.8896,
+      "step": 156000
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.6855631893967053e-06,
+      "loss": 0.9145,
+      "step": 156500
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.63024549990596e-06,
+      "loss": 0.9796,
+      "step": 157000
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.5749278104152147e-06,
+      "loss": 0.9451,
+      "step": 157500
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.5196101209244694e-06,
+      "loss": 0.9386,
+      "step": 158000
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.464292431433724e-06,
+      "loss": 0.9607,
+      "step": 158500
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 2.408974741942979e-06,
+      "loss": 0.9343,
+      "step": 159000
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.3536570524522335e-06,
+      "loss": 0.9211,
+      "step": 159500
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.298339362961488e-06,
+      "loss": 0.9131,
+      "step": 160000
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.243021673470743e-06,
+      "loss": 0.9738,
+      "step": 160500
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 2.187703983979997e-06,
+      "loss": 0.9394,
+      "step": 161000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.132386294489252e-06,
+      "loss": 0.9542,
+      "step": 161500
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.0770686049985066e-06,
+      "loss": 0.9253,
+      "step": 162000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.0217509155077613e-06,
+      "loss": 0.8945,
+      "step": 162500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.9664332260170156e-06,
+      "loss": 0.9523,
+      "step": 163000
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.9111155365262703e-06,
+      "loss": 0.8762,
+      "step": 163500
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.855797847035525e-06,
+      "loss": 0.8997,
+      "step": 164000
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.8004801575447797e-06,
+      "loss": 0.9035,
+      "step": 164500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.7451624680540344e-06,
+      "loss": 0.9174,
+      "step": 165000
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.689844778563289e-06,
+      "loss": 0.8572,
+      "step": 165500
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.6345270890725438e-06,
+      "loss": 0.9433,
+      "step": 166000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.5792093995817983e-06,
+      "loss": 0.9124,
+      "step": 166500
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.523891710091053e-06,
+      "loss": 0.8998,
+      "step": 167000
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.4685740206003077e-06,
+      "loss": 0.9118,
+      "step": 167500
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.4132563311095624e-06,
+      "loss": 0.9285,
+      "step": 168000
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.357938641618817e-06,
+      "loss": 0.929,
+      "step": 168500
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.3026209521280716e-06,
+      "loss": 0.9209,
+      "step": 169000
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.2473032626373263e-06,
+      "loss": 0.8879,
+      "step": 169500
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.191985573146581e-06,
+      "loss": 0.9313,
+      "step": 170000
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.1366678836558355e-06,
+      "loss": 0.9262,
+      "step": 170500
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.0813501941650902e-06,
+      "loss": 0.8769,
+      "step": 171000
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.026032504674345e-06,
+      "loss": 0.926,
+      "step": 171500
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 9.707148151835994e-07,
+      "loss": 0.8786,
+      "step": 172000
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 9.153971256928541e-07,
+      "loss": 0.8979,
+      "step": 172500
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.600794362021088e-07,
+      "loss": 0.9133,
+      "step": 173000
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.047617467113634e-07,
+      "loss": 0.9347,
+      "step": 173500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 7.494440572206181e-07,
+      "loss": 0.9344,
+      "step": 174000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 6.941263677298728e-07,
+      "loss": 0.9326,
+      "step": 174500
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 6.388086782391274e-07,
+      "loss": 0.8966,
+      "step": 175000
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 5.83490988748382e-07,
+      "loss": 0.9497,
+      "step": 175500
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 5.281732992576367e-07,
+      "loss": 0.9086,
+      "step": 176000
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 4.7285560976689126e-07,
+      "loss": 0.9177,
+      "step": 176500
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.175379202761459e-07,
+      "loss": 0.9335,
+      "step": 177000
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.6222023078540056e-07,
+      "loss": 0.8967,
+      "step": 177500
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.0690254129465527e-07,
+      "loss": 0.9551,
+      "step": 178000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.5158485180390987e-07,
+      "loss": 0.9525,
+      "step": 178500
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.9626716231316452e-07,
+      "loss": 0.9046,
+      "step": 179000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.4094947282241917e-07,
+      "loss": 0.9163,
+      "step": 179500
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 8.56317833316738e-08,
+      "loss": 0.935,
+      "step": 180000
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.0314093840928456e-08,
+      "loss": 0.8872,
+      "step": 180500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 180774,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "total_flos": 2.9686234485024e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-180774/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5ddd77616b4f164cac0cdef0d02b6f4236309ff39a32650cbcc8037f91ac321
+size 4920

checkpoint-60258/config.json

@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "/workspace/GOT_weights",
+  "architectures": [
+    "GOTQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "freeze_vision_tower": false,
+  "hidden_act": "silu",
+  "hidden_size": 1024,
+  "im_end_token": 151858,
+  "im_patch_token": 151859,
+  "im_start_token": 151857,
+  "image_token_len": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 2816,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "GOT",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 16,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.37.2",
+  "use_cache": true,
+  "use_im_start_end": true,
+  "use_sliding_window": false,
+  "vision_select_layer": -1,
+  "vision_tower": "~/.cache/huggingface/hub/models--openai--clip-vit-large-patch14/snapshots/8d052a0f05efbaefbc9e8786ba291cfdf93e5bff/",
+  "vocab_size": 151860
+}

checkpoint-60258/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.37.2"
+}

checkpoint-60258/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce7bd3dccb01ad12dfca743dfbb19f8b9cd1798a96171f2e4a5403cbc8bebde5
+size 1121111672

checkpoint-60258/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6284ee456467ea0c6cec00bf96276222a2673ecb6e42cac4cd7f56a2c14973ea
+size 1860083898

checkpoint-60258/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2e155718eaa0de9768e57816e6e6874ba931ec4489afb1c2874db91c97f7a96
+size 14244

checkpoint-60258/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87adb9a682daab18db833dd78cfcd89411bb16376e4bd3e0c3780eab48d646c7
+size 1064

checkpoint-60258/special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-60258/tokenization_qwen.py

@@ -0,0 +1,264 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Tokenization classes for QWen."""
+
+import base64
+import logging
+import os
+import unicodedata
+from typing import Collection, Dict, List, Set, Tuple, Union
+
+import tiktoken
+from transformers import PreTrainedTokenizer, AddedToken
+
+logger = logging.getLogger(__name__)
+
+
+VOCAB_FILES_NAMES = {"vocab_file": "qwen.tiktoken"}
+
+PAT_STR = r"""(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+"""
+ENDOFTEXT = "<|endoftext|>"
+IMSTART = "<|im_start|>"
+IMEND = "<|im_end|>"
+# as the default behavior is changed to allow special tokens in
+# regular texts, the surface forms of special tokens need to be
+# as different as possible to minimize the impact
+EXTRAS = tuple((f"<|extra_{i}|>" for i in range(205)))
+SPECIAL_TOKENS = (
+    ENDOFTEXT,
+    IMSTART,
+    IMEND,
+) + EXTRAS
+
+
+def _load_tiktoken_bpe(tiktoken_bpe_file: str) -> Dict[bytes, int]:
+    with open(tiktoken_bpe_file, "rb") as f:
+        contents = f.read()
+    return {
+        base64.b64decode(token): int(rank)
+        for token, rank in (line.split() for line in contents.splitlines() if line)
+    }
+
+class QWenTokenizer(PreTrainedTokenizer):
+    """QWen tokenizer."""
+
+    vocab_files_names = VOCAB_FILES_NAMES
+
+    def __init__(
+        self,
+        vocab_file,
+        errors="replace",
+        image_start_tag='<img>',
+        image_end_tag='</img>',
+        image_pad_tag='<imgpad>',
+        ref_start_tag='<ref>',
+        ref_end_tag='</ref>',
+        box_start_tag='<box>',
+        box_end_tag='</box>',
+        quad_start_tag='<quad>',
+        quad_end_tag='</quad>',
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        
+        self.image_start_tag = image_start_tag
+        self.image_end_tag = image_end_tag
+        self.image_pad_tag = image_pad_tag
+        self.ref_start_tag = ref_start_tag
+        self.ref_end_tag = ref_end_tag
+        self.box_start_tag = box_start_tag
+        self.box_end_tag = box_end_tag
+        self.quad_start_tag = quad_start_tag
+        self.quad_end_tag = quad_end_tag
+        self.IMAGE_ST = (
+            ref_start_tag, ref_end_tag,
+            box_start_tag, box_end_tag,
+            quad_start_tag, quad_end_tag,
+            image_start_tag, image_end_tag,
+            image_pad_tag
+        )
+
+        self.errors = errors  # how to handle errors in decoding
+
+        self.mergeable_ranks = _load_tiktoken_bpe(vocab_file)  # type: dict[bytes, int]
+        self.special_tokens = {
+            token: index
+            for index, token in enumerate(
+                SPECIAL_TOKENS + self.IMAGE_ST, start=len(self.mergeable_ranks)
+            )
+        }
+        
+        self.img_start_id = self.special_tokens[self.image_start_tag]
+        self.img_end_id = self.special_tokens[self.image_end_tag]
+        self.img_pad_id = self.special_tokens[self.image_pad_tag]
+        self.ref_start_id = self.special_tokens[self.ref_start_tag]
+        self.ref_end_id = self.special_tokens[self.ref_end_tag]
+        self.box_start_id = self.special_tokens[self.box_start_tag]
+        self.box_end_id = self.special_tokens[self.box_end_tag]
+        self.quad_start_id = self.special_tokens[self.quad_start_tag]
+        self.quad_end_id = self.special_tokens[self.quad_end_tag]
+
+        enc = tiktoken.Encoding(
+            "Qwen",
+            pat_str=PAT_STR,
+            mergeable_ranks=self.mergeable_ranks,
+            special_tokens=self.special_tokens,
+        )
+        assert (
+            len(self.mergeable_ranks) + len(self.special_tokens) == enc.n_vocab
+        ), f"{len(self.mergeable_ranks) + len(self.special_tokens)} != {enc.n_vocab} in encoding"
+
+        self.decoder = {
+            v: k for k, v in self.mergeable_ranks.items()
+        }  # type: dict[int, bytes|str]
+        self.decoder.update({v: k for k, v in self.special_tokens.items()})
+
+        self.tokenizer = enc  # type: tiktoken.Encoding
+
+        self.eod_id = self.tokenizer.eot_token
+        self.im_start_id = self.special_tokens[IMSTART]
+        self.im_end_id = self.special_tokens[IMEND]
+
+    def __len__(self) -> int:
+        return self.tokenizer.n_vocab
+
+    def get_vocab(self) -> Dict[bytes, int]:
+        return self.mergeable_ranks
+
+    def convert_tokens_to_ids(
+        self, tokens: Union[bytes, str, List[Union[bytes, str]]]
+    ) -> List[int]:
+        ids = []
+        if isinstance(tokens, (str, bytes)):
+            if tokens in self.special_tokens:
+                return self.special_tokens[tokens]
+            else:
+                return self.mergeable_ranks.get(tokens)
+        for token in tokens:
+            if token in self.special_tokens:
+                ids.append(self.special_tokens[token])
+            else:
+                ids.append(self.mergeable_ranks.get(token))
+        return ids
+
+    def _add_tokens(self, new_tokens: Union[List[str], List[AddedToken]], special_tokens: bool = False) -> int:
+        if not special_tokens and new_tokens:
+            raise ValueError('Adding regular tokens is not supported')
+        for token in new_tokens:
+            surface_form = token.content if isinstance(token, AddedToken) else token
+            if surface_form not in SPECIAL_TOKENS:
+                raise ValueError('Adding unknown special tokens is not supported')
+        return 0
+
+    def save_vocabulary(self, save_directory: str, **kwargs) -> Tuple[str]:
+        """
+        Save only the vocabulary of the tokenizer (vocabulary).
+
+        Returns:
+            `Tuple(str)`: Paths to the files saved.
+        """
+        file_path = os.path.join(save_directory, "qwen.tiktoken")
+        with open(file_path, "w", encoding="utf8") as w:
+            for k, v in self.mergeable_ranks.items():
+                line = base64.b64encode(k).decode("utf8") + " " + str(v) + "\n"
+                w.write(line)
+        return (file_path,)
+
+    def tokenize(
+        self,
+        text: str,
+        allowed_special: Union[Set, str] = "all",
+        disallowed_special: Union[Collection, str] = (),
+        **kwargs,
+    ) -> List[Union[bytes, str]]:
+        """
+        Converts a string in a sequence of tokens.
+
+        Args:
+            text (`str`):
+                The sequence to be encoded.
+            allowed_special (`Literal["all"]` or `set`):
+                The surface forms of the tokens to be encoded as special tokens in regular texts.
+                Default to "all".
+            disallowed_special (`Literal["all"]` or `Collection`):
+                The surface forms of the tokens that should not be in regular texts and trigger errors.
+                Default to an empty tuple.
+
+            kwargs (additional keyword arguments, *optional*):
+                Will be passed to the underlying model specific encode method.
+
+        Returns:
+            `List[bytes|str]`: The list of tokens.
+        """
+        tokens = []
+        text = unicodedata.normalize("NFC", text)
+
+        # this implementation takes a detour: text -> token id -> token surface forms
+        for t in self.tokenizer.encode(
+            text, allowed_special=allowed_special, disallowed_special=disallowed_special
+        ):
+            tokens.append(self.decoder[t])
+        return tokens
+
+    def convert_tokens_to_string(self, tokens: List[Union[bytes, str]]) -> str:
+        """
+        Converts a sequence of tokens in a single string.
+        """
+        text = ""
+        temp = b""
+        for t in tokens:
+            if isinstance(t, str):
+                if temp:
+                    text += temp.decode("utf-8", errors=self.errors)
+                    temp = b""
+                text += t
+            elif isinstance(t, bytes):
+                temp += t
+            else:
+                raise TypeError("token should only be of type types or str")
+        if temp:
+            text += temp.decode("utf-8", errors=self.errors)
+        return text
+
+    @property
+    def vocab_size(self):
+        return self.tokenizer.n_vocab
+
+    def _convert_id_to_token(self, index: int) -> Union[bytes, str]:
+        """Converts an id to a token, special tokens included"""
+        if index in self.decoder:
+            return self.decoder[index]
+        raise ValueError("unknown ids")
+
+    def _convert_token_to_id(self, token: Union[bytes, str]) -> int:
+        """Converts a token to an id using the vocab, special tokens included"""
+        if token in self.special_tokens:
+            return self.special_tokens[token]
+        if token in self.mergeable_ranks:
+            return self.mergeable_ranks[token]
+        raise ValueError("unknown token")
+
+    def _tokenize(self, text: str, **kwargs):
+        """
+        Converts a string in a sequence of tokens (string), using the tokenizer. Split in words for word-based
+        vocabulary or sub-words for sub-word-based vocabularies (BPE/SentencePieces/WordPieces).
+
+        Do NOT take care of added tokens.
+        """
+        raise NotImplementedError
+
+    def _decode(
+        self,
+        token_ids: Union[int, List[int]],
+        skip_special_tokens: bool = False,
+        errors: str = None,
+        **kwargs,
+    ) -> str:
+        if isinstance(token_ids, int):
+            token_ids = [token_ids]
+        if skip_special_tokens:
+            token_ids = [i for i in token_ids if i < self.eod_id]
+        return self.tokenizer.decode(token_ids, errors=errors or self.errors)

checkpoint-60258/tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "added_tokens_decoder": {},
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_qwen.QWenTokenizer",
+      null
+    ]
+  },
+  "clean_up_tokenization_spaces": true,
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "tokenizer_class": "QWenTokenizer"
+}

checkpoint-60258/trainer_state.json

@@ -0,0 +1,741 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 60258,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9944682310509254e-05,
+      "loss": 1.5408,
+      "step": 500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.988936462101851e-05,
+      "loss": 1.2376,
+      "step": 1000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9834046931527766e-05,
+      "loss": 1.1307,
+      "step": 1500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.977872924203702e-05,
+      "loss": 1.103,
+      "step": 2000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9723411552546275e-05,
+      "loss": 1.0708,
+      "step": 2500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9668093863055527e-05,
+      "loss": 1.0288,
+      "step": 3000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9612776173564783e-05,
+      "loss": 1.0089,
+      "step": 3500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.955745848407404e-05,
+      "loss": 1.0096,
+      "step": 4000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9502140794583292e-05,
+      "loss": 1.0213,
+      "step": 4500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9446823105092548e-05,
+      "loss": 0.9962,
+      "step": 5000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9391505415601804e-05,
+      "loss": 0.9709,
+      "step": 5500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9336187726111056e-05,
+      "loss": 0.9898,
+      "step": 6000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9280870036620312e-05,
+      "loss": 0.9536,
+      "step": 6500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9225552347129565e-05,
+      "loss": 0.9591,
+      "step": 7000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.917023465763882e-05,
+      "loss": 0.9851,
+      "step": 7500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9114916968148077e-05,
+      "loss": 0.9548,
+      "step": 8000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.905959927865733e-05,
+      "loss": 1.0179,
+      "step": 8500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9004281589166585e-05,
+      "loss": 0.9737,
+      "step": 9000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.894896389967584e-05,
+      "loss": 0.9592,
+      "step": 9500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8893646210185094e-05,
+      "loss": 0.9488,
+      "step": 10000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.883832852069435e-05,
+      "loss": 0.9598,
+      "step": 10500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8783010831203603e-05,
+      "loss": 0.9572,
+      "step": 11000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.872769314171286e-05,
+      "loss": 0.9789,
+      "step": 11500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8672375452222115e-05,
+      "loss": 0.9503,
+      "step": 12000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8617057762731367e-05,
+      "loss": 0.9296,
+      "step": 12500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8561740073240623e-05,
+      "loss": 0.933,
+      "step": 13000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.850642238374988e-05,
+      "loss": 0.9438,
+      "step": 13500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.845110469425913e-05,
+      "loss": 0.919,
+      "step": 14000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.8395787004768388e-05,
+      "loss": 0.9704,
+      "step": 14500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.834046931527764e-05,
+      "loss": 0.9099,
+      "step": 15000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.8285151625786896e-05,
+      "loss": 0.9517,
+      "step": 15500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8229833936296152e-05,
+      "loss": 0.9463,
+      "step": 16000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8174516246805405e-05,
+      "loss": 0.9535,
+      "step": 16500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.8119198557314657e-05,
+      "loss": 0.9372,
+      "step": 17000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.8063880867823917e-05,
+      "loss": 0.9515,
+      "step": 17500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.800856317833317e-05,
+      "loss": 0.9675,
+      "step": 18000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.7953245488842422e-05,
+      "loss": 0.9579,
+      "step": 18500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.7897927799351678e-05,
+      "loss": 0.961,
+      "step": 19000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.7842610109860934e-05,
+      "loss": 0.9397,
+      "step": 19500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.7787292420370186e-05,
+      "loss": 0.9518,
+      "step": 20000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.7731974730879442e-05,
+      "loss": 0.9745,
+      "step": 20500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.7676657041388695e-05,
+      "loss": 0.8979,
+      "step": 21000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.762133935189795e-05,
+      "loss": 0.9224,
+      "step": 21500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.7566021662407207e-05,
+      "loss": 0.9556,
+      "step": 22000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.751070397291646e-05,
+      "loss": 0.978,
+      "step": 22500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.7455386283425715e-05,
+      "loss": 0.9102,
+      "step": 23000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.7400068593934968e-05,
+      "loss": 0.9124,
+      "step": 23500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.7344750904444224e-05,
+      "loss": 0.9089,
+      "step": 24000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.728943321495348e-05,
+      "loss": 0.9296,
+      "step": 24500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.7234115525462733e-05,
+      "loss": 0.9406,
+      "step": 25000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.717879783597199e-05,
+      "loss": 0.9625,
+      "step": 25500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7123480146481245e-05,
+      "loss": 0.9196,
+      "step": 26000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7068162456990497e-05,
+      "loss": 0.9402,
+      "step": 26500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7012844767499753e-05,
+      "loss": 0.977,
+      "step": 27000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6957527078009006e-05,
+      "loss": 0.9082,
+      "step": 27500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.690220938851826e-05,
+      "loss": 0.937,
+      "step": 28000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.6846891699027518e-05,
+      "loss": 0.8986,
+      "step": 28500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.679157400953677e-05,
+      "loss": 0.9493,
+      "step": 29000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6736256320046026e-05,
+      "loss": 0.9245,
+      "step": 29500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.6680938630555282e-05,
+      "loss": 0.8917,
+      "step": 30000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6625620941064535e-05,
+      "loss": 0.9436,
+      "step": 30500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.657030325157379e-05,
+      "loss": 0.9,
+      "step": 31000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6514985562083043e-05,
+      "loss": 0.9481,
+      "step": 31500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.64596678725923e-05,
+      "loss": 0.9146,
+      "step": 32000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6404350183101555e-05,
+      "loss": 0.9303,
+      "step": 32500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6349032493610808e-05,
+      "loss": 0.9305,
+      "step": 33000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.629371480412006e-05,
+      "loss": 0.9222,
+      "step": 33500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.623839711462932e-05,
+      "loss": 0.9495,
+      "step": 34000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6183079425138572e-05,
+      "loss": 0.8726,
+      "step": 34500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6127761735647825e-05,
+      "loss": 0.9479,
+      "step": 35000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.607244404615708e-05,
+      "loss": 0.9395,
+      "step": 35500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6017126356666337e-05,
+      "loss": 0.9261,
+      "step": 36000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.596180866717559e-05,
+      "loss": 0.8737,
+      "step": 36500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5906490977684845e-05,
+      "loss": 0.91,
+      "step": 37000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5851173288194098e-05,
+      "loss": 0.9309,
+      "step": 37500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5795855598703354e-05,
+      "loss": 0.9278,
+      "step": 38000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.574053790921261e-05,
+      "loss": 0.9506,
+      "step": 38500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5685220219721863e-05,
+      "loss": 0.8787,
+      "step": 39000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.562990253023112e-05,
+      "loss": 0.9097,
+      "step": 39500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.557458484074037e-05,
+      "loss": 0.8812,
+      "step": 40000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5519267151249627e-05,
+      "loss": 0.8798,
+      "step": 40500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5463949461758883e-05,
+      "loss": 0.9398,
+      "step": 41000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5408631772268136e-05,
+      "loss": 0.9148,
+      "step": 41500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.535331408277739e-05,
+      "loss": 0.9398,
+      "step": 42000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5297996393286648e-05,
+      "loss": 0.8984,
+      "step": 42500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5242678703795902e-05,
+      "loss": 0.8947,
+      "step": 43000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.5187361014305155e-05,
+      "loss": 0.8767,
+      "step": 43500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.5132043324814409e-05,
+      "loss": 0.906,
+      "step": 44000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.5076725635323666e-05,
+      "loss": 0.9188,
+      "step": 44500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.5021407945832919e-05,
+      "loss": 0.9534,
+      "step": 45000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4966090256342173e-05,
+      "loss": 0.913,
+      "step": 45500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4910772566851428e-05,
+      "loss": 0.8824,
+      "step": 46000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4855454877360684e-05,
+      "loss": 0.9155,
+      "step": 46500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.4800137187869938e-05,
+      "loss": 0.9142,
+      "step": 47000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.4744819498379192e-05,
+      "loss": 0.9195,
+      "step": 47500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.4689501808888446e-05,
+      "loss": 0.9207,
+      "step": 48000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.4634184119397702e-05,
+      "loss": 0.9057,
+      "step": 48500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.4578866429906957e-05,
+      "loss": 0.9273,
+      "step": 49000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.4523548740416211e-05,
+      "loss": 0.9333,
+      "step": 49500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4468231050925465e-05,
+      "loss": 0.8996,
+      "step": 50000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.4412913361434721e-05,
+      "loss": 0.9298,
+      "step": 50500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.4357595671943975e-05,
+      "loss": 0.933,
+      "step": 51000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.430227798245323e-05,
+      "loss": 0.8895,
+      "step": 51500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4246960292962484e-05,
+      "loss": 0.9297,
+      "step": 52000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.419164260347174e-05,
+      "loss": 0.9256,
+      "step": 52500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.4136324913980994e-05,
+      "loss": 0.9088,
+      "step": 53000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.4081007224490249e-05,
+      "loss": 0.9162,
+      "step": 53500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.4025689534999503e-05,
+      "loss": 0.9249,
+      "step": 54000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.3970371845508759e-05,
+      "loss": 0.9561,
+      "step": 54500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3915054156018013e-05,
+      "loss": 0.9376,
+      "step": 55000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.3859736466527267e-05,
+      "loss": 0.9404,
+      "step": 55500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.3804418777036522e-05,
+      "loss": 0.9654,
+      "step": 56000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.3749101087545778e-05,
+      "loss": 0.8642,
+      "step": 56500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3693783398055032e-05,
+      "loss": 0.9651,
+      "step": 57000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3638465708564286e-05,
+      "loss": 0.928,
+      "step": 57500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.358314801907354e-05,
+      "loss": 0.9215,
+      "step": 58000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.3527830329582796e-05,
+      "loss": 0.8917,
+      "step": 58500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.347251264009205e-05,
+      "loss": 0.8971,
+      "step": 59000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3417194950601305e-05,
+      "loss": 0.9151,
+      "step": 59500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.336187726111056e-05,
+      "loss": 0.8999,
+      "step": 60000
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 180774,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "total_flos": 9.895411495008e+16,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-60258/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5ddd77616b4f164cac0cdef0d02b6f4236309ff39a32650cbcc8037f91ac321
+size 4920

config.json

@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "/workspace/GOT_weights",
+  "architectures": [
+    "GOTQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "freeze_vision_tower": false,
+  "hidden_act": "silu",
+  "hidden_size": 1024,
+  "im_end_token": 151858,
+  "im_patch_token": 151859,
+  "im_start_token": 151857,
+  "image_token_len": 256,
+  "initializer_range": 0.02,
+  "intermediate_size": 2816,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "GOT",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 16,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.37.2",
+  "use_cache": true,
+  "use_im_start_end": true,
+  "use_sliding_window": false,
+  "vision_select_layer": -1,
+  "vision_tower": "~/.cache/huggingface/hub/models--openai--clip-vit-large-patch14/snapshots/8d052a0f05efbaefbc9e8786ba291cfdf93e5bff/",
+  "vocab_size": 151860
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.37.2"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39c5cc3daa858bf5d4aa8de5c6971c5f2fef649226407b8ee24f200b2fc62f8d
+size 1432121416

special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenization_qwen.py

@@ -0,0 +1,264 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Tokenization classes for QWen."""
+
+import base64
+import logging
+import os
+import unicodedata
+from typing import Collection, Dict, List, Set, Tuple, Union
+
+import tiktoken
+from transformers import PreTrainedTokenizer, AddedToken
+
+logger = logging.getLogger(__name__)
+
+
+VOCAB_FILES_NAMES = {"vocab_file": "qwen.tiktoken"}
+
+PAT_STR = r"""(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+"""
+ENDOFTEXT = "<|endoftext|>"
+IMSTART = "<|im_start|>"
+IMEND = "<|im_end|>"
+# as the default behavior is changed to allow special tokens in
+# regular texts, the surface forms of special tokens need to be
+# as different as possible to minimize the impact
+EXTRAS = tuple((f"<|extra_{i}|>" for i in range(205)))
+SPECIAL_TOKENS = (
+    ENDOFTEXT,
+    IMSTART,
+    IMEND,
+) + EXTRAS
+
+
+def _load_tiktoken_bpe(tiktoken_bpe_file: str) -> Dict[bytes, int]:
+    with open(tiktoken_bpe_file, "rb") as f:
+        contents = f.read()
+    return {
+        base64.b64decode(token): int(rank)
+        for token, rank in (line.split() for line in contents.splitlines() if line)
+    }
+
+class QWenTokenizer(PreTrainedTokenizer):
+    """QWen tokenizer."""
+
+    vocab_files_names = VOCAB_FILES_NAMES
+
+    def __init__(
+        self,
+        vocab_file,
+        errors="replace",
+        image_start_tag='<img>',
+        image_end_tag='</img>',
+        image_pad_tag='<imgpad>',
+        ref_start_tag='<ref>',
+        ref_end_tag='</ref>',
+        box_start_tag='<box>',
+        box_end_tag='</box>',
+        quad_start_tag='<quad>',
+        quad_end_tag='</quad>',
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        
+        self.image_start_tag = image_start_tag
+        self.image_end_tag = image_end_tag
+        self.image_pad_tag = image_pad_tag
+        self.ref_start_tag = ref_start_tag
+        self.ref_end_tag = ref_end_tag
+        self.box_start_tag = box_start_tag
+        self.box_end_tag = box_end_tag
+        self.quad_start_tag = quad_start_tag
+        self.quad_end_tag = quad_end_tag
+        self.IMAGE_ST = (
+            ref_start_tag, ref_end_tag,
+            box_start_tag, box_end_tag,
+            quad_start_tag, quad_end_tag,
+            image_start_tag, image_end_tag,
+            image_pad_tag
+        )
+
+        self.errors = errors  # how to handle errors in decoding
+
+        self.mergeable_ranks = _load_tiktoken_bpe(vocab_file)  # type: dict[bytes, int]
+        self.special_tokens = {
+            token: index
+            for index, token in enumerate(
+                SPECIAL_TOKENS + self.IMAGE_ST, start=len(self.mergeable_ranks)
+            )
+        }
+        
+        self.img_start_id = self.special_tokens[self.image_start_tag]
+        self.img_end_id = self.special_tokens[self.image_end_tag]
+        self.img_pad_id = self.special_tokens[self.image_pad_tag]
+        self.ref_start_id = self.special_tokens[self.ref_start_tag]
+        self.ref_end_id = self.special_tokens[self.ref_end_tag]
+        self.box_start_id = self.special_tokens[self.box_start_tag]
+        self.box_end_id = self.special_tokens[self.box_end_tag]
+        self.quad_start_id = self.special_tokens[self.quad_start_tag]
+        self.quad_end_id = self.special_tokens[self.quad_end_tag]
+
+        enc = tiktoken.Encoding(
+            "Qwen",
+            pat_str=PAT_STR,
+            mergeable_ranks=self.mergeable_ranks,
+            special_tokens=self.special_tokens,
+        )
+        assert (
+            len(self.mergeable_ranks) + len(self.special_tokens) == enc.n_vocab
+        ), f"{len(self.mergeable_ranks) + len(self.special_tokens)} != {enc.n_vocab} in encoding"
+
+        self.decoder = {
+            v: k for k, v in self.mergeable_ranks.items()
+        }  # type: dict[int, bytes|str]
+        self.decoder.update({v: k for k, v in self.special_tokens.items()})
+
+        self.tokenizer = enc  # type: tiktoken.Encoding
+
+        self.eod_id = self.tokenizer.eot_token
+        self.im_start_id = self.special_tokens[IMSTART]
+        self.im_end_id = self.special_tokens[IMEND]
+
+    def __len__(self) -> int:
+        return self.tokenizer.n_vocab
+
+    def get_vocab(self) -> Dict[bytes, int]:
+        return self.mergeable_ranks
+
+    def convert_tokens_to_ids(
+        self, tokens: Union[bytes, str, List[Union[bytes, str]]]
+    ) -> List[int]:
+        ids = []
+        if isinstance(tokens, (str, bytes)):
+            if tokens in self.special_tokens:
+                return self.special_tokens[tokens]
+            else:
+                return self.mergeable_ranks.get(tokens)
+        for token in tokens:
+            if token in self.special_tokens:
+                ids.append(self.special_tokens[token])
+            else:
+                ids.append(self.mergeable_ranks.get(token))
+        return ids
+
+    def _add_tokens(self, new_tokens: Union[List[str], List[AddedToken]], special_tokens: bool = False) -> int:
+        if not special_tokens and new_tokens:
+            raise ValueError('Adding regular tokens is not supported')
+        for token in new_tokens:
+            surface_form = token.content if isinstance(token, AddedToken) else token
+            if surface_form not in SPECIAL_TOKENS:
+                raise ValueError('Adding unknown special tokens is not supported')
+        return 0
+
+    def save_vocabulary(self, save_directory: str, **kwargs) -> Tuple[str]:
+        """
+        Save only the vocabulary of the tokenizer (vocabulary).
+
+        Returns:
+            `Tuple(str)`: Paths to the files saved.
+        """
+        file_path = os.path.join(save_directory, "qwen.tiktoken")
+        with open(file_path, "w", encoding="utf8") as w:
+            for k, v in self.mergeable_ranks.items():
+                line = base64.b64encode(k).decode("utf8") + " " + str(v) + "\n"
+                w.write(line)
+        return (file_path,)
+
+    def tokenize(
+        self,
+        text: str,
+        allowed_special: Union[Set, str] = "all",
+        disallowed_special: Union[Collection, str] = (),
+        **kwargs,
+    ) -> List[Union[bytes, str]]:
+        """
+        Converts a string in a sequence of tokens.
+
+        Args:
+            text (`str`):
+                The sequence to be encoded.
+            allowed_special (`Literal["all"]` or `set`):
+                The surface forms of the tokens to be encoded as special tokens in regular texts.
+                Default to "all".
+            disallowed_special (`Literal["all"]` or `Collection`):
+                The surface forms of the tokens that should not be in regular texts and trigger errors.
+                Default to an empty tuple.
+
+            kwargs (additional keyword arguments, *optional*):
+                Will be passed to the underlying model specific encode method.
+
+        Returns:
+            `List[bytes|str]`: The list of tokens.
+        """
+        tokens = []
+        text = unicodedata.normalize("NFC", text)
+
+        # this implementation takes a detour: text -> token id -> token surface forms
+        for t in self.tokenizer.encode(
+            text, allowed_special=allowed_special, disallowed_special=disallowed_special
+        ):
+            tokens.append(self.decoder[t])
+        return tokens
+
+    def convert_tokens_to_string(self, tokens: List[Union[bytes, str]]) -> str:
+        """
+        Converts a sequence of tokens in a single string.
+        """
+        text = ""
+        temp = b""
+        for t in tokens:
+            if isinstance(t, str):
+                if temp:
+                    text += temp.decode("utf-8", errors=self.errors)
+                    temp = b""
+                text += t
+            elif isinstance(t, bytes):
+                temp += t
+            else:
+                raise TypeError("token should only be of type types or str")
+        if temp:
+            text += temp.decode("utf-8", errors=self.errors)
+        return text
+
+    @property
+    def vocab_size(self):
+        return self.tokenizer.n_vocab
+
+    def _convert_id_to_token(self, index: int) -> Union[bytes, str]:
+        """Converts an id to a token, special tokens included"""
+        if index in self.decoder:
+            return self.decoder[index]
+        raise ValueError("unknown ids")
+
+    def _convert_token_to_id(self, token: Union[bytes, str]) -> int:
+        """Converts a token to an id using the vocab, special tokens included"""
+        if token in self.special_tokens:
+            return self.special_tokens[token]
+        if token in self.mergeable_ranks:
+            return self.mergeable_ranks[token]
+        raise ValueError("unknown token")
+
+    def _tokenize(self, text: str, **kwargs):
+        """
+        Converts a string in a sequence of tokens (string), using the tokenizer. Split in words for word-based
+        vocabulary or sub-words for sub-word-based vocabularies (BPE/SentencePieces/WordPieces).
+
+        Do NOT take care of added tokens.
+        """
+        raise NotImplementedError
+
+    def _decode(
+        self,
+        token_ids: Union[int, List[int]],
+        skip_special_tokens: bool = False,
+        errors: str = None,
+        **kwargs,
+    ) -> str:
+        if isinstance(token_ids, int):
+            token_ids = [token_ids]
+        if skip_special_tokens:
+            token_ids = [i for i in token_ids if i < self.eod_id]
+        return self.tokenizer.decode(token_ids, errors=errors or self.errors)

tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "added_tokens_decoder": {},
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_qwen.QWenTokenizer",
+      null
+    ]
+  },
+  "clean_up_tokenization_spaces": true,
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "tokenizer_class": "QWenTokenizer"
+}

trainer_state.json

@@ -0,0 +1,2196 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 180774,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9944682310509254e-05,
+      "loss": 1.5408,
+      "step": 500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.988936462101851e-05,
+      "loss": 1.2376,
+      "step": 1000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.9834046931527766e-05,
+      "loss": 1.1307,
+      "step": 1500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.977872924203702e-05,
+      "loss": 1.103,
+      "step": 2000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9723411552546275e-05,
+      "loss": 1.0708,
+      "step": 2500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9668093863055527e-05,
+      "loss": 1.0288,
+      "step": 3000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9612776173564783e-05,
+      "loss": 1.0089,
+      "step": 3500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.955745848407404e-05,
+      "loss": 1.0096,
+      "step": 4000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9502140794583292e-05,
+      "loss": 1.0213,
+      "step": 4500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9446823105092548e-05,
+      "loss": 0.9962,
+      "step": 5000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9391505415601804e-05,
+      "loss": 0.9709,
+      "step": 5500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9336187726111056e-05,
+      "loss": 0.9898,
+      "step": 6000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9280870036620312e-05,
+      "loss": 0.9536,
+      "step": 6500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9225552347129565e-05,
+      "loss": 0.9591,
+      "step": 7000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.917023465763882e-05,
+      "loss": 0.9851,
+      "step": 7500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9114916968148077e-05,
+      "loss": 0.9548,
+      "step": 8000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.905959927865733e-05,
+      "loss": 1.0179,
+      "step": 8500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9004281589166585e-05,
+      "loss": 0.9737,
+      "step": 9000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.894896389967584e-05,
+      "loss": 0.9592,
+      "step": 9500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8893646210185094e-05,
+      "loss": 0.9488,
+      "step": 10000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.883832852069435e-05,
+      "loss": 0.9598,
+      "step": 10500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8783010831203603e-05,
+      "loss": 0.9572,
+      "step": 11000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.872769314171286e-05,
+      "loss": 0.9789,
+      "step": 11500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8672375452222115e-05,
+      "loss": 0.9503,
+      "step": 12000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8617057762731367e-05,
+      "loss": 0.9296,
+      "step": 12500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8561740073240623e-05,
+      "loss": 0.933,
+      "step": 13000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.850642238374988e-05,
+      "loss": 0.9438,
+      "step": 13500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.845110469425913e-05,
+      "loss": 0.919,
+      "step": 14000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.8395787004768388e-05,
+      "loss": 0.9704,
+      "step": 14500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.834046931527764e-05,
+      "loss": 0.9099,
+      "step": 15000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.8285151625786896e-05,
+      "loss": 0.9517,
+      "step": 15500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8229833936296152e-05,
+      "loss": 0.9463,
+      "step": 16000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8174516246805405e-05,
+      "loss": 0.9535,
+      "step": 16500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.8119198557314657e-05,
+      "loss": 0.9372,
+      "step": 17000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.8063880867823917e-05,
+      "loss": 0.9515,
+      "step": 17500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.800856317833317e-05,
+      "loss": 0.9675,
+      "step": 18000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.7953245488842422e-05,
+      "loss": 0.9579,
+      "step": 18500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.7897927799351678e-05,
+      "loss": 0.961,
+      "step": 19000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.7842610109860934e-05,
+      "loss": 0.9397,
+      "step": 19500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.7787292420370186e-05,
+      "loss": 0.9518,
+      "step": 20000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.7731974730879442e-05,
+      "loss": 0.9745,
+      "step": 20500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.7676657041388695e-05,
+      "loss": 0.8979,
+      "step": 21000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.762133935189795e-05,
+      "loss": 0.9224,
+      "step": 21500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.7566021662407207e-05,
+      "loss": 0.9556,
+      "step": 22000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.751070397291646e-05,
+      "loss": 0.978,
+      "step": 22500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.7455386283425715e-05,
+      "loss": 0.9102,
+      "step": 23000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.7400068593934968e-05,
+      "loss": 0.9124,
+      "step": 23500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.7344750904444224e-05,
+      "loss": 0.9089,
+      "step": 24000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.728943321495348e-05,
+      "loss": 0.9296,
+      "step": 24500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.7234115525462733e-05,
+      "loss": 0.9406,
+      "step": 25000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.717879783597199e-05,
+      "loss": 0.9625,
+      "step": 25500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.7123480146481245e-05,
+      "loss": 0.9196,
+      "step": 26000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.7068162456990497e-05,
+      "loss": 0.9402,
+      "step": 26500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7012844767499753e-05,
+      "loss": 0.977,
+      "step": 27000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6957527078009006e-05,
+      "loss": 0.9082,
+      "step": 27500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.690220938851826e-05,
+      "loss": 0.937,
+      "step": 28000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.6846891699027518e-05,
+      "loss": 0.8986,
+      "step": 28500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.679157400953677e-05,
+      "loss": 0.9493,
+      "step": 29000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.6736256320046026e-05,
+      "loss": 0.9245,
+      "step": 29500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.6680938630555282e-05,
+      "loss": 0.8917,
+      "step": 30000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.6625620941064535e-05,
+      "loss": 0.9436,
+      "step": 30500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.657030325157379e-05,
+      "loss": 0.9,
+      "step": 31000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.6514985562083043e-05,
+      "loss": 0.9481,
+      "step": 31500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.64596678725923e-05,
+      "loss": 0.9146,
+      "step": 32000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.6404350183101555e-05,
+      "loss": 0.9303,
+      "step": 32500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6349032493610808e-05,
+      "loss": 0.9305,
+      "step": 33000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.629371480412006e-05,
+      "loss": 0.9222,
+      "step": 33500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.623839711462932e-05,
+      "loss": 0.9495,
+      "step": 34000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6183079425138572e-05,
+      "loss": 0.8726,
+      "step": 34500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6127761735647825e-05,
+      "loss": 0.9479,
+      "step": 35000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.607244404615708e-05,
+      "loss": 0.9395,
+      "step": 35500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6017126356666337e-05,
+      "loss": 0.9261,
+      "step": 36000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.596180866717559e-05,
+      "loss": 0.8737,
+      "step": 36500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.5906490977684845e-05,
+      "loss": 0.91,
+      "step": 37000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.5851173288194098e-05,
+      "loss": 0.9309,
+      "step": 37500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5795855598703354e-05,
+      "loss": 0.9278,
+      "step": 38000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.574053790921261e-05,
+      "loss": 0.9506,
+      "step": 38500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5685220219721863e-05,
+      "loss": 0.8787,
+      "step": 39000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.562990253023112e-05,
+      "loss": 0.9097,
+      "step": 39500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.557458484074037e-05,
+      "loss": 0.8812,
+      "step": 40000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5519267151249627e-05,
+      "loss": 0.8798,
+      "step": 40500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5463949461758883e-05,
+      "loss": 0.9398,
+      "step": 41000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5408631772268136e-05,
+      "loss": 0.9148,
+      "step": 41500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.535331408277739e-05,
+      "loss": 0.9398,
+      "step": 42000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5297996393286648e-05,
+      "loss": 0.8984,
+      "step": 42500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5242678703795902e-05,
+      "loss": 0.8947,
+      "step": 43000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.5187361014305155e-05,
+      "loss": 0.8767,
+      "step": 43500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.5132043324814409e-05,
+      "loss": 0.906,
+      "step": 44000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.5076725635323666e-05,
+      "loss": 0.9188,
+      "step": 44500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.5021407945832919e-05,
+      "loss": 0.9534,
+      "step": 45000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4966090256342173e-05,
+      "loss": 0.913,
+      "step": 45500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4910772566851428e-05,
+      "loss": 0.8824,
+      "step": 46000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4855454877360684e-05,
+      "loss": 0.9155,
+      "step": 46500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.4800137187869938e-05,
+      "loss": 0.9142,
+      "step": 47000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.4744819498379192e-05,
+      "loss": 0.9195,
+      "step": 47500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.4689501808888446e-05,
+      "loss": 0.9207,
+      "step": 48000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.4634184119397702e-05,
+      "loss": 0.9057,
+      "step": 48500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.4578866429906957e-05,
+      "loss": 0.9273,
+      "step": 49000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.4523548740416211e-05,
+      "loss": 0.9333,
+      "step": 49500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4468231050925465e-05,
+      "loss": 0.8996,
+      "step": 50000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.4412913361434721e-05,
+      "loss": 0.9298,
+      "step": 50500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.4357595671943975e-05,
+      "loss": 0.933,
+      "step": 51000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.430227798245323e-05,
+      "loss": 0.8895,
+      "step": 51500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.4246960292962484e-05,
+      "loss": 0.9297,
+      "step": 52000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.419164260347174e-05,
+      "loss": 0.9256,
+      "step": 52500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.4136324913980994e-05,
+      "loss": 0.9088,
+      "step": 53000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.4081007224490249e-05,
+      "loss": 0.9162,
+      "step": 53500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.4025689534999503e-05,
+      "loss": 0.9249,
+      "step": 54000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.3970371845508759e-05,
+      "loss": 0.9561,
+      "step": 54500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3915054156018013e-05,
+      "loss": 0.9376,
+      "step": 55000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.3859736466527267e-05,
+      "loss": 0.9404,
+      "step": 55500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.3804418777036522e-05,
+      "loss": 0.9654,
+      "step": 56000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.3749101087545778e-05,
+      "loss": 0.8642,
+      "step": 56500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3693783398055032e-05,
+      "loss": 0.9651,
+      "step": 57000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3638465708564286e-05,
+      "loss": 0.928,
+      "step": 57500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.358314801907354e-05,
+      "loss": 0.9215,
+      "step": 58000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.3527830329582796e-05,
+      "loss": 0.8917,
+      "step": 58500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.347251264009205e-05,
+      "loss": 0.8971,
+      "step": 59000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3417194950601305e-05,
+      "loss": 0.9151,
+      "step": 59500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.336187726111056e-05,
+      "loss": 0.8999,
+      "step": 60000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.3306559571619812e-05,
+      "loss": 0.9411,
+      "step": 60500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.325124188212907e-05,
+      "loss": 0.9059,
+      "step": 61000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.3195924192638324e-05,
+      "loss": 0.8741,
+      "step": 61500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.3140606503147576e-05,
+      "loss": 0.93,
+      "step": 62000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.308528881365683e-05,
+      "loss": 0.9157,
+      "step": 62500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.3029971124166088e-05,
+      "loss": 0.8834,
+      "step": 63000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.2974653434675341e-05,
+      "loss": 0.9047,
+      "step": 63500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.2919335745184595e-05,
+      "loss": 0.9396,
+      "step": 64000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.286401805569385e-05,
+      "loss": 0.9448,
+      "step": 64500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.2808700366203106e-05,
+      "loss": 0.9332,
+      "step": 65000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.275338267671236e-05,
+      "loss": 0.9394,
+      "step": 65500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.2698064987221614e-05,
+      "loss": 0.9217,
+      "step": 66000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.2642747297730868e-05,
+      "loss": 0.893,
+      "step": 66500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.2587429608240124e-05,
+      "loss": 0.8987,
+      "step": 67000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.2532111918749379e-05,
+      "loss": 0.8831,
+      "step": 67500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.2476794229258633e-05,
+      "loss": 0.9192,
+      "step": 68000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.2421476539767887e-05,
+      "loss": 0.9175,
+      "step": 68500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.2366158850277143e-05,
+      "loss": 0.9207,
+      "step": 69000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.2310841160786397e-05,
+      "loss": 0.8792,
+      "step": 69500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.2255523471295652e-05,
+      "loss": 0.9052,
+      "step": 70000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.2200205781804906e-05,
+      "loss": 0.9362,
+      "step": 70500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.2144888092314162e-05,
+      "loss": 0.8689,
+      "step": 71000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.2089570402823416e-05,
+      "loss": 0.8964,
+      "step": 71500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.203425271333267e-05,
+      "loss": 0.9057,
+      "step": 72000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.1978935023841925e-05,
+      "loss": 0.9527,
+      "step": 72500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.192361733435118e-05,
+      "loss": 0.9061,
+      "step": 73000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.1868299644860435e-05,
+      "loss": 0.9662,
+      "step": 73500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.181298195536969e-05,
+      "loss": 0.9286,
+      "step": 74000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.1757664265878944e-05,
+      "loss": 0.953,
+      "step": 74500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.17023465763882e-05,
+      "loss": 0.9109,
+      "step": 75000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.1647028886897454e-05,
+      "loss": 0.9167,
+      "step": 75500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.1591711197406708e-05,
+      "loss": 0.9382,
+      "step": 76000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.1536393507915962e-05,
+      "loss": 0.9144,
+      "step": 76500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.1481075818425218e-05,
+      "loss": 0.8935,
+      "step": 77000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.1425758128934473e-05,
+      "loss": 0.9181,
+      "step": 77500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.1370440439443727e-05,
+      "loss": 0.9167,
+      "step": 78000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.131512274995298e-05,
+      "loss": 0.8912,
+      "step": 78500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.1259805060462234e-05,
+      "loss": 0.9078,
+      "step": 79000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.1204487370971491e-05,
+      "loss": 0.9164,
+      "step": 79500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.1149169681480744e-05,
+      "loss": 0.9178,
+      "step": 80000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.1093851991989998e-05,
+      "loss": 0.8802,
+      "step": 80500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.1038534302499253e-05,
+      "loss": 0.8861,
+      "step": 81000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.0983216613008509e-05,
+      "loss": 0.9514,
+      "step": 81500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.0927898923517763e-05,
+      "loss": 0.9349,
+      "step": 82000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.0872581234027017e-05,
+      "loss": 0.8864,
+      "step": 82500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.0817263544536271e-05,
+      "loss": 0.9031,
+      "step": 83000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.0761945855045527e-05,
+      "loss": 0.9202,
+      "step": 83500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.0706628165554782e-05,
+      "loss": 0.964,
+      "step": 84000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.0651310476064036e-05,
+      "loss": 0.8934,
+      "step": 84500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.059599278657329e-05,
+      "loss": 0.8834,
+      "step": 85000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.0540675097082546e-05,
+      "loss": 0.9234,
+      "step": 85500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.04853574075918e-05,
+      "loss": 0.9243,
+      "step": 86000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.0430039718101055e-05,
+      "loss": 0.9032,
+      "step": 86500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.0374722028610309e-05,
+      "loss": 0.8856,
+      "step": 87000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.0319404339119565e-05,
+      "loss": 0.918,
+      "step": 87500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.026408664962882e-05,
+      "loss": 0.9737,
+      "step": 88000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.0208768960138074e-05,
+      "loss": 0.8813,
+      "step": 88500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.0153451270647328e-05,
+      "loss": 0.8957,
+      "step": 89000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.0098133581156584e-05,
+      "loss": 0.9095,
+      "step": 89500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.0042815891665838e-05,
+      "loss": 0.9336,
+      "step": 90000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 9.987498202175092e-06,
+      "loss": 0.8921,
+      "step": 90500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 9.932180512684347e-06,
+      "loss": 0.8593,
+      "step": 91000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 9.876862823193601e-06,
+      "loss": 0.9258,
+      "step": 91500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 9.821545133702857e-06,
+      "loss": 0.927,
+      "step": 92000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 9.766227444212111e-06,
+      "loss": 0.9453,
+      "step": 92500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 9.710909754721366e-06,
+      "loss": 0.9271,
+      "step": 93000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 9.65559206523062e-06,
+      "loss": 0.9393,
+      "step": 93500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 9.600274375739876e-06,
+      "loss": 0.9376,
+      "step": 94000
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 9.54495668624913e-06,
+      "loss": 0.929,
+      "step": 94500
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.489638996758384e-06,
+      "loss": 0.9438,
+      "step": 95000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.434321307267639e-06,
+      "loss": 0.9369,
+      "step": 95500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 9.379003617776895e-06,
+      "loss": 0.9364,
+      "step": 96000
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 9.323685928286149e-06,
+      "loss": 0.9375,
+      "step": 96500
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 9.268368238795403e-06,
+      "loss": 0.8816,
+      "step": 97000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 9.213050549304657e-06,
+      "loss": 0.916,
+      "step": 97500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.157732859813913e-06,
+      "loss": 0.994,
+      "step": 98000
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.102415170323166e-06,
+      "loss": 0.9202,
+      "step": 98500
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 9.04709748083242e-06,
+      "loss": 0.9044,
+      "step": 99000
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 8.991779791341676e-06,
+      "loss": 0.8644,
+      "step": 99500
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 8.93646210185093e-06,
+      "loss": 0.8926,
+      "step": 100000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.881144412360185e-06,
+      "loss": 0.9647,
+      "step": 100500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.825826722869439e-06,
+      "loss": 0.9277,
+      "step": 101000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.770509033378695e-06,
+      "loss": 0.8776,
+      "step": 101500
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.71519134388795e-06,
+      "loss": 0.8863,
+      "step": 102000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.659873654397204e-06,
+      "loss": 0.9286,
+      "step": 102500
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.604555964906458e-06,
+      "loss": 0.9329,
+      "step": 103000
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.549238275415714e-06,
+      "loss": 0.9091,
+      "step": 103500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.493920585924968e-06,
+      "loss": 0.9215,
+      "step": 104000
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.438602896434222e-06,
+      "loss": 0.9013,
+      "step": 104500
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.383285206943477e-06,
+      "loss": 0.9064,
+      "step": 105000
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.327967517452733e-06,
+      "loss": 0.8798,
+      "step": 105500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 8.272649827961985e-06,
+      "loss": 0.9142,
+      "step": 106000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 8.217332138471241e-06,
+      "loss": 0.9415,
+      "step": 106500
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 8.162014448980496e-06,
+      "loss": 0.8729,
+      "step": 107000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 8.10669675948975e-06,
+      "loss": 0.9351,
+      "step": 107500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 8.051379069999004e-06,
+      "loss": 0.9186,
+      "step": 108000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.99606138050826e-06,
+      "loss": 0.9713,
+      "step": 108500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.940743691017514e-06,
+      "loss": 0.9179,
+      "step": 109000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.885426001526769e-06,
+      "loss": 0.9631,
+      "step": 109500
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.830108312036023e-06,
+      "loss": 0.891,
+      "step": 110000
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.774790622545279e-06,
+      "loss": 0.9189,
+      "step": 110500
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 7.719472933054533e-06,
+      "loss": 0.8631,
+      "step": 111000
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 7.664155243563787e-06,
+      "loss": 0.9599,
+      "step": 111500
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 7.608837554073042e-06,
+      "loss": 0.9375,
+      "step": 112000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 7.553519864582297e-06,
+      "loss": 0.921,
+      "step": 112500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 7.498202175091551e-06,
+      "loss": 0.8958,
+      "step": 113000
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 7.442884485600806e-06,
+      "loss": 0.9505,
+      "step": 113500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 7.3875667961100605e-06,
+      "loss": 0.9405,
+      "step": 114000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 7.332249106619316e-06,
+      "loss": 0.8996,
+      "step": 114500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 7.27693141712857e-06,
+      "loss": 0.8697,
+      "step": 115000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.221613727637825e-06,
+      "loss": 0.9324,
+      "step": 115500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 7.166296038147079e-06,
+      "loss": 0.8865,
+      "step": 116000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 7.1109783486563345e-06,
+      "loss": 0.9143,
+      "step": 116500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 7.055660659165589e-06,
+      "loss": 0.8974,
+      "step": 117000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 7.000342969674844e-06,
+      "loss": 0.926,
+      "step": 117500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 6.945025280184098e-06,
+      "loss": 0.9321,
+      "step": 118000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 6.889707590693352e-06,
+      "loss": 0.9175,
+      "step": 118500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 6.834389901202607e-06,
+      "loss": 0.9066,
+      "step": 119000
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 6.779072211711861e-06,
+      "loss": 0.9149,
+      "step": 119500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 6.723754522221116e-06,
+      "loss": 0.9249,
+      "step": 120000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 6.66843683273037e-06,
+      "loss": 0.8616,
+      "step": 120500
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 6.6131191432396255e-06,
+      "loss": 0.933,
+      "step": 121000
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 6.55780145374888e-06,
+      "loss": 0.9157,
+      "step": 121500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 6.502483764258135e-06,
+      "loss": 0.924,
+      "step": 122000
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 6.447166074767389e-06,
+      "loss": 0.8637,
+      "step": 122500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 6.391848385276644e-06,
+      "loss": 0.9215,
+      "step": 123000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 6.336530695785899e-06,
+      "loss": 0.9098,
+      "step": 123500
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 6.281213006295154e-06,
+      "loss": 0.921,
+      "step": 124000
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 6.225895316804408e-06,
+      "loss": 0.8698,
+      "step": 124500
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 6.170577627313663e-06,
+      "loss": 0.9047,
+      "step": 125000
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 6.1152599378229175e-06,
+      "loss": 0.9108,
+      "step": 125500
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 6.059942248332173e-06,
+      "loss": 0.8666,
+      "step": 126000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 6.004624558841427e-06,
+      "loss": 0.909,
+      "step": 126500
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 5.949306869350682e-06,
+      "loss": 0.9238,
+      "step": 127000
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 5.8939891798599354e-06,
+      "loss": 0.8946,
+      "step": 127500
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 5.838671490369191e-06,
+      "loss": 0.9025,
+      "step": 128000
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 5.783353800878445e-06,
+      "loss": 0.9355,
+      "step": 128500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 5.7280361113877e-06,
+      "loss": 0.8881,
+      "step": 129000
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 5.672718421896954e-06,
+      "loss": 0.9549,
+      "step": 129500
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 5.617400732406209e-06,
+      "loss": 0.9646,
+      "step": 130000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 5.562083042915464e-06,
+      "loss": 0.9239,
+      "step": 130500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 5.506765353424719e-06,
+      "loss": 0.9045,
+      "step": 131000
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 5.451447663933973e-06,
+      "loss": 0.9572,
+      "step": 131500
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 5.396129974443228e-06,
+      "loss": 0.8585,
+      "step": 132000
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 5.3408122849524825e-06,
+      "loss": 0.9026,
+      "step": 132500
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 5.285494595461738e-06,
+      "loss": 0.9043,
+      "step": 133000
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 5.230176905970992e-06,
+      "loss": 0.9234,
+      "step": 133500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 5.174859216480247e-06,
+      "loss": 0.8722,
+      "step": 134000
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 5.119541526989501e-06,
+      "loss": 0.9212,
+      "step": 134500
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 5.064223837498756e-06,
+      "loss": 0.8981,
+      "step": 135000
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 5.008906148008011e-06,
+      "loss": 0.9532,
+      "step": 135500
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 4.953588458517265e-06,
+      "loss": 0.9129,
+      "step": 136000
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.898270769026519e-06,
+      "loss": 0.9075,
+      "step": 136500
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 4.842953079535774e-06,
+      "loss": 0.9206,
+      "step": 137000
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 4.787635390045029e-06,
+      "loss": 0.8512,
+      "step": 137500
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 4.732317700554284e-06,
+      "loss": 0.9086,
+      "step": 138000
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 4.677000011063538e-06,
+      "loss": 0.8828,
+      "step": 138500
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 4.621682321572793e-06,
+      "loss": 0.9071,
+      "step": 139000
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.5663646320820475e-06,
+      "loss": 0.8957,
+      "step": 139500
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 4.511046942591303e-06,
+      "loss": 0.9506,
+      "step": 140000
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 4.455729253100557e-06,
+      "loss": 0.9241,
+      "step": 140500
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 4.400411563609812e-06,
+      "loss": 0.9283,
+      "step": 141000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 4.345093874119066e-06,
+      "loss": 0.9078,
+      "step": 141500
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.2897761846283206e-06,
+      "loss": 0.8944,
+      "step": 142000
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 4.234458495137576e-06,
+      "loss": 0.9283,
+      "step": 142500
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 4.17914080564683e-06,
+      "loss": 0.9263,
+      "step": 143000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 4.123823116156085e-06,
+      "loss": 0.9381,
+      "step": 143500
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 4.068505426665339e-06,
+      "loss": 0.9135,
+      "step": 144000
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.0131877371745945e-06,
+      "loss": 0.8499,
+      "step": 144500
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.957870047683849e-06,
+      "loss": 0.9364,
+      "step": 145000
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.902552358193104e-06,
+      "loss": 0.8836,
+      "step": 145500
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3.847234668702358e-06,
+      "loss": 0.8774,
+      "step": 146000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.791916979211613e-06,
+      "loss": 0.9224,
+      "step": 146500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.7365992897208676e-06,
+      "loss": 0.8753,
+      "step": 147000
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 3.6812816002301223e-06,
+      "loss": 0.9015,
+      "step": 147500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.625963910739376e-06,
+      "loss": 0.9001,
+      "step": 148000
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.570646221248631e-06,
+      "loss": 0.9434,
+      "step": 148500
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.5153285317578856e-06,
+      "loss": 0.9276,
+      "step": 149000
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.4600108422671403e-06,
+      "loss": 0.9056,
+      "step": 149500
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.404693152776395e-06,
+      "loss": 0.884,
+      "step": 150000
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.3493754632856497e-06,
+      "loss": 0.9658,
+      "step": 150500
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.2940577737949044e-06,
+      "loss": 0.9596,
+      "step": 151000
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.2387400843041587e-06,
+      "loss": 0.9951,
+      "step": 151500
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.1834223948134134e-06,
+      "loss": 0.8763,
+      "step": 152000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.128104705322668e-06,
+      "loss": 0.8953,
+      "step": 152500
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.0727870158319228e-06,
+      "loss": 0.9155,
+      "step": 153000
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 3.0174693263411775e-06,
+      "loss": 0.89,
+      "step": 153500
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.962151636850432e-06,
+      "loss": 0.8809,
+      "step": 154000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.906833947359687e-06,
+      "loss": 0.8995,
+      "step": 154500
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.8515162578689416e-06,
+      "loss": 0.9092,
+      "step": 155000
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.7961985683781963e-06,
+      "loss": 0.9475,
+      "step": 155500
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.740880878887451e-06,
+      "loss": 0.8896,
+      "step": 156000
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.6855631893967053e-06,
+      "loss": 0.9145,
+      "step": 156500
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.63024549990596e-06,
+      "loss": 0.9796,
+      "step": 157000
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.5749278104152147e-06,
+      "loss": 0.9451,
+      "step": 157500
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.5196101209244694e-06,
+      "loss": 0.9386,
+      "step": 158000
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 2.464292431433724e-06,
+      "loss": 0.9607,
+      "step": 158500
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 2.408974741942979e-06,
+      "loss": 0.9343,
+      "step": 159000
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 2.3536570524522335e-06,
+      "loss": 0.9211,
+      "step": 159500
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.298339362961488e-06,
+      "loss": 0.9131,
+      "step": 160000
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 2.243021673470743e-06,
+      "loss": 0.9738,
+      "step": 160500
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 2.187703983979997e-06,
+      "loss": 0.9394,
+      "step": 161000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.132386294489252e-06,
+      "loss": 0.9542,
+      "step": 161500
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.0770686049985066e-06,
+      "loss": 0.9253,
+      "step": 162000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.0217509155077613e-06,
+      "loss": 0.8945,
+      "step": 162500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.9664332260170156e-06,
+      "loss": 0.9523,
+      "step": 163000
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.9111155365262703e-06,
+      "loss": 0.8762,
+      "step": 163500
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.855797847035525e-06,
+      "loss": 0.8997,
+      "step": 164000
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.8004801575447797e-06,
+      "loss": 0.9035,
+      "step": 164500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.7451624680540344e-06,
+      "loss": 0.9174,
+      "step": 165000
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.689844778563289e-06,
+      "loss": 0.8572,
+      "step": 165500
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.6345270890725438e-06,
+      "loss": 0.9433,
+      "step": 166000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.5792093995817983e-06,
+      "loss": 0.9124,
+      "step": 166500
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.523891710091053e-06,
+      "loss": 0.8998,
+      "step": 167000
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.4685740206003077e-06,
+      "loss": 0.9118,
+      "step": 167500
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.4132563311095624e-06,
+      "loss": 0.9285,
+      "step": 168000
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.357938641618817e-06,
+      "loss": 0.929,
+      "step": 168500
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.3026209521280716e-06,
+      "loss": 0.9209,
+      "step": 169000
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.2473032626373263e-06,
+      "loss": 0.8879,
+      "step": 169500
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.191985573146581e-06,
+      "loss": 0.9313,
+      "step": 170000
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.1366678836558355e-06,
+      "loss": 0.9262,
+      "step": 170500
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.0813501941650902e-06,
+      "loss": 0.8769,
+      "step": 171000
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.026032504674345e-06,
+      "loss": 0.926,
+      "step": 171500
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 9.707148151835994e-07,
+      "loss": 0.8786,
+      "step": 172000
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 9.153971256928541e-07,
+      "loss": 0.8979,
+      "step": 172500
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 8.600794362021088e-07,
+      "loss": 0.9133,
+      "step": 173000
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 8.047617467113634e-07,
+      "loss": 0.9347,
+      "step": 173500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 7.494440572206181e-07,
+      "loss": 0.9344,
+      "step": 174000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 6.941263677298728e-07,
+      "loss": 0.9326,
+      "step": 174500
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 6.388086782391274e-07,
+      "loss": 0.8966,
+      "step": 175000
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 5.83490988748382e-07,
+      "loss": 0.9497,
+      "step": 175500
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 5.281732992576367e-07,
+      "loss": 0.9086,
+      "step": 176000
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 4.7285560976689126e-07,
+      "loss": 0.9177,
+      "step": 176500
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.175379202761459e-07,
+      "loss": 0.9335,
+      "step": 177000
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.6222023078540056e-07,
+      "loss": 0.8967,
+      "step": 177500
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.0690254129465527e-07,
+      "loss": 0.9551,
+      "step": 178000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.5158485180390987e-07,
+      "loss": 0.9525,
+      "step": 178500
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.9626716231316452e-07,
+      "loss": 0.9046,
+      "step": 179000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.4094947282241917e-07,
+      "loss": 0.9163,
+      "step": 179500
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 8.56317833316738e-08,
+      "loss": 0.935,
+      "step": 180000
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.0314093840928456e-08,
+      "loss": 0.8872,
+      "step": 180500
+    },
+    {
+      "epoch": 3.0,
+      "step": 180774,
+      "total_flos": 2.9686234485024e+17,
+      "train_loss": 0.925638329537373,
+      "train_runtime": 31623.5954,
+      "train_samples_per_second": 5.716,
+      "train_steps_per_second": 5.716
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 180774,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "total_flos": 2.9686234485024e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5ddd77616b4f164cac0cdef0d02b6f4236309ff39a32650cbcc8037f91ac321
+size 4920