End of training
Browse files- last-checkpoint/generation_config.json → generation_config.json +0 -0
- last-checkpoint/config.json +0 -31
- 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/special_tokens_map.json +0 -5
- last-checkpoint/spiece.model +0 -3
- last-checkpoint/tokenizer_config.json +0 -12
- last-checkpoint/trainer_state.json +0 -412
- last-checkpoint/training_args.bin +0 -3
- pytorch_model.bin +1 -1
- runs/Jan26_09-04-05_e112ce585155/events.out.tfevents.1674723897.e112ce585155.6073.0 +2 -2
last-checkpoint/generation_config.json → generation_config.json
RENAMED
|
File without changes
|
last-checkpoint/config.json
DELETED
|
@@ -1,31 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"_name_or_path": "google/mt5-small",
|
| 3 |
-
"architectures": [
|
| 4 |
-
"MT5ForConditionalGeneration"
|
| 5 |
-
],
|
| 6 |
-
"d_ff": 1024,
|
| 7 |
-
"d_kv": 64,
|
| 8 |
-
"d_model": 512,
|
| 9 |
-
"decoder_start_token_id": 0,
|
| 10 |
-
"dense_act_fn": "gelu_new",
|
| 11 |
-
"dropout_rate": 0.1,
|
| 12 |
-
"eos_token_id": 1,
|
| 13 |
-
"feed_forward_proj": "gated-gelu",
|
| 14 |
-
"initializer_factor": 1.0,
|
| 15 |
-
"is_encoder_decoder": true,
|
| 16 |
-
"is_gated_act": true,
|
| 17 |
-
"layer_norm_epsilon": 1e-06,
|
| 18 |
-
"model_type": "mt5",
|
| 19 |
-
"num_decoder_layers": 8,
|
| 20 |
-
"num_heads": 6,
|
| 21 |
-
"num_layers": 8,
|
| 22 |
-
"pad_token_id": 0,
|
| 23 |
-
"relative_attention_max_distance": 128,
|
| 24 |
-
"relative_attention_num_buckets": 32,
|
| 25 |
-
"tie_word_embeddings": false,
|
| 26 |
-
"tokenizer_class": "T5Tokenizer",
|
| 27 |
-
"torch_dtype": "float32",
|
| 28 |
-
"transformers_version": "4.26.0",
|
| 29 |
-
"use_cache": true,
|
| 30 |
-
"vocab_size": 250112
|
| 31 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/optimizer.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:92cbacc594fc4d03f876f678c8368ba06af5f801fe986d52c0c77eea78d9d059
|
| 3 |
-
size 2401526789
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/pytorch_model.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:9171600923799fbd736ef09a97dae709568b02cf9feabddcab70945814d96c31
|
| 3 |
-
size 1200772485
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/rng_state.pth
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:7c2171553dfbbc27f33003f3725022aa0ab9cd0873edefc1d01d9617498a1ba4
|
| 3 |
-
size 14575
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/scheduler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:9937e95b70c0680ecf96691c44cc3b737c61cf8ec7c37caf61a209d12a4a70e7
|
| 3 |
-
size 627
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/special_tokens_map.json
DELETED
|
@@ -1,5 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"eos_token": "</s>",
|
| 3 |
-
"pad_token": "<pad>",
|
| 4 |
-
"unk_token": "<unk>"
|
| 5 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/spiece.model
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6
|
| 3 |
-
size 4309802
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/tokenizer_config.json
DELETED
|
@@ -1,12 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"additional_special_tokens": null,
|
| 3 |
-
"eos_token": "</s>",
|
| 4 |
-
"extra_ids": 0,
|
| 5 |
-
"model_max_length": 1000000000000000019884624838656,
|
| 6 |
-
"name_or_path": "google/mt5-small",
|
| 7 |
-
"pad_token": "<pad>",
|
| 8 |
-
"sp_model_kwargs": {},
|
| 9 |
-
"special_tokens_map_file": "/root/.cache/huggingface/hub/models--google--mt5-small/snapshots/38f23af8ec210eb6c376d40e9c56bd25a80f195d/special_tokens_map.json",
|
| 10 |
-
"tokenizer_class": "T5Tokenizer",
|
| 11 |
-
"unk_token": "<unk>"
|
| 12 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/trainer_state.json
DELETED
|
@@ -1,412 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_metric": null,
|
| 3 |
-
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 2.65152,
|
| 5 |
-
"global_step": 33144,
|
| 6 |
-
"is_hyper_param_search": false,
|
| 7 |
-
"is_local_process_zero": true,
|
| 8 |
-
"is_world_process_zero": true,
|
| 9 |
-
"log_history": [
|
| 10 |
-
{
|
| 11 |
-
"epoch": 0.04,
|
| 12 |
-
"learning_rate": 1.9733333333333336e-05,
|
| 13 |
-
"loss": 14.1239,
|
| 14 |
-
"step": 500
|
| 15 |
-
},
|
| 16 |
-
{
|
| 17 |
-
"epoch": 0.08,
|
| 18 |
-
"learning_rate": 1.9466666666666668e-05,
|
| 19 |
-
"loss": 6.6454,
|
| 20 |
-
"step": 1000
|
| 21 |
-
},
|
| 22 |
-
{
|
| 23 |
-
"epoch": 0.12,
|
| 24 |
-
"learning_rate": 1.9200000000000003e-05,
|
| 25 |
-
"loss": 5.9138,
|
| 26 |
-
"step": 1500
|
| 27 |
-
},
|
| 28 |
-
{
|
| 29 |
-
"epoch": 0.16,
|
| 30 |
-
"learning_rate": 1.8933333333333334e-05,
|
| 31 |
-
"loss": 5.5971,
|
| 32 |
-
"step": 2000
|
| 33 |
-
},
|
| 34 |
-
{
|
| 35 |
-
"epoch": 0.2,
|
| 36 |
-
"learning_rate": 1.866666666666667e-05,
|
| 37 |
-
"loss": 5.4037,
|
| 38 |
-
"step": 2500
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.24,
|
| 42 |
-
"learning_rate": 1.8400000000000003e-05,
|
| 43 |
-
"loss": 5.2453,
|
| 44 |
-
"step": 3000
|
| 45 |
-
},
|
| 46 |
-
{
|
| 47 |
-
"epoch": 0.28,
|
| 48 |
-
"learning_rate": 1.8133333333333335e-05,
|
| 49 |
-
"loss": 5.1776,
|
| 50 |
-
"step": 3500
|
| 51 |
-
},
|
| 52 |
-
{
|
| 53 |
-
"epoch": 0.32,
|
| 54 |
-
"learning_rate": 1.7866666666666666e-05,
|
| 55 |
-
"loss": 5.0923,
|
| 56 |
-
"step": 4000
|
| 57 |
-
},
|
| 58 |
-
{
|
| 59 |
-
"epoch": 0.36,
|
| 60 |
-
"learning_rate": 1.76e-05,
|
| 61 |
-
"loss": 5.0219,
|
| 62 |
-
"step": 4500
|
| 63 |
-
},
|
| 64 |
-
{
|
| 65 |
-
"epoch": 0.4,
|
| 66 |
-
"learning_rate": 1.7333333333333336e-05,
|
| 67 |
-
"loss": 4.9679,
|
| 68 |
-
"step": 5000
|
| 69 |
-
},
|
| 70 |
-
{
|
| 71 |
-
"epoch": 0.44,
|
| 72 |
-
"learning_rate": 1.706666666666667e-05,
|
| 73 |
-
"loss": 4.8737,
|
| 74 |
-
"step": 5500
|
| 75 |
-
},
|
| 76 |
-
{
|
| 77 |
-
"epoch": 0.48,
|
| 78 |
-
"learning_rate": 1.6800000000000002e-05,
|
| 79 |
-
"loss": 4.8347,
|
| 80 |
-
"step": 6000
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.52,
|
| 84 |
-
"learning_rate": 1.6533333333333333e-05,
|
| 85 |
-
"loss": 4.7651,
|
| 86 |
-
"step": 6500
|
| 87 |
-
},
|
| 88 |
-
{
|
| 89 |
-
"epoch": 0.56,
|
| 90 |
-
"learning_rate": 1.6266666666666668e-05,
|
| 91 |
-
"loss": 4.7169,
|
| 92 |
-
"step": 7000
|
| 93 |
-
},
|
| 94 |
-
{
|
| 95 |
-
"epoch": 0.6,
|
| 96 |
-
"learning_rate": 1.6000000000000003e-05,
|
| 97 |
-
"loss": 4.7314,
|
| 98 |
-
"step": 7500
|
| 99 |
-
},
|
| 100 |
-
{
|
| 101 |
-
"epoch": 0.64,
|
| 102 |
-
"learning_rate": 1.5733333333333334e-05,
|
| 103 |
-
"loss": 4.6803,
|
| 104 |
-
"step": 8000
|
| 105 |
-
},
|
| 106 |
-
{
|
| 107 |
-
"epoch": 0.68,
|
| 108 |
-
"learning_rate": 1.546666666666667e-05,
|
| 109 |
-
"loss": 4.627,
|
| 110 |
-
"step": 8500
|
| 111 |
-
},
|
| 112 |
-
{
|
| 113 |
-
"epoch": 0.72,
|
| 114 |
-
"learning_rate": 1.5200000000000002e-05,
|
| 115 |
-
"loss": 4.5979,
|
| 116 |
-
"step": 9000
|
| 117 |
-
},
|
| 118 |
-
{
|
| 119 |
-
"epoch": 0.76,
|
| 120 |
-
"learning_rate": 1.4933333333333335e-05,
|
| 121 |
-
"loss": 4.546,
|
| 122 |
-
"step": 9500
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.8,
|
| 126 |
-
"learning_rate": 1.4666666666666666e-05,
|
| 127 |
-
"loss": 4.557,
|
| 128 |
-
"step": 10000
|
| 129 |
-
},
|
| 130 |
-
{
|
| 131 |
-
"epoch": 0.84,
|
| 132 |
-
"learning_rate": 1.4400000000000001e-05,
|
| 133 |
-
"loss": 4.5273,
|
| 134 |
-
"step": 10500
|
| 135 |
-
},
|
| 136 |
-
{
|
| 137 |
-
"epoch": 0.88,
|
| 138 |
-
"learning_rate": 1.4133333333333334e-05,
|
| 139 |
-
"loss": 4.4533,
|
| 140 |
-
"step": 11000
|
| 141 |
-
},
|
| 142 |
-
{
|
| 143 |
-
"epoch": 0.92,
|
| 144 |
-
"learning_rate": 1.3866666666666669e-05,
|
| 145 |
-
"loss": 4.4465,
|
| 146 |
-
"step": 11500
|
| 147 |
-
},
|
| 148 |
-
{
|
| 149 |
-
"epoch": 0.96,
|
| 150 |
-
"learning_rate": 1.3600000000000002e-05,
|
| 151 |
-
"loss": 4.4248,
|
| 152 |
-
"step": 12000
|
| 153 |
-
},
|
| 154 |
-
{
|
| 155 |
-
"epoch": 1.0,
|
| 156 |
-
"learning_rate": 1.3333333333333333e-05,
|
| 157 |
-
"loss": 4.414,
|
| 158 |
-
"step": 12500
|
| 159 |
-
},
|
| 160 |
-
{
|
| 161 |
-
"epoch": 1.04,
|
| 162 |
-
"learning_rate": 1.3066666666666668e-05,
|
| 163 |
-
"loss": 4.3951,
|
| 164 |
-
"step": 13000
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 1.08,
|
| 168 |
-
"learning_rate": 1.2800000000000001e-05,
|
| 169 |
-
"loss": 4.3573,
|
| 170 |
-
"step": 13500
|
| 171 |
-
},
|
| 172 |
-
{
|
| 173 |
-
"epoch": 1.12,
|
| 174 |
-
"learning_rate": 1.2533333333333336e-05,
|
| 175 |
-
"loss": 4.338,
|
| 176 |
-
"step": 14000
|
| 177 |
-
},
|
| 178 |
-
{
|
| 179 |
-
"epoch": 1.16,
|
| 180 |
-
"learning_rate": 1.2266666666666667e-05,
|
| 181 |
-
"loss": 4.3254,
|
| 182 |
-
"step": 14500
|
| 183 |
-
},
|
| 184 |
-
{
|
| 185 |
-
"epoch": 1.2,
|
| 186 |
-
"learning_rate": 1.2e-05,
|
| 187 |
-
"loss": 4.3212,
|
| 188 |
-
"step": 15000
|
| 189 |
-
},
|
| 190 |
-
{
|
| 191 |
-
"epoch": 1.24,
|
| 192 |
-
"learning_rate": 1.1733333333333335e-05,
|
| 193 |
-
"loss": 4.2952,
|
| 194 |
-
"step": 15500
|
| 195 |
-
},
|
| 196 |
-
{
|
| 197 |
-
"epoch": 1.28,
|
| 198 |
-
"learning_rate": 1.1466666666666668e-05,
|
| 199 |
-
"loss": 4.2643,
|
| 200 |
-
"step": 16000
|
| 201 |
-
},
|
| 202 |
-
{
|
| 203 |
-
"epoch": 1.32,
|
| 204 |
-
"learning_rate": 1.1200000000000001e-05,
|
| 205 |
-
"loss": 4.2386,
|
| 206 |
-
"step": 16500
|
| 207 |
-
},
|
| 208 |
-
{
|
| 209 |
-
"epoch": 1.36,
|
| 210 |
-
"learning_rate": 1.0933333333333334e-05,
|
| 211 |
-
"loss": 4.2501,
|
| 212 |
-
"step": 17000
|
| 213 |
-
},
|
| 214 |
-
{
|
| 215 |
-
"epoch": 1.4,
|
| 216 |
-
"learning_rate": 1.0666666666666667e-05,
|
| 217 |
-
"loss": 4.2668,
|
| 218 |
-
"step": 17500
|
| 219 |
-
},
|
| 220 |
-
{
|
| 221 |
-
"epoch": 1.44,
|
| 222 |
-
"learning_rate": 1.04e-05,
|
| 223 |
-
"loss": 4.2539,
|
| 224 |
-
"step": 18000
|
| 225 |
-
},
|
| 226 |
-
{
|
| 227 |
-
"epoch": 1.48,
|
| 228 |
-
"learning_rate": 1.0133333333333335e-05,
|
| 229 |
-
"loss": 4.1889,
|
| 230 |
-
"step": 18500
|
| 231 |
-
},
|
| 232 |
-
{
|
| 233 |
-
"epoch": 1.52,
|
| 234 |
-
"learning_rate": 9.866666666666668e-06,
|
| 235 |
-
"loss": 4.1985,
|
| 236 |
-
"step": 19000
|
| 237 |
-
},
|
| 238 |
-
{
|
| 239 |
-
"epoch": 1.56,
|
| 240 |
-
"learning_rate": 9.600000000000001e-06,
|
| 241 |
-
"loss": 4.2305,
|
| 242 |
-
"step": 19500
|
| 243 |
-
},
|
| 244 |
-
{
|
| 245 |
-
"epoch": 1.6,
|
| 246 |
-
"learning_rate": 9.333333333333334e-06,
|
| 247 |
-
"loss": 4.1754,
|
| 248 |
-
"step": 20000
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 1.64,
|
| 252 |
-
"learning_rate": 9.066666666666667e-06,
|
| 253 |
-
"loss": 4.1607,
|
| 254 |
-
"step": 20500
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 1.68,
|
| 258 |
-
"learning_rate": 8.8e-06,
|
| 259 |
-
"loss": 4.1668,
|
| 260 |
-
"step": 21000
|
| 261 |
-
},
|
| 262 |
-
{
|
| 263 |
-
"epoch": 1.72,
|
| 264 |
-
"learning_rate": 8.533333333333335e-06,
|
| 265 |
-
"loss": 4.1535,
|
| 266 |
-
"step": 21500
|
| 267 |
-
},
|
| 268 |
-
{
|
| 269 |
-
"epoch": 1.76,
|
| 270 |
-
"learning_rate": 8.266666666666667e-06,
|
| 271 |
-
"loss": 4.1269,
|
| 272 |
-
"step": 22000
|
| 273 |
-
},
|
| 274 |
-
{
|
| 275 |
-
"epoch": 1.8,
|
| 276 |
-
"learning_rate": 8.000000000000001e-06,
|
| 277 |
-
"loss": 4.135,
|
| 278 |
-
"step": 22500
|
| 279 |
-
},
|
| 280 |
-
{
|
| 281 |
-
"epoch": 1.84,
|
| 282 |
-
"learning_rate": 7.733333333333334e-06,
|
| 283 |
-
"loss": 4.1194,
|
| 284 |
-
"step": 23000
|
| 285 |
-
},
|
| 286 |
-
{
|
| 287 |
-
"epoch": 1.88,
|
| 288 |
-
"learning_rate": 7.4666666666666675e-06,
|
| 289 |
-
"loss": 4.1276,
|
| 290 |
-
"step": 23500
|
| 291 |
-
},
|
| 292 |
-
{
|
| 293 |
-
"epoch": 1.92,
|
| 294 |
-
"learning_rate": 7.2000000000000005e-06,
|
| 295 |
-
"loss": 4.1375,
|
| 296 |
-
"step": 24000
|
| 297 |
-
},
|
| 298 |
-
{
|
| 299 |
-
"epoch": 1.96,
|
| 300 |
-
"learning_rate": 6.9333333333333344e-06,
|
| 301 |
-
"loss": 4.1004,
|
| 302 |
-
"step": 24500
|
| 303 |
-
},
|
| 304 |
-
{
|
| 305 |
-
"epoch": 2.0,
|
| 306 |
-
"learning_rate": 6.666666666666667e-06,
|
| 307 |
-
"loss": 4.1036,
|
| 308 |
-
"step": 25000
|
| 309 |
-
},
|
| 310 |
-
{
|
| 311 |
-
"epoch": 2.04,
|
| 312 |
-
"learning_rate": 6.4000000000000006e-06,
|
| 313 |
-
"loss": 4.0862,
|
| 314 |
-
"step": 25500
|
| 315 |
-
},
|
| 316 |
-
{
|
| 317 |
-
"epoch": 2.08,
|
| 318 |
-
"learning_rate": 6.133333333333334e-06,
|
| 319 |
-
"loss": 4.0681,
|
| 320 |
-
"step": 26000
|
| 321 |
-
},
|
| 322 |
-
{
|
| 323 |
-
"epoch": 2.12,
|
| 324 |
-
"learning_rate": 5.8666666666666675e-06,
|
| 325 |
-
"loss": 4.0587,
|
| 326 |
-
"step": 26500
|
| 327 |
-
},
|
| 328 |
-
{
|
| 329 |
-
"epoch": 2.16,
|
| 330 |
-
"learning_rate": 5.600000000000001e-06,
|
| 331 |
-
"loss": 4.0543,
|
| 332 |
-
"step": 27000
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 2.2,
|
| 336 |
-
"learning_rate": 5.333333333333334e-06,
|
| 337 |
-
"loss": 4.0743,
|
| 338 |
-
"step": 27500
|
| 339 |
-
},
|
| 340 |
-
{
|
| 341 |
-
"epoch": 2.24,
|
| 342 |
-
"learning_rate": 5.0666666666666676e-06,
|
| 343 |
-
"loss": 4.0691,
|
| 344 |
-
"step": 28000
|
| 345 |
-
},
|
| 346 |
-
{
|
| 347 |
-
"epoch": 2.28,
|
| 348 |
-
"learning_rate": 4.800000000000001e-06,
|
| 349 |
-
"loss": 4.048,
|
| 350 |
-
"step": 28500
|
| 351 |
-
},
|
| 352 |
-
{
|
| 353 |
-
"epoch": 2.32,
|
| 354 |
-
"learning_rate": 4.533333333333334e-06,
|
| 355 |
-
"loss": 4.0455,
|
| 356 |
-
"step": 29000
|
| 357 |
-
},
|
| 358 |
-
{
|
| 359 |
-
"epoch": 2.36,
|
| 360 |
-
"learning_rate": 4.266666666666668e-06,
|
| 361 |
-
"loss": 4.0462,
|
| 362 |
-
"step": 29500
|
| 363 |
-
},
|
| 364 |
-
{
|
| 365 |
-
"epoch": 2.4,
|
| 366 |
-
"learning_rate": 4.000000000000001e-06,
|
| 367 |
-
"loss": 4.0445,
|
| 368 |
-
"step": 30000
|
| 369 |
-
},
|
| 370 |
-
{
|
| 371 |
-
"epoch": 2.44,
|
| 372 |
-
"learning_rate": 3.7333333333333337e-06,
|
| 373 |
-
"loss": 4.0473,
|
| 374 |
-
"step": 30500
|
| 375 |
-
},
|
| 376 |
-
{
|
| 377 |
-
"epoch": 2.48,
|
| 378 |
-
"learning_rate": 3.4666666666666672e-06,
|
| 379 |
-
"loss": 4.0302,
|
| 380 |
-
"step": 31000
|
| 381 |
-
},
|
| 382 |
-
{
|
| 383 |
-
"epoch": 2.52,
|
| 384 |
-
"learning_rate": 3.2000000000000003e-06,
|
| 385 |
-
"loss": 4.0641,
|
| 386 |
-
"step": 31500
|
| 387 |
-
},
|
| 388 |
-
{
|
| 389 |
-
"epoch": 2.56,
|
| 390 |
-
"learning_rate": 2.9333333333333338e-06,
|
| 391 |
-
"loss": 4.0391,
|
| 392 |
-
"step": 32000
|
| 393 |
-
},
|
| 394 |
-
{
|
| 395 |
-
"epoch": 2.6,
|
| 396 |
-
"learning_rate": 2.666666666666667e-06,
|
| 397 |
-
"loss": 4.0097,
|
| 398 |
-
"step": 32500
|
| 399 |
-
},
|
| 400 |
-
{
|
| 401 |
-
"epoch": 2.64,
|
| 402 |
-
"learning_rate": 2.4000000000000003e-06,
|
| 403 |
-
"loss": 4.0379,
|
| 404 |
-
"step": 33000
|
| 405 |
-
}
|
| 406 |
-
],
|
| 407 |
-
"max_steps": 37500,
|
| 408 |
-
"num_train_epochs": 3,
|
| 409 |
-
"total_flos": 3.663240647442432e+16,
|
| 410 |
-
"trial_name": null,
|
| 411 |
-
"trial_params": null
|
| 412 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/training_args.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:ad52c16538319a22a2678757517656a5b363f33d0a783ef91256d9af1c175a3f
|
| 3 |
-
size 3643
|
|
|
|
|
|
|
|
|
|
|
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1200772485
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b577154032eb078377be41ead41bc92034c6f36100e61033337adce202fa5c87
|
| 3 |
size 1200772485
|
runs/Jan26_09-04-05_e112ce585155/events.out.tfevents.1674723897.e112ce585155.6073.0
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 sha256:27ef48880cd80a699f9589646bd87d4f797011cc3778dbac18c89fb679ef8009
|
| 3 |
+
size 16333
|