Model save
Browse files- last-checkpoint/config.json +0 -27
- last-checkpoint/merges.txt +0 -0
- 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 -15
- last-checkpoint/tokenizer.json +0 -0
- last-checkpoint/tokenizer_config.json +0 -17
- last-checkpoint/trainer_state.json +0 -982
- last-checkpoint/training_args.bin +0 -3
- last-checkpoint/vocab.json +0 -0
- pytorch_model.bin +1 -1
- runs/Feb06_10-35-49_turing/events.out.tfevents.1675676161.turing.964098.1 +2 -2
last-checkpoint/config.json
DELETED
|
@@ -1,27 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"_name_or_path": "/home/pcjf/CESGA/works/lmodels/models/large",
|
| 3 |
-
"architectures": [
|
| 4 |
-
"RobertaForMaskedLM"
|
| 5 |
-
],
|
| 6 |
-
"attention_probs_dropout_prob": 0.1,
|
| 7 |
-
"bos_token_id": 0,
|
| 8 |
-
"classifier_dropout": null,
|
| 9 |
-
"eos_token_id": 2,
|
| 10 |
-
"hidden_act": "gelu",
|
| 11 |
-
"hidden_dropout_prob": 0.1,
|
| 12 |
-
"hidden_size": 768,
|
| 13 |
-
"initializer_range": 0.02,
|
| 14 |
-
"intermediate_size": 3072,
|
| 15 |
-
"layer_norm_eps": 1e-05,
|
| 16 |
-
"max_position_embeddings": 514,
|
| 17 |
-
"model_type": "roberta",
|
| 18 |
-
"num_attention_heads": 12,
|
| 19 |
-
"num_hidden_layers": 12,
|
| 20 |
-
"pad_token_id": 1,
|
| 21 |
-
"position_embedding_type": "absolute",
|
| 22 |
-
"torch_dtype": "float32",
|
| 23 |
-
"transformers_version": "4.24.0",
|
| 24 |
-
"type_vocab_size": 1,
|
| 25 |
-
"use_cache": true,
|
| 26 |
-
"vocab_size": 50265
|
| 27 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/merges.txt
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
last-checkpoint/optimizer.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:cd68ef232e3e68b2a523145fb4aab44e1178593f4ecebfe4a7fb7c2a61d39159
|
| 3 |
-
size 997747845
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/pytorch_model.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:9c293560287e12e9f8d6f988947bc53ce911420ce65ca91ba9813acb4185c488
|
| 3 |
-
size 498863417
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/rng_state.pth
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:a06bfe35557371cd4030124443e6805028c2a839df314289636cf0caa8997b79
|
| 3 |
-
size 14575
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/scheduler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:0ff1be153872ceab362bc8f896bf3f611b155e54edf151eccfc448653a32209d
|
| 3 |
-
size 627
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/special_tokens_map.json
DELETED
|
@@ -1,15 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"bos_token": "<s>",
|
| 3 |
-
"cls_token": "<s>",
|
| 4 |
-
"eos_token": "</s>",
|
| 5 |
-
"mask_token": {
|
| 6 |
-
"content": "<mask>",
|
| 7 |
-
"lstrip": true,
|
| 8 |
-
"normalized": false,
|
| 9 |
-
"rstrip": false,
|
| 10 |
-
"single_word": false
|
| 11 |
-
},
|
| 12 |
-
"pad_token": "<pad>",
|
| 13 |
-
"sep_token": "</s>",
|
| 14 |
-
"unk_token": "<unk>"
|
| 15 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/tokenizer.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
last-checkpoint/tokenizer_config.json
DELETED
|
@@ -1,17 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"add_prefix_space": false,
|
| 3 |
-
"bos_token": "<s>",
|
| 4 |
-
"cls_token": "<s>",
|
| 5 |
-
"eos_token": "</s>",
|
| 6 |
-
"errors": "replace",
|
| 7 |
-
"mask_token": "<mask>",
|
| 8 |
-
"max_len": 512,
|
| 9 |
-
"model_max_length": 512,
|
| 10 |
-
"name_or_path": "/home/pcjf/CESGA/works/lmodels/models/large",
|
| 11 |
-
"pad_token": "<pad>",
|
| 12 |
-
"sep_token": "</s>",
|
| 13 |
-
"special_tokens_map_file": null,
|
| 14 |
-
"tokenizer_class": "RobertaTokenizer",
|
| 15 |
-
"trim_offsets": true,
|
| 16 |
-
"unk_token": "<unk>"
|
| 17 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/trainer_state.json
DELETED
|
@@ -1,982 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_metric": 1.0537773370742798,
|
| 3 |
-
"best_model_checkpoint": "/home/pcjf/CESGA/works/lmodels/models/large/checkpoint-102000",
|
| 4 |
-
"epoch": 14.902730598086016,
|
| 5 |
-
"global_step": 103500,
|
| 6 |
-
"is_hyper_param_search": false,
|
| 7 |
-
"is_local_process_zero": true,
|
| 8 |
-
"is_world_process_zero": true,
|
| 9 |
-
"log_history": [
|
| 10 |
-
{
|
| 11 |
-
"epoch": 0.22,
|
| 12 |
-
"learning_rate": 9.856011519078475e-06,
|
| 13 |
-
"loss": 3.6976,
|
| 14 |
-
"step": 1500
|
| 15 |
-
},
|
| 16 |
-
{
|
| 17 |
-
"epoch": 0.22,
|
| 18 |
-
"eval_loss": 2.2865896224975586,
|
| 19 |
-
"eval_runtime": 84.2237,
|
| 20 |
-
"eval_samples_per_second": 215.45,
|
| 21 |
-
"eval_steps_per_second": 26.94,
|
| 22 |
-
"step": 1500
|
| 23 |
-
},
|
| 24 |
-
{
|
| 25 |
-
"epoch": 0.43,
|
| 26 |
-
"learning_rate": 9.712023038156948e-06,
|
| 27 |
-
"loss": 2.3057,
|
| 28 |
-
"step": 3000
|
| 29 |
-
},
|
| 30 |
-
{
|
| 31 |
-
"epoch": 0.43,
|
| 32 |
-
"eval_loss": 1.9275821447372437,
|
| 33 |
-
"eval_runtime": 83.9089,
|
| 34 |
-
"eval_samples_per_second": 216.258,
|
| 35 |
-
"eval_steps_per_second": 27.041,
|
| 36 |
-
"step": 3000
|
| 37 |
-
},
|
| 38 |
-
{
|
| 39 |
-
"epoch": 0.65,
|
| 40 |
-
"learning_rate": 9.568034557235422e-06,
|
| 41 |
-
"loss": 2.0428,
|
| 42 |
-
"step": 4500
|
| 43 |
-
},
|
| 44 |
-
{
|
| 45 |
-
"epoch": 0.65,
|
| 46 |
-
"eval_loss": 1.7688409090042114,
|
| 47 |
-
"eval_runtime": 83.2995,
|
| 48 |
-
"eval_samples_per_second": 217.841,
|
| 49 |
-
"eval_steps_per_second": 27.239,
|
| 50 |
-
"step": 4500
|
| 51 |
-
},
|
| 52 |
-
{
|
| 53 |
-
"epoch": 0.86,
|
| 54 |
-
"learning_rate": 9.424046076313895e-06,
|
| 55 |
-
"loss": 1.8963,
|
| 56 |
-
"step": 6000
|
| 57 |
-
},
|
| 58 |
-
{
|
| 59 |
-
"epoch": 0.86,
|
| 60 |
-
"eval_loss": 1.651775598526001,
|
| 61 |
-
"eval_runtime": 84.0186,
|
| 62 |
-
"eval_samples_per_second": 215.976,
|
| 63 |
-
"eval_steps_per_second": 27.006,
|
| 64 |
-
"step": 6000
|
| 65 |
-
},
|
| 66 |
-
{
|
| 67 |
-
"epoch": 1.08,
|
| 68 |
-
"learning_rate": 9.28005759539237e-06,
|
| 69 |
-
"loss": 1.8031,
|
| 70 |
-
"step": 7500
|
| 71 |
-
},
|
| 72 |
-
{
|
| 73 |
-
"epoch": 1.08,
|
| 74 |
-
"eval_loss": 1.593921184539795,
|
| 75 |
-
"eval_runtime": 84.6892,
|
| 76 |
-
"eval_samples_per_second": 214.266,
|
| 77 |
-
"eval_steps_per_second": 26.792,
|
| 78 |
-
"step": 7500
|
| 79 |
-
},
|
| 80 |
-
{
|
| 81 |
-
"epoch": 1.3,
|
| 82 |
-
"learning_rate": 9.136069114470844e-06,
|
| 83 |
-
"loss": 1.7288,
|
| 84 |
-
"step": 9000
|
| 85 |
-
},
|
| 86 |
-
{
|
| 87 |
-
"epoch": 1.3,
|
| 88 |
-
"eval_loss": 1.5367190837860107,
|
| 89 |
-
"eval_runtime": 89.3272,
|
| 90 |
-
"eval_samples_per_second": 203.141,
|
| 91 |
-
"eval_steps_per_second": 25.401,
|
| 92 |
-
"step": 9000
|
| 93 |
-
},
|
| 94 |
-
{
|
| 95 |
-
"epoch": 1.51,
|
| 96 |
-
"learning_rate": 8.992080633549316e-06,
|
| 97 |
-
"loss": 1.6743,
|
| 98 |
-
"step": 10500
|
| 99 |
-
},
|
| 100 |
-
{
|
| 101 |
-
"epoch": 1.51,
|
| 102 |
-
"eval_loss": 1.4779834747314453,
|
| 103 |
-
"eval_runtime": 81.597,
|
| 104 |
-
"eval_samples_per_second": 222.386,
|
| 105 |
-
"eval_steps_per_second": 27.807,
|
| 106 |
-
"step": 10500
|
| 107 |
-
},
|
| 108 |
-
{
|
| 109 |
-
"epoch": 1.73,
|
| 110 |
-
"learning_rate": 8.84809215262779e-06,
|
| 111 |
-
"loss": 1.6304,
|
| 112 |
-
"step": 12000
|
| 113 |
-
},
|
| 114 |
-
{
|
| 115 |
-
"epoch": 1.73,
|
| 116 |
-
"eval_loss": 1.4477195739746094,
|
| 117 |
-
"eval_runtime": 84.4637,
|
| 118 |
-
"eval_samples_per_second": 214.838,
|
| 119 |
-
"eval_steps_per_second": 26.864,
|
| 120 |
-
"step": 12000
|
| 121 |
-
},
|
| 122 |
-
{
|
| 123 |
-
"epoch": 1.94,
|
| 124 |
-
"learning_rate": 8.704103671706265e-06,
|
| 125 |
-
"loss": 1.5896,
|
| 126 |
-
"step": 13500
|
| 127 |
-
},
|
| 128 |
-
{
|
| 129 |
-
"epoch": 1.94,
|
| 130 |
-
"eval_loss": 1.4108401536941528,
|
| 131 |
-
"eval_runtime": 84.7175,
|
| 132 |
-
"eval_samples_per_second": 214.194,
|
| 133 |
-
"eval_steps_per_second": 26.783,
|
| 134 |
-
"step": 13500
|
| 135 |
-
},
|
| 136 |
-
{
|
| 137 |
-
"epoch": 2.16,
|
| 138 |
-
"learning_rate": 8.560115190784738e-06,
|
| 139 |
-
"loss": 1.5581,
|
| 140 |
-
"step": 15000
|
| 141 |
-
},
|
| 142 |
-
{
|
| 143 |
-
"epoch": 2.16,
|
| 144 |
-
"eval_loss": 1.3877344131469727,
|
| 145 |
-
"eval_runtime": 84.5715,
|
| 146 |
-
"eval_samples_per_second": 214.564,
|
| 147 |
-
"eval_steps_per_second": 26.829,
|
| 148 |
-
"step": 15000
|
| 149 |
-
},
|
| 150 |
-
{
|
| 151 |
-
"epoch": 2.38,
|
| 152 |
-
"learning_rate": 8.416126709863212e-06,
|
| 153 |
-
"loss": 1.5267,
|
| 154 |
-
"step": 16500
|
| 155 |
-
},
|
| 156 |
-
{
|
| 157 |
-
"epoch": 2.38,
|
| 158 |
-
"eval_loss": 1.3697640895843506,
|
| 159 |
-
"eval_runtime": 84.7025,
|
| 160 |
-
"eval_samples_per_second": 214.232,
|
| 161 |
-
"eval_steps_per_second": 26.788,
|
| 162 |
-
"step": 16500
|
| 163 |
-
},
|
| 164 |
-
{
|
| 165 |
-
"epoch": 2.59,
|
| 166 |
-
"learning_rate": 8.272138228941685e-06,
|
| 167 |
-
"loss": 1.5027,
|
| 168 |
-
"step": 18000
|
| 169 |
-
},
|
| 170 |
-
{
|
| 171 |
-
"epoch": 2.59,
|
| 172 |
-
"eval_loss": 1.3324816226959229,
|
| 173 |
-
"eval_runtime": 84.1197,
|
| 174 |
-
"eval_samples_per_second": 215.716,
|
| 175 |
-
"eval_steps_per_second": 26.973,
|
| 176 |
-
"step": 18000
|
| 177 |
-
},
|
| 178 |
-
{
|
| 179 |
-
"epoch": 2.81,
|
| 180 |
-
"learning_rate": 8.12814974802016e-06,
|
| 181 |
-
"loss": 1.4798,
|
| 182 |
-
"step": 19500
|
| 183 |
-
},
|
| 184 |
-
{
|
| 185 |
-
"epoch": 2.81,
|
| 186 |
-
"eval_loss": 1.3138675689697266,
|
| 187 |
-
"eval_runtime": 85.4337,
|
| 188 |
-
"eval_samples_per_second": 212.399,
|
| 189 |
-
"eval_steps_per_second": 26.559,
|
| 190 |
-
"step": 19500
|
| 191 |
-
},
|
| 192 |
-
{
|
| 193 |
-
"epoch": 3.02,
|
| 194 |
-
"learning_rate": 7.984161267098632e-06,
|
| 195 |
-
"loss": 1.461,
|
| 196 |
-
"step": 21000
|
| 197 |
-
},
|
| 198 |
-
{
|
| 199 |
-
"epoch": 3.02,
|
| 200 |
-
"eval_loss": 1.298140525817871,
|
| 201 |
-
"eval_runtime": 84.4026,
|
| 202 |
-
"eval_samples_per_second": 214.993,
|
| 203 |
-
"eval_steps_per_second": 26.883,
|
| 204 |
-
"step": 21000
|
| 205 |
-
},
|
| 206 |
-
{
|
| 207 |
-
"epoch": 3.24,
|
| 208 |
-
"learning_rate": 7.840172786177106e-06,
|
| 209 |
-
"loss": 1.4403,
|
| 210 |
-
"step": 22500
|
| 211 |
-
},
|
| 212 |
-
{
|
| 213 |
-
"epoch": 3.24,
|
| 214 |
-
"eval_loss": 1.2826968431472778,
|
| 215 |
-
"eval_runtime": 85.2074,
|
| 216 |
-
"eval_samples_per_second": 212.963,
|
| 217 |
-
"eval_steps_per_second": 26.629,
|
| 218 |
-
"step": 22500
|
| 219 |
-
},
|
| 220 |
-
{
|
| 221 |
-
"epoch": 3.46,
|
| 222 |
-
"learning_rate": 7.69618430525558e-06,
|
| 223 |
-
"loss": 1.4267,
|
| 224 |
-
"step": 24000
|
| 225 |
-
},
|
| 226 |
-
{
|
| 227 |
-
"epoch": 3.46,
|
| 228 |
-
"eval_loss": 1.270477294921875,
|
| 229 |
-
"eval_runtime": 84.6854,
|
| 230 |
-
"eval_samples_per_second": 214.275,
|
| 231 |
-
"eval_steps_per_second": 26.793,
|
| 232 |
-
"step": 24000
|
| 233 |
-
},
|
| 234 |
-
{
|
| 235 |
-
"epoch": 3.67,
|
| 236 |
-
"learning_rate": 7.552195824334054e-06,
|
| 237 |
-
"loss": 1.4095,
|
| 238 |
-
"step": 25500
|
| 239 |
-
},
|
| 240 |
-
{
|
| 241 |
-
"epoch": 3.67,
|
| 242 |
-
"eval_loss": 1.2709885835647583,
|
| 243 |
-
"eval_runtime": 84.9755,
|
| 244 |
-
"eval_samples_per_second": 213.544,
|
| 245 |
-
"eval_steps_per_second": 26.702,
|
| 246 |
-
"step": 25500
|
| 247 |
-
},
|
| 248 |
-
{
|
| 249 |
-
"epoch": 3.89,
|
| 250 |
-
"learning_rate": 7.408207343412528e-06,
|
| 251 |
-
"loss": 1.3988,
|
| 252 |
-
"step": 27000
|
| 253 |
-
},
|
| 254 |
-
{
|
| 255 |
-
"epoch": 3.89,
|
| 256 |
-
"eval_loss": 1.2431179285049438,
|
| 257 |
-
"eval_runtime": 84.5543,
|
| 258 |
-
"eval_samples_per_second": 214.608,
|
| 259 |
-
"eval_steps_per_second": 26.835,
|
| 260 |
-
"step": 27000
|
| 261 |
-
},
|
| 262 |
-
{
|
| 263 |
-
"epoch": 4.1,
|
| 264 |
-
"learning_rate": 7.264218862491001e-06,
|
| 265 |
-
"loss": 1.386,
|
| 266 |
-
"step": 28500
|
| 267 |
-
},
|
| 268 |
-
{
|
| 269 |
-
"epoch": 4.1,
|
| 270 |
-
"eval_loss": 1.2419956922531128,
|
| 271 |
-
"eval_runtime": 100.1186,
|
| 272 |
-
"eval_samples_per_second": 181.245,
|
| 273 |
-
"eval_steps_per_second": 22.663,
|
| 274 |
-
"step": 28500
|
| 275 |
-
},
|
| 276 |
-
{
|
| 277 |
-
"epoch": 4.32,
|
| 278 |
-
"learning_rate": 7.1202303815694755e-06,
|
| 279 |
-
"loss": 1.371,
|
| 280 |
-
"step": 30000
|
| 281 |
-
},
|
| 282 |
-
{
|
| 283 |
-
"epoch": 4.32,
|
| 284 |
-
"eval_loss": 1.221591591835022,
|
| 285 |
-
"eval_runtime": 84.502,
|
| 286 |
-
"eval_samples_per_second": 214.741,
|
| 287 |
-
"eval_steps_per_second": 26.851,
|
| 288 |
-
"step": 30000
|
| 289 |
-
},
|
| 290 |
-
{
|
| 291 |
-
"epoch": 4.54,
|
| 292 |
-
"learning_rate": 6.976241900647949e-06,
|
| 293 |
-
"loss": 1.3598,
|
| 294 |
-
"step": 31500
|
| 295 |
-
},
|
| 296 |
-
{
|
| 297 |
-
"epoch": 4.54,
|
| 298 |
-
"eval_loss": 1.2219996452331543,
|
| 299 |
-
"eval_runtime": 84.5085,
|
| 300 |
-
"eval_samples_per_second": 214.724,
|
| 301 |
-
"eval_steps_per_second": 26.849,
|
| 302 |
-
"step": 31500
|
| 303 |
-
},
|
| 304 |
-
{
|
| 305 |
-
"epoch": 4.75,
|
| 306 |
-
"learning_rate": 6.8322534197264226e-06,
|
| 307 |
-
"loss": 1.3537,
|
| 308 |
-
"step": 33000
|
| 309 |
-
},
|
| 310 |
-
{
|
| 311 |
-
"epoch": 4.75,
|
| 312 |
-
"eval_loss": 1.2087223529815674,
|
| 313 |
-
"eval_runtime": 85.2139,
|
| 314 |
-
"eval_samples_per_second": 212.946,
|
| 315 |
-
"eval_steps_per_second": 26.627,
|
| 316 |
-
"step": 33000
|
| 317 |
-
},
|
| 318 |
-
{
|
| 319 |
-
"epoch": 4.97,
|
| 320 |
-
"learning_rate": 6.688264938804896e-06,
|
| 321 |
-
"loss": 1.3435,
|
| 322 |
-
"step": 34500
|
| 323 |
-
},
|
| 324 |
-
{
|
| 325 |
-
"epoch": 4.97,
|
| 326 |
-
"eval_loss": 1.1993805170059204,
|
| 327 |
-
"eval_runtime": 84.8549,
|
| 328 |
-
"eval_samples_per_second": 213.847,
|
| 329 |
-
"eval_steps_per_second": 26.74,
|
| 330 |
-
"step": 34500
|
| 331 |
-
},
|
| 332 |
-
{
|
| 333 |
-
"epoch": 5.18,
|
| 334 |
-
"learning_rate": 6.54427645788337e-06,
|
| 335 |
-
"loss": 1.3324,
|
| 336 |
-
"step": 36000
|
| 337 |
-
},
|
| 338 |
-
{
|
| 339 |
-
"epoch": 5.18,
|
| 340 |
-
"eval_loss": 1.1966261863708496,
|
| 341 |
-
"eval_runtime": 85.1605,
|
| 342 |
-
"eval_samples_per_second": 213.08,
|
| 343 |
-
"eval_steps_per_second": 26.644,
|
| 344 |
-
"step": 36000
|
| 345 |
-
},
|
| 346 |
-
{
|
| 347 |
-
"epoch": 5.4,
|
| 348 |
-
"learning_rate": 6.400287976961843e-06,
|
| 349 |
-
"loss": 1.3247,
|
| 350 |
-
"step": 37500
|
| 351 |
-
},
|
| 352 |
-
{
|
| 353 |
-
"epoch": 5.4,
|
| 354 |
-
"eval_loss": 1.179038643836975,
|
| 355 |
-
"eval_runtime": 84.5668,
|
| 356 |
-
"eval_samples_per_second": 214.576,
|
| 357 |
-
"eval_steps_per_second": 26.831,
|
| 358 |
-
"step": 37500
|
| 359 |
-
},
|
| 360 |
-
{
|
| 361 |
-
"epoch": 5.62,
|
| 362 |
-
"learning_rate": 6.2562994960403175e-06,
|
| 363 |
-
"loss": 1.3189,
|
| 364 |
-
"step": 39000
|
| 365 |
-
},
|
| 366 |
-
{
|
| 367 |
-
"epoch": 5.62,
|
| 368 |
-
"eval_loss": 1.1733150482177734,
|
| 369 |
-
"eval_runtime": 86.5829,
|
| 370 |
-
"eval_samples_per_second": 209.579,
|
| 371 |
-
"eval_steps_per_second": 26.206,
|
| 372 |
-
"step": 39000
|
| 373 |
-
},
|
| 374 |
-
{
|
| 375 |
-
"epoch": 5.83,
|
| 376 |
-
"learning_rate": 6.112311015118791e-06,
|
| 377 |
-
"loss": 1.3118,
|
| 378 |
-
"step": 40500
|
| 379 |
-
},
|
| 380 |
-
{
|
| 381 |
-
"epoch": 5.83,
|
| 382 |
-
"eval_loss": 1.1638059616088867,
|
| 383 |
-
"eval_runtime": 85.3947,
|
| 384 |
-
"eval_samples_per_second": 212.496,
|
| 385 |
-
"eval_steps_per_second": 26.571,
|
| 386 |
-
"step": 40500
|
| 387 |
-
},
|
| 388 |
-
{
|
| 389 |
-
"epoch": 6.05,
|
| 390 |
-
"learning_rate": 5.968322534197265e-06,
|
| 391 |
-
"loss": 1.3033,
|
| 392 |
-
"step": 42000
|
| 393 |
-
},
|
| 394 |
-
{
|
| 395 |
-
"epoch": 6.05,
|
| 396 |
-
"eval_loss": 1.166013240814209,
|
| 397 |
-
"eval_runtime": 85.1093,
|
| 398 |
-
"eval_samples_per_second": 213.208,
|
| 399 |
-
"eval_steps_per_second": 26.66,
|
| 400 |
-
"step": 42000
|
| 401 |
-
},
|
| 402 |
-
{
|
| 403 |
-
"epoch": 6.26,
|
| 404 |
-
"learning_rate": 5.824334053275739e-06,
|
| 405 |
-
"loss": 1.2962,
|
| 406 |
-
"step": 43500
|
| 407 |
-
},
|
| 408 |
-
{
|
| 409 |
-
"epoch": 6.26,
|
| 410 |
-
"eval_loss": 1.1626156568527222,
|
| 411 |
-
"eval_runtime": 84.6549,
|
| 412 |
-
"eval_samples_per_second": 214.353,
|
| 413 |
-
"eval_steps_per_second": 26.803,
|
| 414 |
-
"step": 43500
|
| 415 |
-
},
|
| 416 |
-
{
|
| 417 |
-
"epoch": 6.48,
|
| 418 |
-
"learning_rate": 5.6803455723542124e-06,
|
| 419 |
-
"loss": 1.2939,
|
| 420 |
-
"step": 45000
|
| 421 |
-
},
|
| 422 |
-
{
|
| 423 |
-
"epoch": 6.48,
|
| 424 |
-
"eval_loss": 1.1594172716140747,
|
| 425 |
-
"eval_runtime": 85.287,
|
| 426 |
-
"eval_samples_per_second": 212.764,
|
| 427 |
-
"eval_steps_per_second": 26.604,
|
| 428 |
-
"step": 45000
|
| 429 |
-
},
|
| 430 |
-
{
|
| 431 |
-
"epoch": 6.7,
|
| 432 |
-
"learning_rate": 5.536357091432686e-06,
|
| 433 |
-
"loss": 1.2871,
|
| 434 |
-
"step": 46500
|
| 435 |
-
},
|
| 436 |
-
{
|
| 437 |
-
"epoch": 6.7,
|
| 438 |
-
"eval_loss": 1.1451094150543213,
|
| 439 |
-
"eval_runtime": 84.769,
|
| 440 |
-
"eval_samples_per_second": 214.064,
|
| 441 |
-
"eval_steps_per_second": 26.767,
|
| 442 |
-
"step": 46500
|
| 443 |
-
},
|
| 444 |
-
{
|
| 445 |
-
"epoch": 6.91,
|
| 446 |
-
"learning_rate": 5.3923686105111595e-06,
|
| 447 |
-
"loss": 1.2801,
|
| 448 |
-
"step": 48000
|
| 449 |
-
},
|
| 450 |
-
{
|
| 451 |
-
"epoch": 6.91,
|
| 452 |
-
"eval_loss": 1.1349542140960693,
|
| 453 |
-
"eval_runtime": 85.0192,
|
| 454 |
-
"eval_samples_per_second": 213.434,
|
| 455 |
-
"eval_steps_per_second": 26.688,
|
| 456 |
-
"step": 48000
|
| 457 |
-
},
|
| 458 |
-
{
|
| 459 |
-
"epoch": 7.13,
|
| 460 |
-
"learning_rate": 5.248380129589633e-06,
|
| 461 |
-
"loss": 1.2732,
|
| 462 |
-
"step": 49500
|
| 463 |
-
},
|
| 464 |
-
{
|
| 465 |
-
"epoch": 7.13,
|
| 466 |
-
"eval_loss": 1.1414066553115845,
|
| 467 |
-
"eval_runtime": 83.7422,
|
| 468 |
-
"eval_samples_per_second": 216.689,
|
| 469 |
-
"eval_steps_per_second": 27.095,
|
| 470 |
-
"step": 49500
|
| 471 |
-
},
|
| 472 |
-
{
|
| 473 |
-
"epoch": 7.34,
|
| 474 |
-
"learning_rate": 5.1043916486681065e-06,
|
| 475 |
-
"loss": 1.272,
|
| 476 |
-
"step": 51000
|
| 477 |
-
},
|
| 478 |
-
{
|
| 479 |
-
"epoch": 7.34,
|
| 480 |
-
"eval_loss": 1.1416987180709839,
|
| 481 |
-
"eval_runtime": 85.9362,
|
| 482 |
-
"eval_samples_per_second": 211.157,
|
| 483 |
-
"eval_steps_per_second": 26.403,
|
| 484 |
-
"step": 51000
|
| 485 |
-
},
|
| 486 |
-
{
|
| 487 |
-
"epoch": 7.56,
|
| 488 |
-
"learning_rate": 4.960403167746581e-06,
|
| 489 |
-
"loss": 1.2652,
|
| 490 |
-
"step": 52500
|
| 491 |
-
},
|
| 492 |
-
{
|
| 493 |
-
"epoch": 7.56,
|
| 494 |
-
"eval_loss": 1.1306627988815308,
|
| 495 |
-
"eval_runtime": 84.746,
|
| 496 |
-
"eval_samples_per_second": 214.122,
|
| 497 |
-
"eval_steps_per_second": 26.774,
|
| 498 |
-
"step": 52500
|
| 499 |
-
},
|
| 500 |
-
{
|
| 501 |
-
"epoch": 7.78,
|
| 502 |
-
"learning_rate": 4.8164146868250544e-06,
|
| 503 |
-
"loss": 1.2602,
|
| 504 |
-
"step": 54000
|
| 505 |
-
},
|
| 506 |
-
{
|
| 507 |
-
"epoch": 7.78,
|
| 508 |
-
"eval_loss": 1.1231367588043213,
|
| 509 |
-
"eval_runtime": 84.9153,
|
| 510 |
-
"eval_samples_per_second": 213.695,
|
| 511 |
-
"eval_steps_per_second": 26.721,
|
| 512 |
-
"step": 54000
|
| 513 |
-
},
|
| 514 |
-
{
|
| 515 |
-
"epoch": 7.99,
|
| 516 |
-
"learning_rate": 4.672426205903528e-06,
|
| 517 |
-
"loss": 1.2589,
|
| 518 |
-
"step": 55500
|
| 519 |
-
},
|
| 520 |
-
{
|
| 521 |
-
"epoch": 7.99,
|
| 522 |
-
"eval_loss": 1.116618275642395,
|
| 523 |
-
"eval_runtime": 85.1024,
|
| 524 |
-
"eval_samples_per_second": 213.226,
|
| 525 |
-
"eval_steps_per_second": 26.662,
|
| 526 |
-
"step": 55500
|
| 527 |
-
},
|
| 528 |
-
{
|
| 529 |
-
"epoch": 8.21,
|
| 530 |
-
"learning_rate": 4.5284377249820015e-06,
|
| 531 |
-
"loss": 1.2515,
|
| 532 |
-
"step": 57000
|
| 533 |
-
},
|
| 534 |
-
{
|
| 535 |
-
"epoch": 8.21,
|
| 536 |
-
"eval_loss": 1.1146851778030396,
|
| 537 |
-
"eval_runtime": 86.1091,
|
| 538 |
-
"eval_samples_per_second": 210.733,
|
| 539 |
-
"eval_steps_per_second": 26.35,
|
| 540 |
-
"step": 57000
|
| 541 |
-
},
|
| 542 |
-
{
|
| 543 |
-
"epoch": 8.42,
|
| 544 |
-
"learning_rate": 4.384449244060476e-06,
|
| 545 |
-
"loss": 1.2508,
|
| 546 |
-
"step": 58500
|
| 547 |
-
},
|
| 548 |
-
{
|
| 549 |
-
"epoch": 8.42,
|
| 550 |
-
"eval_loss": 1.104642391204834,
|
| 551 |
-
"eval_runtime": 86.8395,
|
| 552 |
-
"eval_samples_per_second": 208.96,
|
| 553 |
-
"eval_steps_per_second": 26.129,
|
| 554 |
-
"step": 58500
|
| 555 |
-
},
|
| 556 |
-
{
|
| 557 |
-
"epoch": 8.64,
|
| 558 |
-
"learning_rate": 4.240460763138949e-06,
|
| 559 |
-
"loss": 1.2483,
|
| 560 |
-
"step": 60000
|
| 561 |
-
},
|
| 562 |
-
{
|
| 563 |
-
"epoch": 8.64,
|
| 564 |
-
"eval_loss": 1.1124722957611084,
|
| 565 |
-
"eval_runtime": 92.3126,
|
| 566 |
-
"eval_samples_per_second": 196.571,
|
| 567 |
-
"eval_steps_per_second": 24.58,
|
| 568 |
-
"step": 60000
|
| 569 |
-
},
|
| 570 |
-
{
|
| 571 |
-
"epoch": 8.86,
|
| 572 |
-
"learning_rate": 4.096472282217423e-06,
|
| 573 |
-
"loss": 1.2439,
|
| 574 |
-
"step": 61500
|
| 575 |
-
},
|
| 576 |
-
{
|
| 577 |
-
"epoch": 8.86,
|
| 578 |
-
"eval_loss": 1.1115002632141113,
|
| 579 |
-
"eval_runtime": 85.7598,
|
| 580 |
-
"eval_samples_per_second": 211.591,
|
| 581 |
-
"eval_steps_per_second": 26.458,
|
| 582 |
-
"step": 61500
|
| 583 |
-
},
|
| 584 |
-
{
|
| 585 |
-
"epoch": 9.07,
|
| 586 |
-
"learning_rate": 3.952483801295896e-06,
|
| 587 |
-
"loss": 1.2393,
|
| 588 |
-
"step": 63000
|
| 589 |
-
},
|
| 590 |
-
{
|
| 591 |
-
"epoch": 9.07,
|
| 592 |
-
"eval_loss": 1.0992404222488403,
|
| 593 |
-
"eval_runtime": 85.8343,
|
| 594 |
-
"eval_samples_per_second": 211.407,
|
| 595 |
-
"eval_steps_per_second": 26.435,
|
| 596 |
-
"step": 63000
|
| 597 |
-
},
|
| 598 |
-
{
|
| 599 |
-
"epoch": 9.29,
|
| 600 |
-
"learning_rate": 3.8084953203743704e-06,
|
| 601 |
-
"loss": 1.2346,
|
| 602 |
-
"step": 64500
|
| 603 |
-
},
|
| 604 |
-
{
|
| 605 |
-
"epoch": 9.29,
|
| 606 |
-
"eval_loss": 1.087247610092163,
|
| 607 |
-
"eval_runtime": 85.1097,
|
| 608 |
-
"eval_samples_per_second": 213.207,
|
| 609 |
-
"eval_steps_per_second": 26.66,
|
| 610 |
-
"step": 64500
|
| 611 |
-
},
|
| 612 |
-
{
|
| 613 |
-
"epoch": 9.5,
|
| 614 |
-
"learning_rate": 3.664506839452844e-06,
|
| 615 |
-
"loss": 1.2319,
|
| 616 |
-
"step": 66000
|
| 617 |
-
},
|
| 618 |
-
{
|
| 619 |
-
"epoch": 9.5,
|
| 620 |
-
"eval_loss": 1.1074174642562866,
|
| 621 |
-
"eval_runtime": 94.0488,
|
| 622 |
-
"eval_samples_per_second": 192.942,
|
| 623 |
-
"eval_steps_per_second": 24.126,
|
| 624 |
-
"step": 66000
|
| 625 |
-
},
|
| 626 |
-
{
|
| 627 |
-
"epoch": 9.72,
|
| 628 |
-
"learning_rate": 3.520518358531318e-06,
|
| 629 |
-
"loss": 1.2275,
|
| 630 |
-
"step": 67500
|
| 631 |
-
},
|
| 632 |
-
{
|
| 633 |
-
"epoch": 9.72,
|
| 634 |
-
"eval_loss": 1.0995101928710938,
|
| 635 |
-
"eval_runtime": 86.1197,
|
| 636 |
-
"eval_samples_per_second": 210.707,
|
| 637 |
-
"eval_steps_per_second": 26.347,
|
| 638 |
-
"step": 67500
|
| 639 |
-
},
|
| 640 |
-
{
|
| 641 |
-
"epoch": 9.94,
|
| 642 |
-
"learning_rate": 3.3765298776097914e-06,
|
| 643 |
-
"loss": 1.2263,
|
| 644 |
-
"step": 69000
|
| 645 |
-
},
|
| 646 |
-
{
|
| 647 |
-
"epoch": 9.94,
|
| 648 |
-
"eval_loss": 1.079862117767334,
|
| 649 |
-
"eval_runtime": 86.1089,
|
| 650 |
-
"eval_samples_per_second": 210.733,
|
| 651 |
-
"eval_steps_per_second": 26.35,
|
| 652 |
-
"step": 69000
|
| 653 |
-
},
|
| 654 |
-
{
|
| 655 |
-
"epoch": 10.15,
|
| 656 |
-
"learning_rate": 3.2325413966882653e-06,
|
| 657 |
-
"loss": 1.2242,
|
| 658 |
-
"step": 70500
|
| 659 |
-
},
|
| 660 |
-
{
|
| 661 |
-
"epoch": 10.15,
|
| 662 |
-
"eval_loss": 1.090984582901001,
|
| 663 |
-
"eval_runtime": 89.7862,
|
| 664 |
-
"eval_samples_per_second": 202.102,
|
| 665 |
-
"eval_steps_per_second": 25.271,
|
| 666 |
-
"step": 70500
|
| 667 |
-
},
|
| 668 |
-
{
|
| 669 |
-
"epoch": 10.37,
|
| 670 |
-
"learning_rate": 3.088552915766739e-06,
|
| 671 |
-
"loss": 1.2189,
|
| 672 |
-
"step": 72000
|
| 673 |
-
},
|
| 674 |
-
{
|
| 675 |
-
"epoch": 10.37,
|
| 676 |
-
"eval_loss": 1.0839877128601074,
|
| 677 |
-
"eval_runtime": 84.5722,
|
| 678 |
-
"eval_samples_per_second": 214.562,
|
| 679 |
-
"eval_steps_per_second": 26.829,
|
| 680 |
-
"step": 72000
|
| 681 |
-
},
|
| 682 |
-
{
|
| 683 |
-
"epoch": 10.58,
|
| 684 |
-
"learning_rate": 2.9445644348452123e-06,
|
| 685 |
-
"loss": 1.2175,
|
| 686 |
-
"step": 73500
|
| 687 |
-
},
|
| 688 |
-
{
|
| 689 |
-
"epoch": 10.58,
|
| 690 |
-
"eval_loss": 1.0865727663040161,
|
| 691 |
-
"eval_runtime": 84.3078,
|
| 692 |
-
"eval_samples_per_second": 215.235,
|
| 693 |
-
"eval_steps_per_second": 26.913,
|
| 694 |
-
"step": 73500
|
| 695 |
-
},
|
| 696 |
-
{
|
| 697 |
-
"epoch": 10.8,
|
| 698 |
-
"learning_rate": 2.8005759539236867e-06,
|
| 699 |
-
"loss": 1.218,
|
| 700 |
-
"step": 75000
|
| 701 |
-
},
|
| 702 |
-
{
|
| 703 |
-
"epoch": 10.8,
|
| 704 |
-
"eval_loss": 1.0715888738632202,
|
| 705 |
-
"eval_runtime": 84.6924,
|
| 706 |
-
"eval_samples_per_second": 214.258,
|
| 707 |
-
"eval_steps_per_second": 26.791,
|
| 708 |
-
"step": 75000
|
| 709 |
-
},
|
| 710 |
-
{
|
| 711 |
-
"epoch": 11.02,
|
| 712 |
-
"learning_rate": 2.6565874730021602e-06,
|
| 713 |
-
"loss": 1.2159,
|
| 714 |
-
"step": 76500
|
| 715 |
-
},
|
| 716 |
-
{
|
| 717 |
-
"epoch": 11.02,
|
| 718 |
-
"eval_loss": 1.0899019241333008,
|
| 719 |
-
"eval_runtime": 85.705,
|
| 720 |
-
"eval_samples_per_second": 211.726,
|
| 721 |
-
"eval_steps_per_second": 26.475,
|
| 722 |
-
"step": 76500
|
| 723 |
-
},
|
| 724 |
-
{
|
| 725 |
-
"epoch": 11.23,
|
| 726 |
-
"learning_rate": 2.5125989920806338e-06,
|
| 727 |
-
"loss": 1.2153,
|
| 728 |
-
"step": 78000
|
| 729 |
-
},
|
| 730 |
-
{
|
| 731 |
-
"epoch": 11.23,
|
| 732 |
-
"eval_loss": 1.087203860282898,
|
| 733 |
-
"eval_runtime": 84.3785,
|
| 734 |
-
"eval_samples_per_second": 215.055,
|
| 735 |
-
"eval_steps_per_second": 26.891,
|
| 736 |
-
"step": 78000
|
| 737 |
-
},
|
| 738 |
-
{
|
| 739 |
-
"epoch": 11.45,
|
| 740 |
-
"learning_rate": 2.3686105111591073e-06,
|
| 741 |
-
"loss": 1.2129,
|
| 742 |
-
"step": 79500
|
| 743 |
-
},
|
| 744 |
-
{
|
| 745 |
-
"epoch": 11.45,
|
| 746 |
-
"eval_loss": 1.0858579874038696,
|
| 747 |
-
"eval_runtime": 95.1152,
|
| 748 |
-
"eval_samples_per_second": 190.779,
|
| 749 |
-
"eval_steps_per_second": 23.855,
|
| 750 |
-
"step": 79500
|
| 751 |
-
},
|
| 752 |
-
{
|
| 753 |
-
"epoch": 11.66,
|
| 754 |
-
"learning_rate": 2.2246220302375812e-06,
|
| 755 |
-
"loss": 1.2105,
|
| 756 |
-
"step": 81000
|
| 757 |
-
},
|
| 758 |
-
{
|
| 759 |
-
"epoch": 11.66,
|
| 760 |
-
"eval_loss": 1.0726720094680786,
|
| 761 |
-
"eval_runtime": 86.0753,
|
| 762 |
-
"eval_samples_per_second": 210.815,
|
| 763 |
-
"eval_steps_per_second": 26.361,
|
| 764 |
-
"step": 81000
|
| 765 |
-
},
|
| 766 |
-
{
|
| 767 |
-
"epoch": 11.88,
|
| 768 |
-
"learning_rate": 2.0806335493160548e-06,
|
| 769 |
-
"loss": 1.2081,
|
| 770 |
-
"step": 82500
|
| 771 |
-
},
|
| 772 |
-
{
|
| 773 |
-
"epoch": 11.88,
|
| 774 |
-
"eval_loss": 1.0759787559509277,
|
| 775 |
-
"eval_runtime": 85.7141,
|
| 776 |
-
"eval_samples_per_second": 211.704,
|
| 777 |
-
"eval_steps_per_second": 26.472,
|
| 778 |
-
"step": 82500
|
| 779 |
-
},
|
| 780 |
-
{
|
| 781 |
-
"epoch": 12.1,
|
| 782 |
-
"learning_rate": 1.9366450683945287e-06,
|
| 783 |
-
"loss": 1.2107,
|
| 784 |
-
"step": 84000
|
| 785 |
-
},
|
| 786 |
-
{
|
| 787 |
-
"epoch": 12.1,
|
| 788 |
-
"eval_loss": 1.0708467960357666,
|
| 789 |
-
"eval_runtime": 84.1645,
|
| 790 |
-
"eval_samples_per_second": 215.602,
|
| 791 |
-
"eval_steps_per_second": 26.959,
|
| 792 |
-
"step": 84000
|
| 793 |
-
},
|
| 794 |
-
{
|
| 795 |
-
"epoch": 12.31,
|
| 796 |
-
"learning_rate": 1.7926565874730022e-06,
|
| 797 |
-
"loss": 1.2033,
|
| 798 |
-
"step": 85500
|
| 799 |
-
},
|
| 800 |
-
{
|
| 801 |
-
"epoch": 12.31,
|
| 802 |
-
"eval_loss": 1.072534441947937,
|
| 803 |
-
"eval_runtime": 85.0352,
|
| 804 |
-
"eval_samples_per_second": 213.394,
|
| 805 |
-
"eval_steps_per_second": 26.683,
|
| 806 |
-
"step": 85500
|
| 807 |
-
},
|
| 808 |
-
{
|
| 809 |
-
"epoch": 12.53,
|
| 810 |
-
"learning_rate": 1.648668106551476e-06,
|
| 811 |
-
"loss": 1.2033,
|
| 812 |
-
"step": 87000
|
| 813 |
-
},
|
| 814 |
-
{
|
| 815 |
-
"epoch": 12.53,
|
| 816 |
-
"eval_loss": 1.0773580074310303,
|
| 817 |
-
"eval_runtime": 94.5251,
|
| 818 |
-
"eval_samples_per_second": 191.97,
|
| 819 |
-
"eval_steps_per_second": 24.004,
|
| 820 |
-
"step": 87000
|
| 821 |
-
},
|
| 822 |
-
{
|
| 823 |
-
"epoch": 12.74,
|
| 824 |
-
"learning_rate": 1.5046796256299497e-06,
|
| 825 |
-
"loss": 1.2052,
|
| 826 |
-
"step": 88500
|
| 827 |
-
},
|
| 828 |
-
{
|
| 829 |
-
"epoch": 12.74,
|
| 830 |
-
"eval_loss": 1.0760116577148438,
|
| 831 |
-
"eval_runtime": 84.4275,
|
| 832 |
-
"eval_samples_per_second": 214.93,
|
| 833 |
-
"eval_steps_per_second": 26.875,
|
| 834 |
-
"step": 88500
|
| 835 |
-
},
|
| 836 |
-
{
|
| 837 |
-
"epoch": 12.96,
|
| 838 |
-
"learning_rate": 1.3606911447084234e-06,
|
| 839 |
-
"loss": 1.2012,
|
| 840 |
-
"step": 90000
|
| 841 |
-
},
|
| 842 |
-
{
|
| 843 |
-
"epoch": 12.96,
|
| 844 |
-
"eval_loss": 1.0765339136123657,
|
| 845 |
-
"eval_runtime": 84.1883,
|
| 846 |
-
"eval_samples_per_second": 215.541,
|
| 847 |
-
"eval_steps_per_second": 26.951,
|
| 848 |
-
"step": 90000
|
| 849 |
-
},
|
| 850 |
-
{
|
| 851 |
-
"epoch": 13.17,
|
| 852 |
-
"learning_rate": 1.2167026637868972e-06,
|
| 853 |
-
"loss": 1.2011,
|
| 854 |
-
"step": 91500
|
| 855 |
-
},
|
| 856 |
-
{
|
| 857 |
-
"epoch": 13.17,
|
| 858 |
-
"eval_loss": 1.0625150203704834,
|
| 859 |
-
"eval_runtime": 84.0684,
|
| 860 |
-
"eval_samples_per_second": 215.848,
|
| 861 |
-
"eval_steps_per_second": 26.99,
|
| 862 |
-
"step": 91500
|
| 863 |
-
},
|
| 864 |
-
{
|
| 865 |
-
"epoch": 13.39,
|
| 866 |
-
"learning_rate": 1.072714182865371e-06,
|
| 867 |
-
"loss": 1.2015,
|
| 868 |
-
"step": 93000
|
| 869 |
-
},
|
| 870 |
-
{
|
| 871 |
-
"epoch": 13.39,
|
| 872 |
-
"eval_loss": 1.0583701133728027,
|
| 873 |
-
"eval_runtime": 84.3705,
|
| 874 |
-
"eval_samples_per_second": 215.075,
|
| 875 |
-
"eval_steps_per_second": 26.893,
|
| 876 |
-
"step": 93000
|
| 877 |
-
},
|
| 878 |
-
{
|
| 879 |
-
"epoch": 13.61,
|
| 880 |
-
"learning_rate": 9.287257019438446e-07,
|
| 881 |
-
"loss": 1.1986,
|
| 882 |
-
"step": 94500
|
| 883 |
-
},
|
| 884 |
-
{
|
| 885 |
-
"epoch": 13.61,
|
| 886 |
-
"eval_loss": 1.0719395875930786,
|
| 887 |
-
"eval_runtime": 84.3375,
|
| 888 |
-
"eval_samples_per_second": 215.159,
|
| 889 |
-
"eval_steps_per_second": 26.904,
|
| 890 |
-
"step": 94500
|
| 891 |
-
},
|
| 892 |
-
{
|
| 893 |
-
"epoch": 13.82,
|
| 894 |
-
"learning_rate": 7.847372210223183e-07,
|
| 895 |
-
"loss": 1.2012,
|
| 896 |
-
"step": 96000
|
| 897 |
-
},
|
| 898 |
-
{
|
| 899 |
-
"epoch": 13.82,
|
| 900 |
-
"eval_loss": 1.0656870603561401,
|
| 901 |
-
"eval_runtime": 84.0312,
|
| 902 |
-
"eval_samples_per_second": 215.944,
|
| 903 |
-
"eval_steps_per_second": 27.002,
|
| 904 |
-
"step": 96000
|
| 905 |
-
},
|
| 906 |
-
{
|
| 907 |
-
"epoch": 14.04,
|
| 908 |
-
"learning_rate": 6.40748740100792e-07,
|
| 909 |
-
"loss": 1.1983,
|
| 910 |
-
"step": 97500
|
| 911 |
-
},
|
| 912 |
-
{
|
| 913 |
-
"epoch": 14.04,
|
| 914 |
-
"eval_loss": 1.07068932056427,
|
| 915 |
-
"eval_runtime": 84.1653,
|
| 916 |
-
"eval_samples_per_second": 215.6,
|
| 917 |
-
"eval_steps_per_second": 26.959,
|
| 918 |
-
"step": 97500
|
| 919 |
-
},
|
| 920 |
-
{
|
| 921 |
-
"epoch": 14.25,
|
| 922 |
-
"learning_rate": 4.967602591792657e-07,
|
| 923 |
-
"loss": 1.1982,
|
| 924 |
-
"step": 99000
|
| 925 |
-
},
|
| 926 |
-
{
|
| 927 |
-
"epoch": 14.25,
|
| 928 |
-
"eval_loss": 1.0600839853286743,
|
| 929 |
-
"eval_runtime": 81.6267,
|
| 930 |
-
"eval_samples_per_second": 222.305,
|
| 931 |
-
"eval_steps_per_second": 27.797,
|
| 932 |
-
"step": 99000
|
| 933 |
-
},
|
| 934 |
-
{
|
| 935 |
-
"epoch": 14.47,
|
| 936 |
-
"learning_rate": 3.5277177825773936e-07,
|
| 937 |
-
"loss": 1.196,
|
| 938 |
-
"step": 100500
|
| 939 |
-
},
|
| 940 |
-
{
|
| 941 |
-
"epoch": 14.47,
|
| 942 |
-
"eval_loss": 1.055431842803955,
|
| 943 |
-
"eval_runtime": 84.0067,
|
| 944 |
-
"eval_samples_per_second": 216.007,
|
| 945 |
-
"eval_steps_per_second": 27.01,
|
| 946 |
-
"step": 100500
|
| 947 |
-
},
|
| 948 |
-
{
|
| 949 |
-
"epoch": 14.69,
|
| 950 |
-
"learning_rate": 2.0878329733621312e-07,
|
| 951 |
-
"loss": 1.1971,
|
| 952 |
-
"step": 102000
|
| 953 |
-
},
|
| 954 |
-
{
|
| 955 |
-
"epoch": 14.69,
|
| 956 |
-
"eval_loss": 1.0537773370742798,
|
| 957 |
-
"eval_runtime": 84.2231,
|
| 958 |
-
"eval_samples_per_second": 215.452,
|
| 959 |
-
"eval_steps_per_second": 26.94,
|
| 960 |
-
"step": 102000
|
| 961 |
-
},
|
| 962 |
-
{
|
| 963 |
-
"epoch": 14.9,
|
| 964 |
-
"learning_rate": 6.479481641468683e-08,
|
| 965 |
-
"loss": 1.1954,
|
| 966 |
-
"step": 103500
|
| 967 |
-
},
|
| 968 |
-
{
|
| 969 |
-
"epoch": 14.9,
|
| 970 |
-
"eval_loss": 1.0612763166427612,
|
| 971 |
-
"eval_runtime": 83.7888,
|
| 972 |
-
"eval_samples_per_second": 216.568,
|
| 973 |
-
"eval_steps_per_second": 27.08,
|
| 974 |
-
"step": 103500
|
| 975 |
-
}
|
| 976 |
-
],
|
| 977 |
-
"max_steps": 104175,
|
| 978 |
-
"num_train_epochs": 15,
|
| 979 |
-
"total_flos": 1.5507603575881085e+18,
|
| 980 |
-
"trial_name": null,
|
| 981 |
-
"trial_params": null
|
| 982 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/training_args.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:ef3b1845b37bbb078cf8ca8a9159180a7e5d40d1d9b5ce146385526c2fd6c125
|
| 3 |
-
size 3451
|
|
|
|
|
|
|
|
|
|
|
|
last-checkpoint/vocab.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 498863417
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2fc030a17cebe49f0bf7d940f707ed90b0474de6a2938d06804e0674cd69601
|
| 3 |
size 498863417
|
runs/Feb06_10-35-49_turing/events.out.tfevents.1675676161.turing.964098.1
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 sha256:5de672299d86b9671a5799a15a47643c3945a326870280d8528dacd442cd3b2f
|
| 3 |
+
size 34217
|