Upload checkpoint-700
Browse files- config.json +33 -0
- generation_config.json +10 -0
- model.safetensors +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +10 -0
- trainer_state.json +524 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Phi3ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_dropout": 0.0,
|
| 6 |
+
"bos_token_id": 0,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"embd_pdrop": 0.0,
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 1024,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 2304,
|
| 14 |
+
"max_position_embeddings": 1024,
|
| 15 |
+
"model_type": "phi3",
|
| 16 |
+
"num_attention_heads": 8,
|
| 17 |
+
"num_hidden_layers": 18,
|
| 18 |
+
"num_key_value_heads": 2,
|
| 19 |
+
"original_max_position_embeddings": 4096,
|
| 20 |
+
"pad_token_id": 4,
|
| 21 |
+
"resid_pdrop": 0.0,
|
| 22 |
+
"rms_norm_eps": 1e-05,
|
| 23 |
+
"rope_parameters": {
|
| 24 |
+
"partial_rotary_factor": 1.0,
|
| 25 |
+
"rope_theta": 10000.0,
|
| 26 |
+
"rope_type": "default"
|
| 27 |
+
},
|
| 28 |
+
"sliding_window": null,
|
| 29 |
+
"tie_word_embeddings": false,
|
| 30 |
+
"transformers_version": "5.5.0",
|
| 31 |
+
"use_cache": false,
|
| 32 |
+
"vocab_size": 50000
|
| 33 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 0,
|
| 5 |
+
"output_attentions": false,
|
| 6 |
+
"output_hidden_states": false,
|
| 7 |
+
"pad_token_id": 4,
|
| 8 |
+
"transformers_version": "5.5.0",
|
| 9 |
+
"use_cache": true
|
| 10 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e023dce32d3fdc3aee65fc0f7325ef5ddb15c490475da7cac7b5ead6b19c5901
|
| 3 |
+
size 1108115808
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"backend": "tokenizers",
|
| 3 |
+
"bos_token": "<|endoftext|>",
|
| 4 |
+
"eos_token": "<|endoftext|>",
|
| 5 |
+
"is_local": true,
|
| 6 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 7 |
+
"pad_token": "<|pad|>",
|
| 8 |
+
"tokenizer_class": "TokenizersBackend",
|
| 9 |
+
"unk_token": "<|endoftext|>"
|
| 10 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,524 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.058333333333333334,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 700,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0008333333333333334,
|
| 14 |
+
"grad_norm": 0.3933861553668976,
|
| 15 |
+
"learning_rate": 1.3499999999999998e-05,
|
| 16 |
+
"loss": 10.793175506591798,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.0016666666666666668,
|
| 21 |
+
"grad_norm": 0.21078218519687653,
|
| 22 |
+
"learning_rate": 2.8499999999999998e-05,
|
| 23 |
+
"loss": 9.94891357421875,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.0025,
|
| 28 |
+
"grad_norm": 0.17593811452388763,
|
| 29 |
+
"learning_rate": 4.3499999999999993e-05,
|
| 30 |
+
"loss": 9.451502990722656,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.0033333333333333335,
|
| 35 |
+
"grad_norm": 0.1287856101989746,
|
| 36 |
+
"learning_rate": 5.85e-05,
|
| 37 |
+
"loss": 9.038980102539062,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.004166666666666667,
|
| 42 |
+
"grad_norm": 0.12334560602903366,
|
| 43 |
+
"learning_rate": 7.35e-05,
|
| 44 |
+
"loss": 8.548064422607421,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.005,
|
| 49 |
+
"grad_norm": 0.08555905520915985,
|
| 50 |
+
"learning_rate": 8.849999999999998e-05,
|
| 51 |
+
"loss": 8.130076599121093,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.005833333333333334,
|
| 56 |
+
"grad_norm": 0.09234856814146042,
|
| 57 |
+
"learning_rate": 0.00010349999999999998,
|
| 58 |
+
"loss": 7.781331634521484,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.006666666666666667,
|
| 63 |
+
"grad_norm": 0.12036211788654327,
|
| 64 |
+
"learning_rate": 0.0001185,
|
| 65 |
+
"loss": 7.612437438964844,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.0075,
|
| 70 |
+
"grad_norm": 0.26068001985549927,
|
| 71 |
+
"learning_rate": 0.0001335,
|
| 72 |
+
"loss": 7.377515411376953,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.008333333333333333,
|
| 77 |
+
"grad_norm": 0.14013038575649261,
|
| 78 |
+
"learning_rate": 0.00014849999999999998,
|
| 79 |
+
"loss": 7.176266479492187,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.009166666666666667,
|
| 84 |
+
"grad_norm": 0.1725226491689682,
|
| 85 |
+
"learning_rate": 0.0001635,
|
| 86 |
+
"loss": 6.9345703125,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.01,
|
| 91 |
+
"grad_norm": 0.1893932819366455,
|
| 92 |
+
"learning_rate": 0.00017849999999999997,
|
| 93 |
+
"loss": 6.603112030029297,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.010833333333333334,
|
| 98 |
+
"grad_norm": 0.10222424566745758,
|
| 99 |
+
"learning_rate": 0.0001935,
|
| 100 |
+
"loss": 6.440335083007812,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.011666666666666667,
|
| 105 |
+
"grad_norm": 0.12809771299362183,
|
| 106 |
+
"learning_rate": 0.00020849999999999997,
|
| 107 |
+
"loss": 6.1724189758300785,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.0125,
|
| 112 |
+
"grad_norm": 0.1817079782485962,
|
| 113 |
+
"learning_rate": 0.00022349999999999998,
|
| 114 |
+
"loss": 5.981137084960937,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.013333333333333334,
|
| 119 |
+
"grad_norm": 0.1294909566640854,
|
| 120 |
+
"learning_rate": 0.0002385,
|
| 121 |
+
"loss": 5.84722900390625,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.014166666666666666,
|
| 126 |
+
"grad_norm": 0.1508026421070099,
|
| 127 |
+
"learning_rate": 0.0002535,
|
| 128 |
+
"loss": 5.6487274169921875,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.015,
|
| 133 |
+
"grad_norm": 0.1253245323896408,
|
| 134 |
+
"learning_rate": 0.00026849999999999997,
|
| 135 |
+
"loss": 5.603432464599609,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.015833333333333335,
|
| 140 |
+
"grad_norm": 0.10764947533607483,
|
| 141 |
+
"learning_rate": 0.00028349999999999995,
|
| 142 |
+
"loss": 5.431602096557617,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.016666666666666666,
|
| 147 |
+
"grad_norm": 0.11333471536636353,
|
| 148 |
+
"learning_rate": 0.0002985,
|
| 149 |
+
"loss": 5.339364624023437,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.0175,
|
| 154 |
+
"grad_norm": 0.12494370341300964,
|
| 155 |
+
"learning_rate": 0.0002999995693922863,
|
| 156 |
+
"loss": 5.225184631347656,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.018333333333333333,
|
| 161 |
+
"grad_norm": 0.11432802677154541,
|
| 162 |
+
"learning_rate": 0.0002999980808749688,
|
| 163 |
+
"loss": 5.181040191650391,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.019166666666666665,
|
| 168 |
+
"grad_norm": 0.10197849571704865,
|
| 169 |
+
"learning_rate": 0.000299995529142451,
|
| 170 |
+
"loss": 5.013928985595703,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.02,
|
| 175 |
+
"grad_norm": 0.11272598057985306,
|
| 176 |
+
"learning_rate": 0.00029999191421282024,
|
| 177 |
+
"loss": 4.891825103759766,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.020833333333333332,
|
| 182 |
+
"grad_norm": 0.09002882242202759,
|
| 183 |
+
"learning_rate": 0.0002999872361116999,
|
| 184 |
+
"loss": 4.906301879882813,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.021666666666666667,
|
| 189 |
+
"grad_norm": 0.0879431664943695,
|
| 190 |
+
"learning_rate": 0.0002999814948722491,
|
| 191 |
+
"loss": 4.827959442138672,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.0225,
|
| 196 |
+
"grad_norm": 0.10233565419912338,
|
| 197 |
+
"learning_rate": 0.0002999746905351629,
|
| 198 |
+
"loss": 4.808346176147461,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.023333333333333334,
|
| 203 |
+
"grad_norm": 0.08734232932329178,
|
| 204 |
+
"learning_rate": 0.00029996682314867176,
|
| 205 |
+
"loss": 4.741819381713867,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.024166666666666666,
|
| 210 |
+
"grad_norm": 0.0991712436079979,
|
| 211 |
+
"learning_rate": 0.0002999578927685413,
|
| 212 |
+
"loss": 4.678402709960937,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.025,
|
| 217 |
+
"grad_norm": 0.0887921079993248,
|
| 218 |
+
"learning_rate": 0.0002999478994580716,
|
| 219 |
+
"loss": 4.653173828125,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.025833333333333333,
|
| 224 |
+
"grad_norm": 0.0922158882021904,
|
| 225 |
+
"learning_rate": 0.0002999368432880973,
|
| 226 |
+
"loss": 4.650460815429687,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.02666666666666667,
|
| 231 |
+
"grad_norm": 0.09760992228984833,
|
| 232 |
+
"learning_rate": 0.0002999247243369867,
|
| 233 |
+
"loss": 4.53588981628418,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.0275,
|
| 238 |
+
"grad_norm": 0.08007848262786865,
|
| 239 |
+
"learning_rate": 0.00029991154269064105,
|
| 240 |
+
"loss": 4.48128662109375,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.028333333333333332,
|
| 245 |
+
"grad_norm": 0.09366138279438019,
|
| 246 |
+
"learning_rate": 0.00029989729844249455,
|
| 247 |
+
"loss": 4.436825561523437,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.029166666666666667,
|
| 252 |
+
"grad_norm": 0.08227650076150894,
|
| 253 |
+
"learning_rate": 0.00029988199169351317,
|
| 254 |
+
"loss": 4.347656631469727,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.03,
|
| 259 |
+
"grad_norm": 0.0964592844247818,
|
| 260 |
+
"learning_rate": 0.000299865622552194,
|
| 261 |
+
"loss": 4.348267364501953,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.030833333333333334,
|
| 266 |
+
"grad_norm": 0.08087096363306046,
|
| 267 |
+
"learning_rate": 0.0002998481911345647,
|
| 268 |
+
"loss": 4.408647918701172,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.03166666666666667,
|
| 273 |
+
"grad_norm": 0.09135255217552185,
|
| 274 |
+
"learning_rate": 0.0002998296975641825,
|
| 275 |
+
"loss": 4.3575798034667965,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.0325,
|
| 280 |
+
"grad_norm": 0.07801107317209244,
|
| 281 |
+
"learning_rate": 0.00029981014197213345,
|
| 282 |
+
"loss": 4.369893646240234,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.03333333333333333,
|
| 287 |
+
"grad_norm": 0.074135921895504,
|
| 288 |
+
"learning_rate": 0.0002997895244970312,
|
| 289 |
+
"loss": 4.275115585327148,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.034166666666666665,
|
| 294 |
+
"grad_norm": 0.08232986181974411,
|
| 295 |
+
"learning_rate": 0.00029976784528501645,
|
| 296 |
+
"loss": 4.228077697753906,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.035,
|
| 301 |
+
"grad_norm": 0.07607673108577728,
|
| 302 |
+
"learning_rate": 0.00029974510448975563,
|
| 303 |
+
"loss": 4.215129089355469,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.035833333333333335,
|
| 308 |
+
"grad_norm": 0.07808370888233185,
|
| 309 |
+
"learning_rate": 0.0002997213022724399,
|
| 310 |
+
"loss": 4.161651229858398,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.03666666666666667,
|
| 315 |
+
"grad_norm": 0.07890745997428894,
|
| 316 |
+
"learning_rate": 0.0002996964388017838,
|
| 317 |
+
"loss": 4.197282409667968,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.0375,
|
| 322 |
+
"grad_norm": 0.071009561419487,
|
| 323 |
+
"learning_rate": 0.00029967051425402455,
|
| 324 |
+
"loss": 4.102042007446289,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.03833333333333333,
|
| 329 |
+
"grad_norm": 0.07312625646591187,
|
| 330 |
+
"learning_rate": 0.0002996435288129202,
|
| 331 |
+
"loss": 4.090036773681641,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.03916666666666667,
|
| 336 |
+
"grad_norm": 0.07782439142465591,
|
| 337 |
+
"learning_rate": 0.0002996154826697489,
|
| 338 |
+
"loss": 4.060888671875,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.04,
|
| 343 |
+
"grad_norm": 0.06876000016927719,
|
| 344 |
+
"learning_rate": 0.00029958637602330685,
|
| 345 |
+
"loss": 4.065421295166016,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.04083333333333333,
|
| 350 |
+
"grad_norm": 0.07044120877981186,
|
| 351 |
+
"learning_rate": 0.00029955620907990777,
|
| 352 |
+
"loss": 4.0340923309326175,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.041666666666666664,
|
| 357 |
+
"grad_norm": 0.06847608089447021,
|
| 358 |
+
"learning_rate": 0.0002995249820533806,
|
| 359 |
+
"loss": 4.013957214355469,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.0425,
|
| 364 |
+
"grad_norm": 0.06946133822202682,
|
| 365 |
+
"learning_rate": 0.00029949269516506863,
|
| 366 |
+
"loss": 4.006973648071289,
|
| 367 |
+
"step": 510
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.043333333333333335,
|
| 371 |
+
"grad_norm": 0.07448018342256546,
|
| 372 |
+
"learning_rate": 0.0002994593486438274,
|
| 373 |
+
"loss": 3.9334747314453127,
|
| 374 |
+
"step": 520
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.04416666666666667,
|
| 378 |
+
"grad_norm": 0.0726313516497612,
|
| 379 |
+
"learning_rate": 0.00029942494272602356,
|
| 380 |
+
"loss": 3.9182762145996093,
|
| 381 |
+
"step": 530
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.045,
|
| 385 |
+
"grad_norm": 0.07197625935077667,
|
| 386 |
+
"learning_rate": 0.0002993894776555328,
|
| 387 |
+
"loss": 3.9256614685058593,
|
| 388 |
+
"step": 540
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.04583333333333333,
|
| 392 |
+
"grad_norm": 0.07247449457645416,
|
| 393 |
+
"learning_rate": 0.00029935295368373844,
|
| 394 |
+
"loss": 3.8697052001953125,
|
| 395 |
+
"step": 550
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.04666666666666667,
|
| 399 |
+
"grad_norm": 0.06503958255052567,
|
| 400 |
+
"learning_rate": 0.00029931537106952925,
|
| 401 |
+
"loss": 3.833271026611328,
|
| 402 |
+
"step": 560
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.0475,
|
| 406 |
+
"grad_norm": 0.07078709453344345,
|
| 407 |
+
"learning_rate": 0.00029927673007929814,
|
| 408 |
+
"loss": 3.807381820678711,
|
| 409 |
+
"step": 570
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.04833333333333333,
|
| 413 |
+
"grad_norm": 0.0642218068242073,
|
| 414 |
+
"learning_rate": 0.0002992370309869398,
|
| 415 |
+
"loss": 3.7920074462890625,
|
| 416 |
+
"step": 580
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.049166666666666664,
|
| 420 |
+
"grad_norm": 0.07087492197751999,
|
| 421 |
+
"learning_rate": 0.00029919627407384906,
|
| 422 |
+
"loss": 3.759918975830078,
|
| 423 |
+
"step": 590
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.05,
|
| 427 |
+
"grad_norm": 0.07459498941898346,
|
| 428 |
+
"learning_rate": 0.0002991544596289187,
|
| 429 |
+
"loss": 3.7198650360107424,
|
| 430 |
+
"step": 600
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.050833333333333335,
|
| 434 |
+
"grad_norm": 0.07105842977762222,
|
| 435 |
+
"learning_rate": 0.0002991115879485376,
|
| 436 |
+
"loss": 3.7618087768554687,
|
| 437 |
+
"step": 610
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.051666666666666666,
|
| 441 |
+
"grad_norm": 0.06777296215295792,
|
| 442 |
+
"learning_rate": 0.00029906765933658845,
|
| 443 |
+
"loss": 3.774607467651367,
|
| 444 |
+
"step": 620
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.0525,
|
| 448 |
+
"grad_norm": 0.07675730437040329,
|
| 449 |
+
"learning_rate": 0.00029902267410444563,
|
| 450 |
+
"loss": 3.7071380615234375,
|
| 451 |
+
"step": 630
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.05333333333333334,
|
| 455 |
+
"grad_norm": 0.07201831042766571,
|
| 456 |
+
"learning_rate": 0.00029897663257097315,
|
| 457 |
+
"loss": 3.689577102661133,
|
| 458 |
+
"step": 640
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.05416666666666667,
|
| 462 |
+
"grad_norm": 0.07140760123729706,
|
| 463 |
+
"learning_rate": 0.00029892953506252225,
|
| 464 |
+
"loss": 3.6408855438232424,
|
| 465 |
+
"step": 650
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.055,
|
| 469 |
+
"grad_norm": 0.06695026159286499,
|
| 470 |
+
"learning_rate": 0.0002988813819129291,
|
| 471 |
+
"loss": 3.6567893981933595,
|
| 472 |
+
"step": 660
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.05583333333333333,
|
| 476 |
+
"grad_norm": 0.06928922981023788,
|
| 477 |
+
"learning_rate": 0.00029883217346351245,
|
| 478 |
+
"loss": 3.611972427368164,
|
| 479 |
+
"step": 670
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.056666666666666664,
|
| 483 |
+
"grad_norm": 0.07292694598436356,
|
| 484 |
+
"learning_rate": 0.00029878191006307126,
|
| 485 |
+
"loss": 3.6120975494384764,
|
| 486 |
+
"step": 680
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.0575,
|
| 490 |
+
"grad_norm": 0.06881766021251678,
|
| 491 |
+
"learning_rate": 0.0002987305920678821,
|
| 492 |
+
"loss": 3.561126708984375,
|
| 493 |
+
"step": 690
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.058333333333333334,
|
| 497 |
+
"grad_norm": 0.06780871003866196,
|
| 498 |
+
"learning_rate": 0.00029867821984169696,
|
| 499 |
+
"loss": 3.608598327636719,
|
| 500 |
+
"step": 700
|
| 501 |
+
}
|
| 502 |
+
],
|
| 503 |
+
"logging_steps": 10,
|
| 504 |
+
"max_steps": 12000,
|
| 505 |
+
"num_input_tokens_seen": 0,
|
| 506 |
+
"num_train_epochs": 9223372036854775807,
|
| 507 |
+
"save_steps": 100,
|
| 508 |
+
"stateful_callbacks": {
|
| 509 |
+
"TrainerControl": {
|
| 510 |
+
"args": {
|
| 511 |
+
"should_epoch_stop": false,
|
| 512 |
+
"should_evaluate": false,
|
| 513 |
+
"should_log": false,
|
| 514 |
+
"should_save": true,
|
| 515 |
+
"should_training_stop": false
|
| 516 |
+
},
|
| 517 |
+
"attributes": {}
|
| 518 |
+
}
|
| 519 |
+
},
|
| 520 |
+
"total_flos": 6.21588202389504e+16,
|
| 521 |
+
"train_batch_size": 8,
|
| 522 |
+
"trial_name": null,
|
| 523 |
+
"trial_params": null
|
| 524 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2f0e8a09caa135cb813a82756543b3a3b94e604f8a69c1c4988db8b9b100237
|
| 3 |
+
size 5201
|