End of training
Browse files- last-checkpoint/config.json +0 -75
- last-checkpoint/optimizer.pt +0 -3
- last-checkpoint/pytorch_model.bin +0 -3
- last-checkpoint/rng_state.pth +0 -3
- last-checkpoint/scheduler.pt +0 -3
- last-checkpoint/trainer_state.json +0 -916
- last-checkpoint/training_args.bin +0 -3
- pytorch_model.bin +1 -1
- runs/Jan01_04-28-26_6ea5fb75c867/events.out.tfevents.1672547319.6ea5fb75c867.24.0 +2 -2
last-checkpoint/config.json
DELETED
|
@@ -1,75 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"_name_or_path": "facebook/bart-base",
|
| 3 |
-
"activation_dropout": 0.1,
|
| 4 |
-
"activation_function": "gelu",
|
| 5 |
-
"add_bias_logits": false,
|
| 6 |
-
"add_final_layer_norm": false,
|
| 7 |
-
"architectures": [
|
| 8 |
-
"BartForConditionalGeneration"
|
| 9 |
-
],
|
| 10 |
-
"attention_dropout": 0.1,
|
| 11 |
-
"bos_token_id": 0,
|
| 12 |
-
"classif_dropout": 0.1,
|
| 13 |
-
"classifier_dropout": 0.0,
|
| 14 |
-
"d_model": 768,
|
| 15 |
-
"decoder_attention_heads": 12,
|
| 16 |
-
"decoder_ffn_dim": 3072,
|
| 17 |
-
"decoder_layerdrop": 0.0,
|
| 18 |
-
"decoder_layers": 6,
|
| 19 |
-
"decoder_start_token_id": 2,
|
| 20 |
-
"dropout": 0.1,
|
| 21 |
-
"early_stopping": true,
|
| 22 |
-
"encoder_attention_heads": 12,
|
| 23 |
-
"encoder_ffn_dim": 3072,
|
| 24 |
-
"encoder_layerdrop": 0.0,
|
| 25 |
-
"encoder_layers": 6,
|
| 26 |
-
"eos_token_id": 2,
|
| 27 |
-
"forced_bos_token_id": 0,
|
| 28 |
-
"forced_eos_token_id": 2,
|
| 29 |
-
"gradient_checkpointing": false,
|
| 30 |
-
"id2label": {
|
| 31 |
-
"0": "LABEL_0",
|
| 32 |
-
"1": "LABEL_1",
|
| 33 |
-
"2": "LABEL_2"
|
| 34 |
-
},
|
| 35 |
-
"init_std": 0.02,
|
| 36 |
-
"is_encoder_decoder": true,
|
| 37 |
-
"label2id": {
|
| 38 |
-
"LABEL_0": 0,
|
| 39 |
-
"LABEL_1": 1,
|
| 40 |
-
"LABEL_2": 2
|
| 41 |
-
},
|
| 42 |
-
"max_position_embeddings": 1024,
|
| 43 |
-
"model_type": "bart",
|
| 44 |
-
"no_repeat_ngram_size": 3,
|
| 45 |
-
"normalize_before": false,
|
| 46 |
-
"normalize_embedding": true,
|
| 47 |
-
"num_beams": 4,
|
| 48 |
-
"num_hidden_layers": 6,
|
| 49 |
-
"pad_token_id": 1,
|
| 50 |
-
"scale_embedding": false,
|
| 51 |
-
"task_specific_params": {
|
| 52 |
-
"summarization": {
|
| 53 |
-
"length_penalty": 1.0,
|
| 54 |
-
"max_length": 128,
|
| 55 |
-
"min_length": 12,
|
| 56 |
-
"num_beams": 4
|
| 57 |
-
},
|
| 58 |
-
"summarization_cnn": {
|
| 59 |
-
"length_penalty": 2.0,
|
| 60 |
-
"max_length": 142,
|
| 61 |
-
"min_length": 56,
|
| 62 |
-
"num_beams": 4
|
| 63 |
-
},
|
| 64 |
-
"summarization_xsum": {
|
| 65 |
-
"length_penalty": 1.0,
|
| 66 |
-
"max_length": 62,
|
| 67 |
-
"min_length": 11,
|
| 68 |
-
"num_beams": 6
|
| 69 |
-
}
|
| 70 |
-
},
|
| 71 |
-
"torch_dtype": "float32",
|
| 72 |
-
"transformers_version": "4.20.1",
|
| 73 |
-
"use_cache": true,
|
| 74 |
-
"vocab_size": 50265
|
| 75 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/optimizer.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:fadf2874e53143ee977132af8fb5bd7d426229c919e522908fe884ecb84f4345
|
| 3 |
-
size 1115513717
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/pytorch_model.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:8c09b8c06d2814b8003bcbda40c13a2533f7e29d29db2312c80a18a4f8d1cdf9
|
| 3 |
-
size 557969145
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/rng_state.pth
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:3006cbdb81bbb28698a27ec8607059ef2927a18060ae014f8abac7d512aa09fb
|
| 3 |
-
size 15523
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/scheduler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:f33807aed54b7d234a4b5d087ac80aee63260ad15758612d4571a720e74319ce
|
| 3 |
-
size 623
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/trainer_state.json
DELETED
|
@@ -1,916 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_metric": null,
|
| 3 |
-
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 1.723543605653223,
|
| 5 |
-
"global_step": 15000,
|
| 6 |
-
"is_hyper_param_search": false,
|
| 7 |
-
"is_local_process_zero": true,
|
| 8 |
-
"is_world_process_zero": true,
|
| 9 |
-
"log_history": [
|
| 10 |
-
{
|
| 11 |
-
"epoch": 0.01,
|
| 12 |
-
"learning_rate": 4.968083748244606e-05,
|
| 13 |
-
"loss": 2.9096,
|
| 14 |
-
"step": 100
|
| 15 |
-
},
|
| 16 |
-
{
|
| 17 |
-
"epoch": 0.02,
|
| 18 |
-
"learning_rate": 4.936167496489212e-05,
|
| 19 |
-
"loss": 2.8631,
|
| 20 |
-
"step": 200
|
| 21 |
-
},
|
| 22 |
-
{
|
| 23 |
-
"epoch": 0.03,
|
| 24 |
-
"learning_rate": 4.904251244733819e-05,
|
| 25 |
-
"loss": 2.7364,
|
| 26 |
-
"step": 300
|
| 27 |
-
},
|
| 28 |
-
{
|
| 29 |
-
"epoch": 0.05,
|
| 30 |
-
"learning_rate": 4.8723349929784253e-05,
|
| 31 |
-
"loss": 2.8226,
|
| 32 |
-
"step": 400
|
| 33 |
-
},
|
| 34 |
-
{
|
| 35 |
-
"epoch": 0.06,
|
| 36 |
-
"learning_rate": 4.840418741223031e-05,
|
| 37 |
-
"loss": 2.7184,
|
| 38 |
-
"step": 500
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.07,
|
| 42 |
-
"learning_rate": 4.808502489467637e-05,
|
| 43 |
-
"loss": 2.6843,
|
| 44 |
-
"step": 600
|
| 45 |
-
},
|
| 46 |
-
{
|
| 47 |
-
"epoch": 0.08,
|
| 48 |
-
"learning_rate": 4.776586237712243e-05,
|
| 49 |
-
"loss": 2.7507,
|
| 50 |
-
"step": 700
|
| 51 |
-
},
|
| 52 |
-
{
|
| 53 |
-
"epoch": 0.09,
|
| 54 |
-
"learning_rate": 4.744669985956849e-05,
|
| 55 |
-
"loss": 2.7261,
|
| 56 |
-
"step": 800
|
| 57 |
-
},
|
| 58 |
-
{
|
| 59 |
-
"epoch": 0.1,
|
| 60 |
-
"learning_rate": 4.712753734201456e-05,
|
| 61 |
-
"loss": 2.7343,
|
| 62 |
-
"step": 900
|
| 63 |
-
},
|
| 64 |
-
{
|
| 65 |
-
"epoch": 0.11,
|
| 66 |
-
"learning_rate": 4.6808374824460616e-05,
|
| 67 |
-
"loss": 2.7005,
|
| 68 |
-
"step": 1000
|
| 69 |
-
},
|
| 70 |
-
{
|
| 71 |
-
"epoch": 0.13,
|
| 72 |
-
"learning_rate": 4.6489212306906676e-05,
|
| 73 |
-
"loss": 2.6462,
|
| 74 |
-
"step": 1100
|
| 75 |
-
},
|
| 76 |
-
{
|
| 77 |
-
"epoch": 0.14,
|
| 78 |
-
"learning_rate": 4.617004978935274e-05,
|
| 79 |
-
"loss": 2.6276,
|
| 80 |
-
"step": 1200
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.15,
|
| 84 |
-
"learning_rate": 4.58508872717988e-05,
|
| 85 |
-
"loss": 2.6493,
|
| 86 |
-
"step": 1300
|
| 87 |
-
},
|
| 88 |
-
{
|
| 89 |
-
"epoch": 0.16,
|
| 90 |
-
"learning_rate": 4.553172475424487e-05,
|
| 91 |
-
"loss": 2.7024,
|
| 92 |
-
"step": 1400
|
| 93 |
-
},
|
| 94 |
-
{
|
| 95 |
-
"epoch": 0.17,
|
| 96 |
-
"learning_rate": 4.521256223669093e-05,
|
| 97 |
-
"loss": 2.6616,
|
| 98 |
-
"step": 1500
|
| 99 |
-
},
|
| 100 |
-
{
|
| 101 |
-
"epoch": 0.18,
|
| 102 |
-
"learning_rate": 4.4893399719136986e-05,
|
| 103 |
-
"loss": 2.7136,
|
| 104 |
-
"step": 1600
|
| 105 |
-
},
|
| 106 |
-
{
|
| 107 |
-
"epoch": 0.2,
|
| 108 |
-
"learning_rate": 4.4574237201583046e-05,
|
| 109 |
-
"loss": 2.6842,
|
| 110 |
-
"step": 1700
|
| 111 |
-
},
|
| 112 |
-
{
|
| 113 |
-
"epoch": 0.21,
|
| 114 |
-
"learning_rate": 4.4255074684029105e-05,
|
| 115 |
-
"loss": 2.5911,
|
| 116 |
-
"step": 1800
|
| 117 |
-
},
|
| 118 |
-
{
|
| 119 |
-
"epoch": 0.22,
|
| 120 |
-
"learning_rate": 4.393591216647517e-05,
|
| 121 |
-
"loss": 2.7281,
|
| 122 |
-
"step": 1900
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.23,
|
| 126 |
-
"learning_rate": 4.361674964892124e-05,
|
| 127 |
-
"loss": 2.6697,
|
| 128 |
-
"step": 2000
|
| 129 |
-
},
|
| 130 |
-
{
|
| 131 |
-
"epoch": 0.24,
|
| 132 |
-
"learning_rate": 4.32975871313673e-05,
|
| 133 |
-
"loss": 2.6282,
|
| 134 |
-
"step": 2100
|
| 135 |
-
},
|
| 136 |
-
{
|
| 137 |
-
"epoch": 0.25,
|
| 138 |
-
"learning_rate": 4.2978424613813356e-05,
|
| 139 |
-
"loss": 2.564,
|
| 140 |
-
"step": 2200
|
| 141 |
-
},
|
| 142 |
-
{
|
| 143 |
-
"epoch": 0.26,
|
| 144 |
-
"learning_rate": 4.2659262096259415e-05,
|
| 145 |
-
"loss": 2.4882,
|
| 146 |
-
"step": 2300
|
| 147 |
-
},
|
| 148 |
-
{
|
| 149 |
-
"epoch": 0.28,
|
| 150 |
-
"learning_rate": 4.2340099578705475e-05,
|
| 151 |
-
"loss": 2.6347,
|
| 152 |
-
"step": 2400
|
| 153 |
-
},
|
| 154 |
-
{
|
| 155 |
-
"epoch": 0.29,
|
| 156 |
-
"learning_rate": 4.202093706115154e-05,
|
| 157 |
-
"loss": 2.5994,
|
| 158 |
-
"step": 2500
|
| 159 |
-
},
|
| 160 |
-
{
|
| 161 |
-
"epoch": 0.3,
|
| 162 |
-
"learning_rate": 4.17017745435976e-05,
|
| 163 |
-
"loss": 2.6479,
|
| 164 |
-
"step": 2600
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 0.31,
|
| 168 |
-
"learning_rate": 4.1382612026043666e-05,
|
| 169 |
-
"loss": 2.7357,
|
| 170 |
-
"step": 2700
|
| 171 |
-
},
|
| 172 |
-
{
|
| 173 |
-
"epoch": 0.32,
|
| 174 |
-
"learning_rate": 4.1063449508489726e-05,
|
| 175 |
-
"loss": 2.6819,
|
| 176 |
-
"step": 2800
|
| 177 |
-
},
|
| 178 |
-
{
|
| 179 |
-
"epoch": 0.33,
|
| 180 |
-
"learning_rate": 4.0744286990935785e-05,
|
| 181 |
-
"loss": 2.5451,
|
| 182 |
-
"step": 2900
|
| 183 |
-
},
|
| 184 |
-
{
|
| 185 |
-
"epoch": 0.34,
|
| 186 |
-
"learning_rate": 4.0425124473381845e-05,
|
| 187 |
-
"loss": 2.6343,
|
| 188 |
-
"step": 3000
|
| 189 |
-
},
|
| 190 |
-
{
|
| 191 |
-
"epoch": 0.36,
|
| 192 |
-
"learning_rate": 4.010596195582791e-05,
|
| 193 |
-
"loss": 2.5168,
|
| 194 |
-
"step": 3100
|
| 195 |
-
},
|
| 196 |
-
{
|
| 197 |
-
"epoch": 0.37,
|
| 198 |
-
"learning_rate": 3.978679943827397e-05,
|
| 199 |
-
"loss": 2.5881,
|
| 200 |
-
"step": 3200
|
| 201 |
-
},
|
| 202 |
-
{
|
| 203 |
-
"epoch": 0.38,
|
| 204 |
-
"learning_rate": 3.946763692072003e-05,
|
| 205 |
-
"loss": 2.6411,
|
| 206 |
-
"step": 3300
|
| 207 |
-
},
|
| 208 |
-
{
|
| 209 |
-
"epoch": 0.39,
|
| 210 |
-
"learning_rate": 3.9148474403166096e-05,
|
| 211 |
-
"loss": 2.5901,
|
| 212 |
-
"step": 3400
|
| 213 |
-
},
|
| 214 |
-
{
|
| 215 |
-
"epoch": 0.4,
|
| 216 |
-
"learning_rate": 3.8829311885612155e-05,
|
| 217 |
-
"loss": 2.5842,
|
| 218 |
-
"step": 3500
|
| 219 |
-
},
|
| 220 |
-
{
|
| 221 |
-
"epoch": 0.41,
|
| 222 |
-
"learning_rate": 3.851014936805822e-05,
|
| 223 |
-
"loss": 2.5436,
|
| 224 |
-
"step": 3600
|
| 225 |
-
},
|
| 226 |
-
{
|
| 227 |
-
"epoch": 0.43,
|
| 228 |
-
"learning_rate": 3.819098685050428e-05,
|
| 229 |
-
"loss": 2.6021,
|
| 230 |
-
"step": 3700
|
| 231 |
-
},
|
| 232 |
-
{
|
| 233 |
-
"epoch": 0.44,
|
| 234 |
-
"learning_rate": 3.787182433295034e-05,
|
| 235 |
-
"loss": 2.6307,
|
| 236 |
-
"step": 3800
|
| 237 |
-
},
|
| 238 |
-
{
|
| 239 |
-
"epoch": 0.45,
|
| 240 |
-
"learning_rate": 3.75526618153964e-05,
|
| 241 |
-
"loss": 2.5632,
|
| 242 |
-
"step": 3900
|
| 243 |
-
},
|
| 244 |
-
{
|
| 245 |
-
"epoch": 0.46,
|
| 246 |
-
"learning_rate": 3.723349929784246e-05,
|
| 247 |
-
"loss": 2.5933,
|
| 248 |
-
"step": 4000
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 0.47,
|
| 252 |
-
"learning_rate": 3.6914336780288525e-05,
|
| 253 |
-
"loss": 2.5458,
|
| 254 |
-
"step": 4100
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 0.48,
|
| 258 |
-
"learning_rate": 3.659517426273459e-05,
|
| 259 |
-
"loss": 2.5212,
|
| 260 |
-
"step": 4200
|
| 261 |
-
},
|
| 262 |
-
{
|
| 263 |
-
"epoch": 0.49,
|
| 264 |
-
"learning_rate": 3.627601174518065e-05,
|
| 265 |
-
"loss": 2.4532,
|
| 266 |
-
"step": 4300
|
| 267 |
-
},
|
| 268 |
-
{
|
| 269 |
-
"epoch": 0.51,
|
| 270 |
-
"learning_rate": 3.595684922762671e-05,
|
| 271 |
-
"loss": 2.5134,
|
| 272 |
-
"step": 4400
|
| 273 |
-
},
|
| 274 |
-
{
|
| 275 |
-
"epoch": 0.52,
|
| 276 |
-
"learning_rate": 3.563768671007277e-05,
|
| 277 |
-
"loss": 2.5368,
|
| 278 |
-
"step": 4500
|
| 279 |
-
},
|
| 280 |
-
{
|
| 281 |
-
"epoch": 0.53,
|
| 282 |
-
"learning_rate": 3.531852419251883e-05,
|
| 283 |
-
"loss": 2.5547,
|
| 284 |
-
"step": 4600
|
| 285 |
-
},
|
| 286 |
-
{
|
| 287 |
-
"epoch": 0.54,
|
| 288 |
-
"learning_rate": 3.4999361674964895e-05,
|
| 289 |
-
"loss": 2.4786,
|
| 290 |
-
"step": 4700
|
| 291 |
-
},
|
| 292 |
-
{
|
| 293 |
-
"epoch": 0.55,
|
| 294 |
-
"learning_rate": 3.4680199157410954e-05,
|
| 295 |
-
"loss": 2.5377,
|
| 296 |
-
"step": 4800
|
| 297 |
-
},
|
| 298 |
-
{
|
| 299 |
-
"epoch": 0.56,
|
| 300 |
-
"learning_rate": 3.436103663985702e-05,
|
| 301 |
-
"loss": 2.5462,
|
| 302 |
-
"step": 4900
|
| 303 |
-
},
|
| 304 |
-
{
|
| 305 |
-
"epoch": 0.57,
|
| 306 |
-
"learning_rate": 3.404187412230308e-05,
|
| 307 |
-
"loss": 2.5644,
|
| 308 |
-
"step": 5000
|
| 309 |
-
},
|
| 310 |
-
{
|
| 311 |
-
"epoch": 0.59,
|
| 312 |
-
"learning_rate": 3.372271160474914e-05,
|
| 313 |
-
"loss": 2.5206,
|
| 314 |
-
"step": 5100
|
| 315 |
-
},
|
| 316 |
-
{
|
| 317 |
-
"epoch": 0.6,
|
| 318 |
-
"learning_rate": 3.3403549087195205e-05,
|
| 319 |
-
"loss": 2.5305,
|
| 320 |
-
"step": 5200
|
| 321 |
-
},
|
| 322 |
-
{
|
| 323 |
-
"epoch": 0.61,
|
| 324 |
-
"learning_rate": 3.3084386569641264e-05,
|
| 325 |
-
"loss": 2.462,
|
| 326 |
-
"step": 5300
|
| 327 |
-
},
|
| 328 |
-
{
|
| 329 |
-
"epoch": 0.62,
|
| 330 |
-
"learning_rate": 3.2765224052087324e-05,
|
| 331 |
-
"loss": 2.4704,
|
| 332 |
-
"step": 5400
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 0.63,
|
| 336 |
-
"learning_rate": 3.244606153453338e-05,
|
| 337 |
-
"loss": 2.5133,
|
| 338 |
-
"step": 5500
|
| 339 |
-
},
|
| 340 |
-
{
|
| 341 |
-
"epoch": 0.64,
|
| 342 |
-
"learning_rate": 3.212689901697944e-05,
|
| 343 |
-
"loss": 2.4805,
|
| 344 |
-
"step": 5600
|
| 345 |
-
},
|
| 346 |
-
{
|
| 347 |
-
"epoch": 0.65,
|
| 348 |
-
"learning_rate": 3.180773649942551e-05,
|
| 349 |
-
"loss": 2.4572,
|
| 350 |
-
"step": 5700
|
| 351 |
-
},
|
| 352 |
-
{
|
| 353 |
-
"epoch": 0.67,
|
| 354 |
-
"learning_rate": 3.1488573981871575e-05,
|
| 355 |
-
"loss": 2.4128,
|
| 356 |
-
"step": 5800
|
| 357 |
-
},
|
| 358 |
-
{
|
| 359 |
-
"epoch": 0.68,
|
| 360 |
-
"learning_rate": 3.1169411464317634e-05,
|
| 361 |
-
"loss": 2.4577,
|
| 362 |
-
"step": 5900
|
| 363 |
-
},
|
| 364 |
-
{
|
| 365 |
-
"epoch": 0.69,
|
| 366 |
-
"learning_rate": 3.0850248946763694e-05,
|
| 367 |
-
"loss": 2.5539,
|
| 368 |
-
"step": 6000
|
| 369 |
-
},
|
| 370 |
-
{
|
| 371 |
-
"epoch": 0.7,
|
| 372 |
-
"learning_rate": 3.053108642920975e-05,
|
| 373 |
-
"loss": 2.472,
|
| 374 |
-
"step": 6100
|
| 375 |
-
},
|
| 376 |
-
{
|
| 377 |
-
"epoch": 0.71,
|
| 378 |
-
"learning_rate": 3.0211923911655816e-05,
|
| 379 |
-
"loss": 2.4468,
|
| 380 |
-
"step": 6200
|
| 381 |
-
},
|
| 382 |
-
{
|
| 383 |
-
"epoch": 0.72,
|
| 384 |
-
"learning_rate": 2.9892761394101875e-05,
|
| 385 |
-
"loss": 2.4831,
|
| 386 |
-
"step": 6300
|
| 387 |
-
},
|
| 388 |
-
{
|
| 389 |
-
"epoch": 0.74,
|
| 390 |
-
"learning_rate": 2.957359887654794e-05,
|
| 391 |
-
"loss": 2.4093,
|
| 392 |
-
"step": 6400
|
| 393 |
-
},
|
| 394 |
-
{
|
| 395 |
-
"epoch": 0.75,
|
| 396 |
-
"learning_rate": 2.9254436358994004e-05,
|
| 397 |
-
"loss": 2.4683,
|
| 398 |
-
"step": 6500
|
| 399 |
-
},
|
| 400 |
-
{
|
| 401 |
-
"epoch": 0.76,
|
| 402 |
-
"learning_rate": 2.8935273841440063e-05,
|
| 403 |
-
"loss": 2.4388,
|
| 404 |
-
"step": 6600
|
| 405 |
-
},
|
| 406 |
-
{
|
| 407 |
-
"epoch": 0.77,
|
| 408 |
-
"learning_rate": 2.8616111323886126e-05,
|
| 409 |
-
"loss": 2.3558,
|
| 410 |
-
"step": 6700
|
| 411 |
-
},
|
| 412 |
-
{
|
| 413 |
-
"epoch": 0.78,
|
| 414 |
-
"learning_rate": 2.8296948806332185e-05,
|
| 415 |
-
"loss": 2.4788,
|
| 416 |
-
"step": 6800
|
| 417 |
-
},
|
| 418 |
-
{
|
| 419 |
-
"epoch": 0.79,
|
| 420 |
-
"learning_rate": 2.7977786288778245e-05,
|
| 421 |
-
"loss": 2.4523,
|
| 422 |
-
"step": 6900
|
| 423 |
-
},
|
| 424 |
-
{
|
| 425 |
-
"epoch": 0.8,
|
| 426 |
-
"learning_rate": 2.7658623771224308e-05,
|
| 427 |
-
"loss": 2.4759,
|
| 428 |
-
"step": 7000
|
| 429 |
-
},
|
| 430 |
-
{
|
| 431 |
-
"epoch": 0.82,
|
| 432 |
-
"learning_rate": 2.7339461253670367e-05,
|
| 433 |
-
"loss": 2.3574,
|
| 434 |
-
"step": 7100
|
| 435 |
-
},
|
| 436 |
-
{
|
| 437 |
-
"epoch": 0.83,
|
| 438 |
-
"learning_rate": 2.7020298736116433e-05,
|
| 439 |
-
"loss": 2.4569,
|
| 440 |
-
"step": 7200
|
| 441 |
-
},
|
| 442 |
-
{
|
| 443 |
-
"epoch": 0.84,
|
| 444 |
-
"learning_rate": 2.6701136218562496e-05,
|
| 445 |
-
"loss": 2.4078,
|
| 446 |
-
"step": 7300
|
| 447 |
-
},
|
| 448 |
-
{
|
| 449 |
-
"epoch": 0.85,
|
| 450 |
-
"learning_rate": 2.6381973701008555e-05,
|
| 451 |
-
"loss": 2.3948,
|
| 452 |
-
"step": 7400
|
| 453 |
-
},
|
| 454 |
-
{
|
| 455 |
-
"epoch": 0.86,
|
| 456 |
-
"learning_rate": 2.6062811183454615e-05,
|
| 457 |
-
"loss": 2.5496,
|
| 458 |
-
"step": 7500
|
| 459 |
-
},
|
| 460 |
-
{
|
| 461 |
-
"epoch": 0.87,
|
| 462 |
-
"learning_rate": 2.5743648665900677e-05,
|
| 463 |
-
"loss": 2.4191,
|
| 464 |
-
"step": 7600
|
| 465 |
-
},
|
| 466 |
-
{
|
| 467 |
-
"epoch": 0.88,
|
| 468 |
-
"learning_rate": 2.5424486148346737e-05,
|
| 469 |
-
"loss": 2.4895,
|
| 470 |
-
"step": 7700
|
| 471 |
-
},
|
| 472 |
-
{
|
| 473 |
-
"epoch": 0.9,
|
| 474 |
-
"learning_rate": 2.51053236307928e-05,
|
| 475 |
-
"loss": 2.4726,
|
| 476 |
-
"step": 7800
|
| 477 |
-
},
|
| 478 |
-
{
|
| 479 |
-
"epoch": 0.91,
|
| 480 |
-
"learning_rate": 2.4786161113238862e-05,
|
| 481 |
-
"loss": 2.3287,
|
| 482 |
-
"step": 7900
|
| 483 |
-
},
|
| 484 |
-
{
|
| 485 |
-
"epoch": 0.92,
|
| 486 |
-
"learning_rate": 2.446699859568492e-05,
|
| 487 |
-
"loss": 2.4278,
|
| 488 |
-
"step": 8000
|
| 489 |
-
},
|
| 490 |
-
{
|
| 491 |
-
"epoch": 0.93,
|
| 492 |
-
"learning_rate": 2.4147836078130988e-05,
|
| 493 |
-
"loss": 2.4548,
|
| 494 |
-
"step": 8100
|
| 495 |
-
},
|
| 496 |
-
{
|
| 497 |
-
"epoch": 0.94,
|
| 498 |
-
"learning_rate": 2.3828673560577047e-05,
|
| 499 |
-
"loss": 2.5173,
|
| 500 |
-
"step": 8200
|
| 501 |
-
},
|
| 502 |
-
{
|
| 503 |
-
"epoch": 0.95,
|
| 504 |
-
"learning_rate": 2.3509511043023107e-05,
|
| 505 |
-
"loss": 2.4883,
|
| 506 |
-
"step": 8300
|
| 507 |
-
},
|
| 508 |
-
{
|
| 509 |
-
"epoch": 0.97,
|
| 510 |
-
"learning_rate": 2.319034852546917e-05,
|
| 511 |
-
"loss": 2.4435,
|
| 512 |
-
"step": 8400
|
| 513 |
-
},
|
| 514 |
-
{
|
| 515 |
-
"epoch": 0.98,
|
| 516 |
-
"learning_rate": 2.2871186007915232e-05,
|
| 517 |
-
"loss": 2.4825,
|
| 518 |
-
"step": 8500
|
| 519 |
-
},
|
| 520 |
-
{
|
| 521 |
-
"epoch": 0.99,
|
| 522 |
-
"learning_rate": 2.2552023490361295e-05,
|
| 523 |
-
"loss": 2.4539,
|
| 524 |
-
"step": 8600
|
| 525 |
-
},
|
| 526 |
-
{
|
| 527 |
-
"epoch": 1.0,
|
| 528 |
-
"learning_rate": 2.2232860972807354e-05,
|
| 529 |
-
"loss": 2.4115,
|
| 530 |
-
"step": 8700
|
| 531 |
-
},
|
| 532 |
-
{
|
| 533 |
-
"epoch": 1.01,
|
| 534 |
-
"learning_rate": 2.1913698455253417e-05,
|
| 535 |
-
"loss": 2.3452,
|
| 536 |
-
"step": 8800
|
| 537 |
-
},
|
| 538 |
-
{
|
| 539 |
-
"epoch": 1.02,
|
| 540 |
-
"learning_rate": 2.159453593769948e-05,
|
| 541 |
-
"loss": 2.3799,
|
| 542 |
-
"step": 8900
|
| 543 |
-
},
|
| 544 |
-
{
|
| 545 |
-
"epoch": 1.03,
|
| 546 |
-
"learning_rate": 2.127537342014554e-05,
|
| 547 |
-
"loss": 2.4019,
|
| 548 |
-
"step": 9000
|
| 549 |
-
},
|
| 550 |
-
{
|
| 551 |
-
"epoch": 1.05,
|
| 552 |
-
"learning_rate": 2.09562109025916e-05,
|
| 553 |
-
"loss": 2.3678,
|
| 554 |
-
"step": 9100
|
| 555 |
-
},
|
| 556 |
-
{
|
| 557 |
-
"epoch": 1.06,
|
| 558 |
-
"learning_rate": 2.0637048385037665e-05,
|
| 559 |
-
"loss": 2.3711,
|
| 560 |
-
"step": 9200
|
| 561 |
-
},
|
| 562 |
-
{
|
| 563 |
-
"epoch": 1.07,
|
| 564 |
-
"learning_rate": 2.0317885867483724e-05,
|
| 565 |
-
"loss": 2.2767,
|
| 566 |
-
"step": 9300
|
| 567 |
-
},
|
| 568 |
-
{
|
| 569 |
-
"epoch": 1.08,
|
| 570 |
-
"learning_rate": 1.9998723349929783e-05,
|
| 571 |
-
"loss": 2.3761,
|
| 572 |
-
"step": 9400
|
| 573 |
-
},
|
| 574 |
-
{
|
| 575 |
-
"epoch": 1.09,
|
| 576 |
-
"learning_rate": 1.9679560832375846e-05,
|
| 577 |
-
"loss": 2.3392,
|
| 578 |
-
"step": 9500
|
| 579 |
-
},
|
| 580 |
-
{
|
| 581 |
-
"epoch": 1.1,
|
| 582 |
-
"learning_rate": 1.936039831482191e-05,
|
| 583 |
-
"loss": 2.3232,
|
| 584 |
-
"step": 9600
|
| 585 |
-
},
|
| 586 |
-
{
|
| 587 |
-
"epoch": 1.11,
|
| 588 |
-
"learning_rate": 1.904123579726797e-05,
|
| 589 |
-
"loss": 2.355,
|
| 590 |
-
"step": 9700
|
| 591 |
-
},
|
| 592 |
-
{
|
| 593 |
-
"epoch": 1.13,
|
| 594 |
-
"learning_rate": 1.872207327971403e-05,
|
| 595 |
-
"loss": 2.3613,
|
| 596 |
-
"step": 9800
|
| 597 |
-
},
|
| 598 |
-
{
|
| 599 |
-
"epoch": 1.14,
|
| 600 |
-
"learning_rate": 1.840291076216009e-05,
|
| 601 |
-
"loss": 2.3475,
|
| 602 |
-
"step": 9900
|
| 603 |
-
},
|
| 604 |
-
{
|
| 605 |
-
"epoch": 1.15,
|
| 606 |
-
"learning_rate": 1.8083748244606157e-05,
|
| 607 |
-
"loss": 2.3654,
|
| 608 |
-
"step": 10000
|
| 609 |
-
},
|
| 610 |
-
{
|
| 611 |
-
"epoch": 1.16,
|
| 612 |
-
"learning_rate": 1.7764585727052216e-05,
|
| 613 |
-
"loss": 2.3825,
|
| 614 |
-
"step": 10100
|
| 615 |
-
},
|
| 616 |
-
{
|
| 617 |
-
"epoch": 1.17,
|
| 618 |
-
"learning_rate": 1.7445423209498275e-05,
|
| 619 |
-
"loss": 2.3245,
|
| 620 |
-
"step": 10200
|
| 621 |
-
},
|
| 622 |
-
{
|
| 623 |
-
"epoch": 1.18,
|
| 624 |
-
"learning_rate": 1.712626069194434e-05,
|
| 625 |
-
"loss": 2.2869,
|
| 626 |
-
"step": 10300
|
| 627 |
-
},
|
| 628 |
-
{
|
| 629 |
-
"epoch": 1.19,
|
| 630 |
-
"learning_rate": 1.68070981743904e-05,
|
| 631 |
-
"loss": 2.3207,
|
| 632 |
-
"step": 10400
|
| 633 |
-
},
|
| 634 |
-
{
|
| 635 |
-
"epoch": 1.21,
|
| 636 |
-
"learning_rate": 1.648793565683646e-05,
|
| 637 |
-
"loss": 2.3041,
|
| 638 |
-
"step": 10500
|
| 639 |
-
},
|
| 640 |
-
{
|
| 641 |
-
"epoch": 1.22,
|
| 642 |
-
"learning_rate": 1.6168773139282523e-05,
|
| 643 |
-
"loss": 2.3543,
|
| 644 |
-
"step": 10600
|
| 645 |
-
},
|
| 646 |
-
{
|
| 647 |
-
"epoch": 1.23,
|
| 648 |
-
"learning_rate": 1.5849610621728586e-05,
|
| 649 |
-
"loss": 2.3026,
|
| 650 |
-
"step": 10700
|
| 651 |
-
},
|
| 652 |
-
{
|
| 653 |
-
"epoch": 1.24,
|
| 654 |
-
"learning_rate": 1.553044810417465e-05,
|
| 655 |
-
"loss": 2.2919,
|
| 656 |
-
"step": 10800
|
| 657 |
-
},
|
| 658 |
-
{
|
| 659 |
-
"epoch": 1.25,
|
| 660 |
-
"learning_rate": 1.5211285586620708e-05,
|
| 661 |
-
"loss": 2.3805,
|
| 662 |
-
"step": 10900
|
| 663 |
-
},
|
| 664 |
-
{
|
| 665 |
-
"epoch": 1.26,
|
| 666 |
-
"learning_rate": 1.4892123069066769e-05,
|
| 667 |
-
"loss": 2.3523,
|
| 668 |
-
"step": 11000
|
| 669 |
-
},
|
| 670 |
-
{
|
| 671 |
-
"epoch": 1.28,
|
| 672 |
-
"learning_rate": 1.4572960551512832e-05,
|
| 673 |
-
"loss": 2.2931,
|
| 674 |
-
"step": 11100
|
| 675 |
-
},
|
| 676 |
-
{
|
| 677 |
-
"epoch": 1.29,
|
| 678 |
-
"learning_rate": 1.4253798033958893e-05,
|
| 679 |
-
"loss": 2.2623,
|
| 680 |
-
"step": 11200
|
| 681 |
-
},
|
| 682 |
-
{
|
| 683 |
-
"epoch": 1.3,
|
| 684 |
-
"learning_rate": 1.3934635516404954e-05,
|
| 685 |
-
"loss": 2.28,
|
| 686 |
-
"step": 11300
|
| 687 |
-
},
|
| 688 |
-
{
|
| 689 |
-
"epoch": 1.31,
|
| 690 |
-
"learning_rate": 1.3615472998851015e-05,
|
| 691 |
-
"loss": 2.3121,
|
| 692 |
-
"step": 11400
|
| 693 |
-
},
|
| 694 |
-
{
|
| 695 |
-
"epoch": 1.32,
|
| 696 |
-
"learning_rate": 1.3296310481297078e-05,
|
| 697 |
-
"loss": 2.3111,
|
| 698 |
-
"step": 11500
|
| 699 |
-
},
|
| 700 |
-
{
|
| 701 |
-
"epoch": 1.33,
|
| 702 |
-
"learning_rate": 1.2977147963743139e-05,
|
| 703 |
-
"loss": 2.332,
|
| 704 |
-
"step": 11600
|
| 705 |
-
},
|
| 706 |
-
{
|
| 707 |
-
"epoch": 1.34,
|
| 708 |
-
"learning_rate": 1.26579854461892e-05,
|
| 709 |
-
"loss": 2.3495,
|
| 710 |
-
"step": 11700
|
| 711 |
-
},
|
| 712 |
-
{
|
| 713 |
-
"epoch": 1.36,
|
| 714 |
-
"learning_rate": 1.233882292863526e-05,
|
| 715 |
-
"loss": 2.2864,
|
| 716 |
-
"step": 11800
|
| 717 |
-
},
|
| 718 |
-
{
|
| 719 |
-
"epoch": 1.37,
|
| 720 |
-
"learning_rate": 1.2019660411081324e-05,
|
| 721 |
-
"loss": 2.3482,
|
| 722 |
-
"step": 11900
|
| 723 |
-
},
|
| 724 |
-
{
|
| 725 |
-
"epoch": 1.38,
|
| 726 |
-
"learning_rate": 1.1700497893527385e-05,
|
| 727 |
-
"loss": 2.2843,
|
| 728 |
-
"step": 12000
|
| 729 |
-
},
|
| 730 |
-
{
|
| 731 |
-
"epoch": 1.39,
|
| 732 |
-
"learning_rate": 1.1381335375973447e-05,
|
| 733 |
-
"loss": 2.3075,
|
| 734 |
-
"step": 12100
|
| 735 |
-
},
|
| 736 |
-
{
|
| 737 |
-
"epoch": 1.4,
|
| 738 |
-
"learning_rate": 1.1062172858419507e-05,
|
| 739 |
-
"loss": 2.3899,
|
| 740 |
-
"step": 12200
|
| 741 |
-
},
|
| 742 |
-
{
|
| 743 |
-
"epoch": 1.41,
|
| 744 |
-
"learning_rate": 1.074301034086557e-05,
|
| 745 |
-
"loss": 2.3496,
|
| 746 |
-
"step": 12300
|
| 747 |
-
},
|
| 748 |
-
{
|
| 749 |
-
"epoch": 1.42,
|
| 750 |
-
"learning_rate": 1.042384782331163e-05,
|
| 751 |
-
"loss": 2.3482,
|
| 752 |
-
"step": 12400
|
| 753 |
-
},
|
| 754 |
-
{
|
| 755 |
-
"epoch": 1.44,
|
| 756 |
-
"learning_rate": 1.0104685305757693e-05,
|
| 757 |
-
"loss": 2.2579,
|
| 758 |
-
"step": 12500
|
| 759 |
-
},
|
| 760 |
-
{
|
| 761 |
-
"epoch": 1.45,
|
| 762 |
-
"learning_rate": 9.785522788203753e-06,
|
| 763 |
-
"loss": 2.2689,
|
| 764 |
-
"step": 12600
|
| 765 |
-
},
|
| 766 |
-
{
|
| 767 |
-
"epoch": 1.46,
|
| 768 |
-
"learning_rate": 9.466360270649816e-06,
|
| 769 |
-
"loss": 2.3111,
|
| 770 |
-
"step": 12700
|
| 771 |
-
},
|
| 772 |
-
{
|
| 773 |
-
"epoch": 1.47,
|
| 774 |
-
"learning_rate": 9.147197753095877e-06,
|
| 775 |
-
"loss": 2.2794,
|
| 776 |
-
"step": 12800
|
| 777 |
-
},
|
| 778 |
-
{
|
| 779 |
-
"epoch": 1.48,
|
| 780 |
-
"learning_rate": 8.82803523554194e-06,
|
| 781 |
-
"loss": 2.2754,
|
| 782 |
-
"step": 12900
|
| 783 |
-
},
|
| 784 |
-
{
|
| 785 |
-
"epoch": 1.49,
|
| 786 |
-
"learning_rate": 8.508872717987999e-06,
|
| 787 |
-
"loss": 2.2837,
|
| 788 |
-
"step": 13000
|
| 789 |
-
},
|
| 790 |
-
{
|
| 791 |
-
"epoch": 1.51,
|
| 792 |
-
"learning_rate": 8.189710200434061e-06,
|
| 793 |
-
"loss": 2.2297,
|
| 794 |
-
"step": 13100
|
| 795 |
-
},
|
| 796 |
-
{
|
| 797 |
-
"epoch": 1.52,
|
| 798 |
-
"learning_rate": 7.870547682880124e-06,
|
| 799 |
-
"loss": 2.2969,
|
| 800 |
-
"step": 13200
|
| 801 |
-
},
|
| 802 |
-
{
|
| 803 |
-
"epoch": 1.53,
|
| 804 |
-
"learning_rate": 7.5513851653261844e-06,
|
| 805 |
-
"loss": 2.3363,
|
| 806 |
-
"step": 13300
|
| 807 |
-
},
|
| 808 |
-
{
|
| 809 |
-
"epoch": 1.54,
|
| 810 |
-
"learning_rate": 7.232222647772246e-06,
|
| 811 |
-
"loss": 2.2079,
|
| 812 |
-
"step": 13400
|
| 813 |
-
},
|
| 814 |
-
{
|
| 815 |
-
"epoch": 1.55,
|
| 816 |
-
"learning_rate": 6.9130601302183074e-06,
|
| 817 |
-
"loss": 2.3057,
|
| 818 |
-
"step": 13500
|
| 819 |
-
},
|
| 820 |
-
{
|
| 821 |
-
"epoch": 1.56,
|
| 822 |
-
"learning_rate": 6.593897612664369e-06,
|
| 823 |
-
"loss": 2.2665,
|
| 824 |
-
"step": 13600
|
| 825 |
-
},
|
| 826 |
-
{
|
| 827 |
-
"epoch": 1.57,
|
| 828 |
-
"learning_rate": 6.27473509511043e-06,
|
| 829 |
-
"loss": 2.2934,
|
| 830 |
-
"step": 13700
|
| 831 |
-
},
|
| 832 |
-
{
|
| 833 |
-
"epoch": 1.59,
|
| 834 |
-
"learning_rate": 5.9555725775564915e-06,
|
| 835 |
-
"loss": 2.2407,
|
| 836 |
-
"step": 13800
|
| 837 |
-
},
|
| 838 |
-
{
|
| 839 |
-
"epoch": 1.6,
|
| 840 |
-
"learning_rate": 5.636410060002553e-06,
|
| 841 |
-
"loss": 2.3309,
|
| 842 |
-
"step": 13900
|
| 843 |
-
},
|
| 844 |
-
{
|
| 845 |
-
"epoch": 1.61,
|
| 846 |
-
"learning_rate": 5.317247542448615e-06,
|
| 847 |
-
"loss": 2.3257,
|
| 848 |
-
"step": 14000
|
| 849 |
-
},
|
| 850 |
-
{
|
| 851 |
-
"epoch": 1.62,
|
| 852 |
-
"learning_rate": 4.998085024894677e-06,
|
| 853 |
-
"loss": 2.2518,
|
| 854 |
-
"step": 14100
|
| 855 |
-
},
|
| 856 |
-
{
|
| 857 |
-
"epoch": 1.63,
|
| 858 |
-
"learning_rate": 4.678922507340738e-06,
|
| 859 |
-
"loss": 2.2359,
|
| 860 |
-
"step": 14200
|
| 861 |
-
},
|
| 862 |
-
{
|
| 863 |
-
"epoch": 1.64,
|
| 864 |
-
"learning_rate": 4.3597599897868e-06,
|
| 865 |
-
"loss": 2.287,
|
| 866 |
-
"step": 14300
|
| 867 |
-
},
|
| 868 |
-
{
|
| 869 |
-
"epoch": 1.65,
|
| 870 |
-
"learning_rate": 4.040597472232861e-06,
|
| 871 |
-
"loss": 2.2462,
|
| 872 |
-
"step": 14400
|
| 873 |
-
},
|
| 874 |
-
{
|
| 875 |
-
"epoch": 1.67,
|
| 876 |
-
"learning_rate": 3.7214349546789228e-06,
|
| 877 |
-
"loss": 2.2815,
|
| 878 |
-
"step": 14500
|
| 879 |
-
},
|
| 880 |
-
{
|
| 881 |
-
"epoch": 1.68,
|
| 882 |
-
"learning_rate": 3.4022724371249842e-06,
|
| 883 |
-
"loss": 2.3361,
|
| 884 |
-
"step": 14600
|
| 885 |
-
},
|
| 886 |
-
{
|
| 887 |
-
"epoch": 1.69,
|
| 888 |
-
"learning_rate": 3.0831099195710457e-06,
|
| 889 |
-
"loss": 2.2774,
|
| 890 |
-
"step": 14700
|
| 891 |
-
},
|
| 892 |
-
{
|
| 893 |
-
"epoch": 1.7,
|
| 894 |
-
"learning_rate": 2.7639474020171072e-06,
|
| 895 |
-
"loss": 2.2895,
|
| 896 |
-
"step": 14800
|
| 897 |
-
},
|
| 898 |
-
{
|
| 899 |
-
"epoch": 1.71,
|
| 900 |
-
"learning_rate": 2.4447848844631687e-06,
|
| 901 |
-
"loss": 2.3155,
|
| 902 |
-
"step": 14900
|
| 903 |
-
},
|
| 904 |
-
{
|
| 905 |
-
"epoch": 1.72,
|
| 906 |
-
"learning_rate": 2.12562236690923e-06,
|
| 907 |
-
"loss": 2.2698,
|
| 908 |
-
"step": 15000
|
| 909 |
-
}
|
| 910 |
-
],
|
| 911 |
-
"max_steps": 15666,
|
| 912 |
-
"num_train_epochs": 2,
|
| 913 |
-
"total_flos": 5.928520220872704e+16,
|
| 914 |
-
"trial_name": null,
|
| 915 |
-
"trial_params": null
|
| 916 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/training_args.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:fb5c03e02cabfaae5aac527cba1260f324dabff8089669a909b39dcb2a587966
|
| 3 |
-
size 3439
|
|
|
|
|
|
|
|
|
|
|
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 557969145
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51feb8f9eb40ac13ea3bcbdaa38ed1fee7ea527683f56fba10d5981cf701327b
|
| 3 |
size 557969145
|
runs/Jan01_04-28-26_6ea5fb75c867/events.out.tfevents.1672547319.6ea5fb75c867.24.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e44496a858f60435d4954eafca743d7925fa568697775815563c3c7d839e61ce
|
| 3 |
+
size 29681
|