Upload folder using huggingface_hub

README.md

@@ -0,0 +1,54 @@
+---
+license: mit
+library_name: transformers
+tags:
+- reward-model
+- reranking
+- literary-style
+- faithfulness
+- longformer
+pipeline_tag: text-classification
+---
+
+# LongformerRM-Unison
+
+Absolute multi-head reward model for literary rewriting.
+
+## Heads
+This model outputs 3 logits in this order:
+
+1. `style`
+2. `faith`
+3. `identifier`
+
+Apply sigmoid to each logit to obtain scores in `[0, 1]`.
+
+## Intended use
+Score a `(source passage, rewrite)` pair for:
+- stylistic quality
+- semantic/content faithfulness
+- identifier preservation (names, dates, numbers, protected spans)
+
+## Important
+This is **not** a comparative Bradley-Terry reward model.  
+It is an **absolute scorer**. Score each candidate independently, then sort externally.
+
+## Input format
+Use exactly:
+
+### Original Draft:
+{prompt}
+
+### Rewritten Version:
+{response}
+
+For Longformer inference, set `global_attention_mask[:, 0] = 1`.
+
+## Recommended composite score
+`overall_score = style_score * (0.5 * faith_score + 0.5 * identifier_score) * (identifier_score ** 1.5)`
+
+## Output head order
+`[style_logit, faith_logit, identifier_logit]`
+
+## Base model
+`./absolute-multitask-reward-model`

config.json

@@ -0,0 +1,55 @@
+{
+  "_name_or_path": "allenai/longformer-base-4096",
+  "architectures": [
+    "LongformerForSequenceClassification"
+  ],
+  "attention_mode": "longformer",
+  "attention_probs_dropout_prob": 0.1,
+  "attention_window": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "bos_token_id": 0,
+  "eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2"
+  },
+  "ignore_attention_mask": false,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_2": 2
+  },
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 4098,
+  "model_type": "longformer",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "onnx_export": false,
+  "pad_token_id": 1,
+  "problem_type": "multi_label_classification",
+  "sep_token_id": 2,
+  "torch_dtype": "float32",
+  "transformers_version": "4.48.3",
+  "type_vocab_size": 1,
+  "use_cache": false,
+  "vocab_size": 50265
+}

inference_config.json

@@ -0,0 +1,18 @@
+{
+  "model_type": "absolute_multihead_reward_model",
+  "base_model": "./absolute-multitask-reward-model",
+  "num_labels": 3,
+  "label_order": [
+    "style",
+    "faith",
+    "identifier"
+  ],
+  "max_length": 1024,
+  "recommended_overall_formula": "style_score * (0.5 * faith_score + 0.5 * identifier_score) * (identifier_score ** 1.5)",
+  "notes": [
+    "Apply sigmoid to each logit before using scores.",
+    "Do not use style_score alone for reranking.",
+    "Keep input formatting consistent with training.",
+    "For Longformer inference, set global_attention_mask[:, 0] = 1."
+  ]
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5e1a73b2288d51d36b6b682b462a60f81e98566de95e9b1e9a477f1c5316f46
+size 594681260

special_tokens_map.json

@@ -0,0 +1,51 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "cls_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,65 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50264": {
+      "content": "<mask>",
+      "lstrip": true,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "mask_token": "<mask>",
+  "max_length": 1024,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_to_multiple_of": null,
+  "pad_token": "<pad>",
+  "pad_token_type_id": 0,
+  "padding_side": "right",
+  "sep_token": "</s>",
+  "stride": 0,
+  "tokenizer_class": "LongformerTokenizer",
+  "trim_offsets": true,
+  "truncation_side": "right",
+  "truncation_strategy": "longest_first",
+  "unk_token": "<unk>"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:832cbd1d8fc4636ee25d0e16ea97e5fc6c54c12068318296ad9d45bd89998043
+size 5777