Training in progress, step 500, checkpoint
Browse files- last-checkpoint/config.json +47 -0
- last-checkpoint/model.safetensors +2 -2
- last-checkpoint/optimizer.pt +2 -2
- last-checkpoint/rng_state.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +104 -1581
- last-checkpoint/training_args.bin +1 -1
last-checkpoint/config.json
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "x2bee/KoModernBERT-base-mlm-v03-retry-ckp03",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ModernBERTSimCSE"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 50281,
|
| 9 |
+
"classifier_activation": "gelu",
|
| 10 |
+
"classifier_bias": false,
|
| 11 |
+
"classifier_dropout": 0.0,
|
| 12 |
+
"classifier_pooling": "mean",
|
| 13 |
+
"cls_token_id": 50281,
|
| 14 |
+
"decoder_bias": true,
|
| 15 |
+
"deterministic_flash_attn": false,
|
| 16 |
+
"embedding_dropout": 0.0,
|
| 17 |
+
"eos_token_id": 50282,
|
| 18 |
+
"global_attn_every_n_layers": 3,
|
| 19 |
+
"global_rope_theta": 160000.0,
|
| 20 |
+
"gradient_checkpointing": false,
|
| 21 |
+
"hidden_activation": "gelu",
|
| 22 |
+
"hidden_size": 768,
|
| 23 |
+
"initializer_cutoff_factor": 2.0,
|
| 24 |
+
"initializer_range": 0.02,
|
| 25 |
+
"intermediate_size": 1152,
|
| 26 |
+
"layer_norm_eps": 1e-05,
|
| 27 |
+
"local_attention": 128,
|
| 28 |
+
"local_rope_theta": 10000.0,
|
| 29 |
+
"max_position_embeddings": 8192,
|
| 30 |
+
"mlp_bias": false,
|
| 31 |
+
"mlp_dropout": 0.0,
|
| 32 |
+
"model_type": "modernbert",
|
| 33 |
+
"norm_bias": false,
|
| 34 |
+
"norm_eps": 1e-05,
|
| 35 |
+
"num_attention_heads": 12,
|
| 36 |
+
"num_hidden_layers": 22,
|
| 37 |
+
"pad_token_id": 50283,
|
| 38 |
+
"position_embedding_type": "absolute",
|
| 39 |
+
"reference_compile": false,
|
| 40 |
+
"repad_logits_with_grad": false,
|
| 41 |
+
"sep_token_id": 50282,
|
| 42 |
+
"sparse_pred_ignore_index": -100,
|
| 43 |
+
"sparse_prediction": false,
|
| 44 |
+
"torch_dtype": "float32",
|
| 45 |
+
"transformers_version": "4.48.0.dev0",
|
| 46 |
+
"vocab_size": 95663
|
| 47 |
+
}
|
last-checkpoint/model.safetensors
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:a902e03f78d73a65a225e2670272f5ab30fa9243753252da0d81bf6ae1ab88f8
|
| 3 |
+
size 737580392
|
last-checkpoint/optimizer.pt
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 sha256:8c2f7579a0d7be4b045d3119455fa8aacab3cc5e2ee7588dffdebcb0ee31366a
|
| 3 |
+
size 1475248442
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:835f869ea325fd6edf27b48b589309fb66641cb92b45f2fc13d1bb6e8814106c
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:228d14efa38075e5075e5f3ea1c158f27661d545dab61c548dfe15e36f9e3d44
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,1839 +1,362 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
-
"eval_steps":
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.004686035613870665,
|
| 13 |
-
"grad_norm":
|
| 14 |
"learning_rate": 4.997071227741331e-05,
|
| 15 |
-
"loss": 2.
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.00937207122774133,
|
| 20 |
-
"grad_norm": 3.
|
| 21 |
"learning_rate": 4.994142455482662e-05,
|
| 22 |
-
"loss": 1.
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.014058106841611996,
|
| 27 |
-
"grad_norm":
|
| 28 |
"learning_rate": 4.991213683223993e-05,
|
| 29 |
-
"loss":
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.01874414245548266,
|
| 34 |
-
"grad_norm": 2.
|
| 35 |
"learning_rate": 4.9882849109653237e-05,
|
| 36 |
-
"loss": 0.
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.023430178069353328,
|
| 41 |
-
"grad_norm": 2.
|
| 42 |
"learning_rate": 4.9853561387066545e-05,
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 50
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.028116213683223992,
|
| 48 |
-
"grad_norm": 2.
|
| 49 |
"learning_rate": 4.9824273664479854e-05,
|
| 50 |
-
"loss": 0.
|
| 51 |
"step": 60
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.03280224929709466,
|
| 55 |
-
"grad_norm": 2.
|
| 56 |
"learning_rate": 4.979498594189316e-05,
|
| 57 |
-
"loss": 0.
|
| 58 |
"step": 70
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.03748828491096532,
|
| 62 |
-
"grad_norm": 2.
|
| 63 |
"learning_rate": 4.9765698219306464e-05,
|
| 64 |
-
"loss": 0.
|
| 65 |
"step": 80
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.04217432052483599,
|
| 69 |
-
"grad_norm":
|
| 70 |
"learning_rate": 4.973641049671978e-05,
|
| 71 |
-
"loss": 0.
|
| 72 |
"step": 90
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.046860356138706656,
|
| 76 |
-
"grad_norm":
|
| 77 |
"learning_rate": 4.970712277413309e-05,
|
| 78 |
-
"loss": 0.
|
| 79 |
-
"step": 100
|
| 80 |
-
},
|
| 81 |
-
{
|
| 82 |
-
"epoch": 0.046860356138706656,
|
| 83 |
-
"eval_loss": 0.09424024820327759,
|
| 84 |
-
"eval_pearson_cosine": 0.7833384278521862,
|
| 85 |
-
"eval_pearson_dot": 0.710012200190576,
|
| 86 |
-
"eval_pearson_euclidean": 0.7816309148374074,
|
| 87 |
-
"eval_pearson_manhattan": 0.7821511170378841,
|
| 88 |
-
"eval_runtime": 5.0357,
|
| 89 |
-
"eval_samples_per_second": 297.871,
|
| 90 |
-
"eval_spearman_cosine": 0.7817326074827855,
|
| 91 |
-
"eval_spearman_dot": 0.7028556042735588,
|
| 92 |
-
"eval_spearman_euclidean": 0.7865135559191494,
|
| 93 |
-
"eval_spearman_manhattan": 0.7875187958651506,
|
| 94 |
-
"eval_steps_per_second": 18.667,
|
| 95 |
"step": 100
|
| 96 |
},
|
| 97 |
{
|
| 98 |
"epoch": 0.05154639175257732,
|
| 99 |
-
"grad_norm": 2.
|
| 100 |
"learning_rate": 4.9677835051546396e-05,
|
| 101 |
-
"loss": 0.
|
| 102 |
"step": 110
|
| 103 |
},
|
| 104 |
{
|
| 105 |
"epoch": 0.056232427366447985,
|
| 106 |
-
"grad_norm": 1.
|
| 107 |
"learning_rate": 4.9648547328959705e-05,
|
| 108 |
-
"loss": 0.
|
| 109 |
"step": 120
|
| 110 |
},
|
| 111 |
{
|
| 112 |
"epoch": 0.06091846298031865,
|
| 113 |
-
"grad_norm":
|
| 114 |
"learning_rate": 4.961925960637301e-05,
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 130
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.06560449859418932,
|
| 120 |
-
"grad_norm": 1.
|
| 121 |
"learning_rate": 4.9589971883786315e-05,
|
| 122 |
-
"loss": 0.
|
| 123 |
"step": 140
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.07029053420805999,
|
| 127 |
-
"grad_norm": 1.
|
| 128 |
"learning_rate": 4.956068416119963e-05,
|
| 129 |
-
"loss": 0.
|
| 130 |
"step": 150
|
| 131 |
},
|
| 132 |
{
|
| 133 |
"epoch": 0.07497656982193064,
|
| 134 |
-
"grad_norm":
|
| 135 |
"learning_rate": 4.953139643861294e-05,
|
| 136 |
-
"loss": 0.
|
| 137 |
"step": 160
|
| 138 |
},
|
| 139 |
{
|
| 140 |
"epoch": 0.07966260543580131,
|
| 141 |
-
"grad_norm": 2.
|
| 142 |
"learning_rate": 4.950210871602625e-05,
|
| 143 |
-
"loss": 0.
|
| 144 |
"step": 170
|
| 145 |
},
|
| 146 |
{
|
| 147 |
"epoch": 0.08434864104967198,
|
| 148 |
-
"grad_norm":
|
| 149 |
"learning_rate": 4.947282099343955e-05,
|
| 150 |
-
"loss": 0.
|
| 151 |
"step": 180
|
| 152 |
},
|
| 153 |
{
|
| 154 |
"epoch": 0.08903467666354264,
|
| 155 |
-
"grad_norm": 2.
|
| 156 |
"learning_rate": 4.944353327085286e-05,
|
| 157 |
-
"loss": 0.
|
| 158 |
"step": 190
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.09372071227741331,
|
| 162 |
-
"grad_norm": 2.
|
| 163 |
"learning_rate": 4.9414245548266166e-05,
|
| 164 |
-
"loss": 0.
|
| 165 |
-
"step": 200
|
| 166 |
-
},
|
| 167 |
-
{
|
| 168 |
-
"epoch": 0.09372071227741331,
|
| 169 |
-
"eval_loss": 0.08082367479801178,
|
| 170 |
-
"eval_pearson_cosine": 0.7856181201462533,
|
| 171 |
-
"eval_pearson_dot": 0.7399407398880271,
|
| 172 |
-
"eval_pearson_euclidean": 0.7873226689476169,
|
| 173 |
-
"eval_pearson_manhattan": 0.7886999419523848,
|
| 174 |
-
"eval_runtime": 4.9124,
|
| 175 |
-
"eval_samples_per_second": 305.349,
|
| 176 |
-
"eval_spearman_cosine": 0.7882418126384709,
|
| 177 |
-
"eval_spearman_dot": 0.7354032626087063,
|
| 178 |
-
"eval_spearman_euclidean": 0.7930010817448562,
|
| 179 |
-
"eval_spearman_manhattan": 0.7945256304098883,
|
| 180 |
-
"eval_steps_per_second": 19.135,
|
| 181 |
"step": 200
|
| 182 |
},
|
| 183 |
{
|
| 184 |
"epoch": 0.09840674789128398,
|
| 185 |
-
"grad_norm": 2.
|
| 186 |
"learning_rate": 4.938495782567948e-05,
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 210
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.10309278350515463,
|
| 192 |
-
"grad_norm": 1.
|
| 193 |
"learning_rate": 4.935567010309279e-05,
|
| 194 |
-
"loss": 0.
|
| 195 |
"step": 220
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 0.1077788191190253,
|
| 199 |
-
"grad_norm": 2.
|
| 200 |
"learning_rate": 4.932638238050609e-05,
|
| 201 |
-
"loss": 0.
|
| 202 |
"step": 230
|
| 203 |
},
|
| 204 |
{
|
| 205 |
"epoch": 0.11246485473289597,
|
| 206 |
-
"grad_norm": 1.
|
| 207 |
"learning_rate": 4.92970946579194e-05,
|
| 208 |
-
"loss": 0.
|
| 209 |
"step": 240
|
| 210 |
},
|
| 211 |
{
|
| 212 |
"epoch": 0.11715089034676664,
|
| 213 |
-
"grad_norm": 1.
|
| 214 |
"learning_rate": 4.926780693533271e-05,
|
| 215 |
-
"loss": 0.
|
| 216 |
"step": 250
|
| 217 |
},
|
| 218 |
{
|
| 219 |
"epoch": 0.1218369259606373,
|
| 220 |
-
"grad_norm": 1.
|
| 221 |
"learning_rate": 4.923851921274602e-05,
|
| 222 |
-
"loss": 0.
|
| 223 |
"step": 260
|
| 224 |
},
|
| 225 |
{
|
| 226 |
"epoch": 0.12652296157450796,
|
| 227 |
-
"grad_norm": 2.
|
| 228 |
"learning_rate": 4.920923149015933e-05,
|
| 229 |
-
"loss": 0.
|
| 230 |
"step": 270
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.13120899718837864,
|
| 234 |
-
"grad_norm": 1.
|
| 235 |
"learning_rate": 4.9179943767572635e-05,
|
| 236 |
-
"loss": 0.
|
| 237 |
"step": 280
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.1358950328022493,
|
| 241 |
-
"grad_norm": 2.
|
| 242 |
"learning_rate": 4.9150656044985943e-05,
|
| 243 |
-
"loss": 0.
|
| 244 |
"step": 290
|
| 245 |
},
|
| 246 |
{
|
| 247 |
"epoch": 0.14058106841611998,
|
| 248 |
-
"grad_norm":
|
| 249 |
"learning_rate": 4.912136832239925e-05,
|
| 250 |
-
"loss": 0.
|
| 251 |
-
"step": 300
|
| 252 |
-
},
|
| 253 |
-
{
|
| 254 |
-
"epoch": 0.14058106841611998,
|
| 255 |
-
"eval_loss": 0.06987904757261276,
|
| 256 |
-
"eval_pearson_cosine": 0.8036511544079303,
|
| 257 |
-
"eval_pearson_dot": 0.7505297006445062,
|
| 258 |
-
"eval_pearson_euclidean": 0.8010123967565619,
|
| 259 |
-
"eval_pearson_manhattan": 0.8020924463783778,
|
| 260 |
-
"eval_runtime": 6.2685,
|
| 261 |
-
"eval_samples_per_second": 239.29,
|
| 262 |
-
"eval_spearman_cosine": 0.802906284496355,
|
| 263 |
-
"eval_spearman_dot": 0.74611539957357,
|
| 264 |
-
"eval_spearman_euclidean": 0.8076860955607743,
|
| 265 |
-
"eval_spearman_manhattan": 0.8089628010364828,
|
| 266 |
-
"eval_steps_per_second": 14.996,
|
| 267 |
"step": 300
|
| 268 |
},
|
| 269 |
{
|
| 270 |
"epoch": 0.14526710402999063,
|
| 271 |
-
"grad_norm":
|
| 272 |
"learning_rate": 4.909208059981256e-05,
|
| 273 |
-
"loss": 0.
|
| 274 |
"step": 310
|
| 275 |
},
|
| 276 |
{
|
| 277 |
"epoch": 0.14995313964386128,
|
| 278 |
-
"grad_norm": 2.
|
| 279 |
"learning_rate": 4.906279287722587e-05,
|
| 280 |
-
"loss": 0.
|
| 281 |
"step": 320
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"epoch": 0.15463917525773196,
|
| 285 |
-
"grad_norm":
|
| 286 |
"learning_rate": 4.903350515463918e-05,
|
| 287 |
-
"loss": 0.
|
| 288 |
"step": 330
|
| 289 |
},
|
| 290 |
{
|
| 291 |
"epoch": 0.15932521087160262,
|
| 292 |
-
"grad_norm": 1.
|
| 293 |
"learning_rate": 4.9004217432052486e-05,
|
| 294 |
-
"loss": 0.
|
| 295 |
"step": 340
|
| 296 |
},
|
| 297 |
{
|
| 298 |
"epoch": 0.1640112464854733,
|
| 299 |
-
"grad_norm": 1.
|
| 300 |
"learning_rate": 4.8974929709465795e-05,
|
| 301 |
-
"loss": 0.
|
| 302 |
"step": 350
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.16869728209934395,
|
| 306 |
-
"grad_norm": 2.
|
| 307 |
"learning_rate": 4.89456419868791e-05,
|
| 308 |
-
"loss": 0.
|
| 309 |
"step": 360
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 0.1733833177132146,
|
| 313 |
-
"grad_norm": 2.
|
| 314 |
"learning_rate": 4.891635426429241e-05,
|
| 315 |
-
"loss": 0.
|
| 316 |
"step": 370
|
| 317 |
},
|
| 318 |
{
|
| 319 |
"epoch": 0.1780693533270853,
|
| 320 |
-
"grad_norm":
|
| 321 |
"learning_rate": 4.888706654170572e-05,
|
| 322 |
-
"loss": 0.
|
| 323 |
"step": 380
|
| 324 |
},
|
| 325 |
{
|
| 326 |
"epoch": 0.18275538894095594,
|
| 327 |
-
"grad_norm": 1.
|
| 328 |
"learning_rate": 4.885777881911903e-05,
|
| 329 |
-
"loss": 0.
|
| 330 |
"step": 390
|
| 331 |
},
|
| 332 |
{
|
| 333 |
"epoch": 0.18744142455482662,
|
| 334 |
-
"grad_norm":
|
| 335 |
"learning_rate": 4.882849109653234e-05,
|
| 336 |
-
"loss": 0.
|
| 337 |
-
"step": 400
|
| 338 |
-
},
|
| 339 |
-
{
|
| 340 |
-
"epoch": 0.18744142455482662,
|
| 341 |
-
"eval_loss": 0.06156951189041138,
|
| 342 |
-
"eval_pearson_cosine": 0.8108872920119552,
|
| 343 |
-
"eval_pearson_dot": 0.7663492619892054,
|
| 344 |
-
"eval_pearson_euclidean": 0.8079787711457271,
|
| 345 |
-
"eval_pearson_manhattan": 0.8094246591845007,
|
| 346 |
-
"eval_runtime": 6.0053,
|
| 347 |
-
"eval_samples_per_second": 249.78,
|
| 348 |
-
"eval_spearman_cosine": 0.8096398271545133,
|
| 349 |
-
"eval_spearman_dot": 0.7619990554899813,
|
| 350 |
-
"eval_spearman_euclidean": 0.8128327163681044,
|
| 351 |
-
"eval_spearman_manhattan": 0.8143278410022097,
|
| 352 |
-
"eval_steps_per_second": 15.653,
|
| 353 |
"step": 400
|
| 354 |
},
|
| 355 |
{
|
| 356 |
"epoch": 0.19212746016869728,
|
| 357 |
-
"grad_norm": 2.
|
| 358 |
"learning_rate": 4.8799203373945646e-05,
|
| 359 |
-
"loss": 0.
|
| 360 |
"step": 410
|
| 361 |
},
|
| 362 |
{
|
| 363 |
"epoch": 0.19681349578256796,
|
| 364 |
-
"grad_norm": 1.
|
| 365 |
"learning_rate": 4.8769915651358954e-05,
|
| 366 |
-
"loss": 0.
|
| 367 |
"step": 420
|
| 368 |
},
|
| 369 |
{
|
| 370 |
"epoch": 0.2014995313964386,
|
| 371 |
-
"grad_norm": 2.
|
| 372 |
"learning_rate": 4.8740627928772256e-05,
|
| 373 |
-
"loss": 0.
|
| 374 |
"step": 430
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 0.20618556701030927,
|
| 378 |
-
"grad_norm": 1.
|
| 379 |
"learning_rate": 4.871134020618557e-05,
|
| 380 |
-
"loss": 0.
|
| 381 |
"step": 440
|
| 382 |
},
|
| 383 |
{
|
| 384 |
"epoch": 0.21087160262417995,
|
| 385 |
-
"grad_norm": 2.
|
| 386 |
"learning_rate": 4.868205248359888e-05,
|
| 387 |
-
"loss": 0.
|
| 388 |
"step": 450
|
| 389 |
},
|
| 390 |
{
|
| 391 |
"epoch": 0.2155576382380506,
|
| 392 |
-
"grad_norm": 1.
|
| 393 |
"learning_rate": 4.865276476101219e-05,
|
| 394 |
-
"loss": 0.
|
| 395 |
"step": 460
|
| 396 |
},
|
| 397 |
{
|
| 398 |
"epoch": 0.22024367385192128,
|
| 399 |
-
"grad_norm": 1.
|
| 400 |
"learning_rate": 4.86234770384255e-05,
|
| 401 |
-
"loss": 0.
|
| 402 |
"step": 470
|
| 403 |
},
|
| 404 |
{
|
| 405 |
"epoch": 0.22492970946579194,
|
| 406 |
-
"grad_norm": 2.
|
| 407 |
"learning_rate": 4.85941893158388e-05,
|
| 408 |
-
"loss": 0.
|
| 409 |
"step": 480
|
| 410 |
},
|
| 411 |
{
|
| 412 |
"epoch": 0.2296157450796626,
|
| 413 |
-
"grad_norm":
|
| 414 |
"learning_rate": 4.856490159325211e-05,
|
| 415 |
-
"loss": 0.
|
| 416 |
"step": 490
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 0.23430178069353327,
|
| 420 |
-
"grad_norm": 1.
|
| 421 |
"learning_rate": 4.853561387066542e-05,
|
| 422 |
-
"loss": 0.
|
| 423 |
-
"step": 500
|
| 424 |
-
},
|
| 425 |
-
{
|
| 426 |
-
"epoch": 0.23430178069353327,
|
| 427 |
-
"eval_loss": 0.0611547976732254,
|
| 428 |
-
"eval_pearson_cosine": 0.8109234308318634,
|
| 429 |
-
"eval_pearson_dot": 0.7739396599918535,
|
| 430 |
-
"eval_pearson_euclidean": 0.8112683151419446,
|
| 431 |
-
"eval_pearson_manhattan": 0.8132717499846649,
|
| 432 |
-
"eval_runtime": 5.9704,
|
| 433 |
-
"eval_samples_per_second": 251.239,
|
| 434 |
-
"eval_spearman_cosine": 0.8134716423487094,
|
| 435 |
-
"eval_spearman_dot": 0.7705186875238829,
|
| 436 |
-
"eval_spearman_euclidean": 0.8167514840119909,
|
| 437 |
-
"eval_spearman_manhattan": 0.8187222998801444,
|
| 438 |
-
"eval_steps_per_second": 15.744,
|
| 439 |
"step": 500
|
| 440 |
-
},
|
| 441 |
-
{
|
| 442 |
-
"epoch": 0.23898781630740393,
|
| 443 |
-
"grad_norm": 1.9643880128860474,
|
| 444 |
-
"learning_rate": 4.850632614807873e-05,
|
| 445 |
-
"loss": 0.4231,
|
| 446 |
-
"step": 510
|
| 447 |
-
},
|
| 448 |
-
{
|
| 449 |
-
"epoch": 0.2436738519212746,
|
| 450 |
-
"grad_norm": 1.7972699403762817,
|
| 451 |
-
"learning_rate": 4.847703842549204e-05,
|
| 452 |
-
"loss": 0.3893,
|
| 453 |
-
"step": 520
|
| 454 |
-
},
|
| 455 |
-
{
|
| 456 |
-
"epoch": 0.24835988753514526,
|
| 457 |
-
"grad_norm": 1.6312799453735352,
|
| 458 |
-
"learning_rate": 4.844775070290534e-05,
|
| 459 |
-
"loss": 0.3869,
|
| 460 |
-
"step": 530
|
| 461 |
-
},
|
| 462 |
-
{
|
| 463 |
-
"epoch": 0.2530459231490159,
|
| 464 |
-
"grad_norm": 1.8009634017944336,
|
| 465 |
-
"learning_rate": 4.841846298031865e-05,
|
| 466 |
-
"loss": 0.4,
|
| 467 |
-
"step": 540
|
| 468 |
-
},
|
| 469 |
-
{
|
| 470 |
-
"epoch": 0.25773195876288657,
|
| 471 |
-
"grad_norm": 1.1453664302825928,
|
| 472 |
-
"learning_rate": 4.838917525773196e-05,
|
| 473 |
-
"loss": 0.3568,
|
| 474 |
-
"step": 550
|
| 475 |
-
},
|
| 476 |
-
{
|
| 477 |
-
"epoch": 0.2624179943767573,
|
| 478 |
-
"grad_norm": 1.683673620223999,
|
| 479 |
-
"learning_rate": 4.8359887535145274e-05,
|
| 480 |
-
"loss": 0.3674,
|
| 481 |
-
"step": 560
|
| 482 |
-
},
|
| 483 |
-
{
|
| 484 |
-
"epoch": 0.26710402999062793,
|
| 485 |
-
"grad_norm": 1.3695913553237915,
|
| 486 |
-
"learning_rate": 4.833059981255858e-05,
|
| 487 |
-
"loss": 0.3817,
|
| 488 |
-
"step": 570
|
| 489 |
-
},
|
| 490 |
-
{
|
| 491 |
-
"epoch": 0.2717900656044986,
|
| 492 |
-
"grad_norm": 1.8631620407104492,
|
| 493 |
-
"learning_rate": 4.8301312089971884e-05,
|
| 494 |
-
"loss": 0.3546,
|
| 495 |
-
"step": 580
|
| 496 |
-
},
|
| 497 |
-
{
|
| 498 |
-
"epoch": 0.27647610121836924,
|
| 499 |
-
"grad_norm": 1.5883185863494873,
|
| 500 |
-
"learning_rate": 4.827202436738519e-05,
|
| 501 |
-
"loss": 0.3973,
|
| 502 |
-
"step": 590
|
| 503 |
-
},
|
| 504 |
-
{
|
| 505 |
-
"epoch": 0.28116213683223995,
|
| 506 |
-
"grad_norm": 1.7056660652160645,
|
| 507 |
-
"learning_rate": 4.82427366447985e-05,
|
| 508 |
-
"loss": 0.3754,
|
| 509 |
-
"step": 600
|
| 510 |
-
},
|
| 511 |
-
{
|
| 512 |
-
"epoch": 0.28116213683223995,
|
| 513 |
-
"eval_loss": 0.059529613703489304,
|
| 514 |
-
"eval_pearson_cosine": 0.8104839357699305,
|
| 515 |
-
"eval_pearson_dot": 0.7706455959096417,
|
| 516 |
-
"eval_pearson_euclidean": 0.8087417307856555,
|
| 517 |
-
"eval_pearson_manhattan": 0.8101299227665919,
|
| 518 |
-
"eval_runtime": 6.0212,
|
| 519 |
-
"eval_samples_per_second": 249.121,
|
| 520 |
-
"eval_spearman_cosine": 0.8125251228747598,
|
| 521 |
-
"eval_spearman_dot": 0.7680727600884657,
|
| 522 |
-
"eval_spearman_euclidean": 0.8146128693278114,
|
| 523 |
-
"eval_spearman_manhattan": 0.8161621120875591,
|
| 524 |
-
"eval_steps_per_second": 15.612,
|
| 525 |
-
"step": 600
|
| 526 |
-
},
|
| 527 |
-
{
|
| 528 |
-
"epoch": 0.2858481724461106,
|
| 529 |
-
"grad_norm": 1.8564058542251587,
|
| 530 |
-
"learning_rate": 4.821344892221181e-05,
|
| 531 |
-
"loss": 0.4016,
|
| 532 |
-
"step": 610
|
| 533 |
-
},
|
| 534 |
-
{
|
| 535 |
-
"epoch": 0.29053420805998126,
|
| 536 |
-
"grad_norm": 1.467993974685669,
|
| 537 |
-
"learning_rate": 4.818416119962512e-05,
|
| 538 |
-
"loss": 0.3858,
|
| 539 |
-
"step": 620
|
| 540 |
-
},
|
| 541 |
-
{
|
| 542 |
-
"epoch": 0.2952202436738519,
|
| 543 |
-
"grad_norm": 2.3624465465545654,
|
| 544 |
-
"learning_rate": 4.815487347703843e-05,
|
| 545 |
-
"loss": 0.3796,
|
| 546 |
-
"step": 630
|
| 547 |
-
},
|
| 548 |
-
{
|
| 549 |
-
"epoch": 0.29990627928772257,
|
| 550 |
-
"grad_norm": 1.588629126548767,
|
| 551 |
-
"learning_rate": 4.8125585754451736e-05,
|
| 552 |
-
"loss": 0.3667,
|
| 553 |
-
"step": 640
|
| 554 |
-
},
|
| 555 |
-
{
|
| 556 |
-
"epoch": 0.3045923149015933,
|
| 557 |
-
"grad_norm": 1.380112886428833,
|
| 558 |
-
"learning_rate": 4.8096298031865044e-05,
|
| 559 |
-
"loss": 0.3453,
|
| 560 |
-
"step": 650
|
| 561 |
-
},
|
| 562 |
-
{
|
| 563 |
-
"epoch": 0.30927835051546393,
|
| 564 |
-
"grad_norm": 1.4270693063735962,
|
| 565 |
-
"learning_rate": 4.806701030927835e-05,
|
| 566 |
-
"loss": 0.3345,
|
| 567 |
-
"step": 660
|
| 568 |
-
},
|
| 569 |
-
{
|
| 570 |
-
"epoch": 0.3139643861293346,
|
| 571 |
-
"grad_norm": 2.204744338989258,
|
| 572 |
-
"learning_rate": 4.803772258669166e-05,
|
| 573 |
-
"loss": 0.395,
|
| 574 |
-
"step": 670
|
| 575 |
-
},
|
| 576 |
-
{
|
| 577 |
-
"epoch": 0.31865042174320524,
|
| 578 |
-
"grad_norm": 1.4480923414230347,
|
| 579 |
-
"learning_rate": 4.800843486410497e-05,
|
| 580 |
-
"loss": 0.3691,
|
| 581 |
-
"step": 680
|
| 582 |
-
},
|
| 583 |
-
{
|
| 584 |
-
"epoch": 0.3233364573570759,
|
| 585 |
-
"grad_norm": 1.8864325284957886,
|
| 586 |
-
"learning_rate": 4.797914714151828e-05,
|
| 587 |
-
"loss": 0.3986,
|
| 588 |
-
"step": 690
|
| 589 |
-
},
|
| 590 |
-
{
|
| 591 |
-
"epoch": 0.3280224929709466,
|
| 592 |
-
"grad_norm": 1.3784370422363281,
|
| 593 |
-
"learning_rate": 4.794985941893159e-05,
|
| 594 |
-
"loss": 0.3729,
|
| 595 |
-
"step": 700
|
| 596 |
-
},
|
| 597 |
-
{
|
| 598 |
-
"epoch": 0.3280224929709466,
|
| 599 |
-
"eval_loss": 0.061924997717142105,
|
| 600 |
-
"eval_pearson_cosine": 0.8155259960165324,
|
| 601 |
-
"eval_pearson_dot": 0.7761366153485074,
|
| 602 |
-
"eval_pearson_euclidean": 0.8127568794877789,
|
| 603 |
-
"eval_pearson_manhattan": 0.8144288026347226,
|
| 604 |
-
"eval_runtime": 6.0626,
|
| 605 |
-
"eval_samples_per_second": 247.42,
|
| 606 |
-
"eval_spearman_cosine": 0.8175981152530937,
|
| 607 |
-
"eval_spearman_dot": 0.7736443532595881,
|
| 608 |
-
"eval_spearman_euclidean": 0.8195662973032031,
|
| 609 |
-
"eval_spearman_manhattan": 0.8212465310439688,
|
| 610 |
-
"eval_steps_per_second": 15.505,
|
| 611 |
-
"step": 700
|
| 612 |
-
},
|
| 613 |
-
{
|
| 614 |
-
"epoch": 0.33270852858481725,
|
| 615 |
-
"grad_norm": 1.7109017372131348,
|
| 616 |
-
"learning_rate": 4.7920571696344895e-05,
|
| 617 |
-
"loss": 0.3345,
|
| 618 |
-
"step": 710
|
| 619 |
-
},
|
| 620 |
-
{
|
| 621 |
-
"epoch": 0.3373945641986879,
|
| 622 |
-
"grad_norm": 1.8547511100769043,
|
| 623 |
-
"learning_rate": 4.7891283973758204e-05,
|
| 624 |
-
"loss": 0.3735,
|
| 625 |
-
"step": 720
|
| 626 |
-
},
|
| 627 |
-
{
|
| 628 |
-
"epoch": 0.34208059981255856,
|
| 629 |
-
"grad_norm": 1.5369923114776611,
|
| 630 |
-
"learning_rate": 4.786199625117151e-05,
|
| 631 |
-
"loss": 0.3304,
|
| 632 |
-
"step": 730
|
| 633 |
-
},
|
| 634 |
-
{
|
| 635 |
-
"epoch": 0.3467666354264292,
|
| 636 |
-
"grad_norm": 1.308568000793457,
|
| 637 |
-
"learning_rate": 4.783270852858482e-05,
|
| 638 |
-
"loss": 0.3717,
|
| 639 |
-
"step": 740
|
| 640 |
-
},
|
| 641 |
-
{
|
| 642 |
-
"epoch": 0.3514526710402999,
|
| 643 |
-
"grad_norm": 1.3743574619293213,
|
| 644 |
-
"learning_rate": 4.780342080599813e-05,
|
| 645 |
-
"loss": 0.3381,
|
| 646 |
-
"step": 750
|
| 647 |
-
},
|
| 648 |
-
{
|
| 649 |
-
"epoch": 0.3561387066541706,
|
| 650 |
-
"grad_norm": 1.874657154083252,
|
| 651 |
-
"learning_rate": 4.777413308341144e-05,
|
| 652 |
-
"loss": 0.3193,
|
| 653 |
-
"step": 760
|
| 654 |
-
},
|
| 655 |
-
{
|
| 656 |
-
"epoch": 0.36082474226804123,
|
| 657 |
-
"grad_norm": 1.4700101613998413,
|
| 658 |
-
"learning_rate": 4.774484536082475e-05,
|
| 659 |
-
"loss": 0.3799,
|
| 660 |
-
"step": 770
|
| 661 |
-
},
|
| 662 |
-
{
|
| 663 |
-
"epoch": 0.3655107778819119,
|
| 664 |
-
"grad_norm": 1.5662988424301147,
|
| 665 |
-
"learning_rate": 4.771555763823805e-05,
|
| 666 |
-
"loss": 0.3453,
|
| 667 |
-
"step": 780
|
| 668 |
-
},
|
| 669 |
-
{
|
| 670 |
-
"epoch": 0.3701968134957826,
|
| 671 |
-
"grad_norm": 1.4666754007339478,
|
| 672 |
-
"learning_rate": 4.768626991565136e-05,
|
| 673 |
-
"loss": 0.3175,
|
| 674 |
-
"step": 790
|
| 675 |
-
},
|
| 676 |
-
{
|
| 677 |
-
"epoch": 0.37488284910965325,
|
| 678 |
-
"grad_norm": 1.3993242979049683,
|
| 679 |
-
"learning_rate": 4.765698219306467e-05,
|
| 680 |
-
"loss": 0.341,
|
| 681 |
-
"step": 800
|
| 682 |
-
},
|
| 683 |
-
{
|
| 684 |
-
"epoch": 0.37488284910965325,
|
| 685 |
-
"eval_loss": 0.05296875163912773,
|
| 686 |
-
"eval_pearson_cosine": 0.8137295797811834,
|
| 687 |
-
"eval_pearson_dot": 0.7695932846417932,
|
| 688 |
-
"eval_pearson_euclidean": 0.8189567419998482,
|
| 689 |
-
"eval_pearson_manhattan": 0.8200667930673546,
|
| 690 |
-
"eval_runtime": 6.3175,
|
| 691 |
-
"eval_samples_per_second": 237.436,
|
| 692 |
-
"eval_spearman_cosine": 0.815541427803139,
|
| 693 |
-
"eval_spearman_dot": 0.7663341686268886,
|
| 694 |
-
"eval_spearman_euclidean": 0.8233566840888671,
|
| 695 |
-
"eval_spearman_manhattan": 0.8246092914965037,
|
| 696 |
-
"eval_steps_per_second": 14.879,
|
| 697 |
-
"step": 800
|
| 698 |
-
},
|
| 699 |
-
{
|
| 700 |
-
"epoch": 0.3795688847235239,
|
| 701 |
-
"grad_norm": 1.4209802150726318,
|
| 702 |
-
"learning_rate": 4.762769447047798e-05,
|
| 703 |
-
"loss": 0.3831,
|
| 704 |
-
"step": 810
|
| 705 |
-
},
|
| 706 |
-
{
|
| 707 |
-
"epoch": 0.38425492033739456,
|
| 708 |
-
"grad_norm": 1.4097892045974731,
|
| 709 |
-
"learning_rate": 4.759840674789129e-05,
|
| 710 |
-
"loss": 0.3318,
|
| 711 |
-
"step": 820
|
| 712 |
-
},
|
| 713 |
-
{
|
| 714 |
-
"epoch": 0.3889409559512652,
|
| 715 |
-
"grad_norm": 1.706900715827942,
|
| 716 |
-
"learning_rate": 4.756911902530459e-05,
|
| 717 |
-
"loss": 0.3328,
|
| 718 |
-
"step": 830
|
| 719 |
-
},
|
| 720 |
-
{
|
| 721 |
-
"epoch": 0.3936269915651359,
|
| 722 |
-
"grad_norm": 1.610275149345398,
|
| 723 |
-
"learning_rate": 4.75398313027179e-05,
|
| 724 |
-
"loss": 0.3575,
|
| 725 |
-
"step": 840
|
| 726 |
-
},
|
| 727 |
-
{
|
| 728 |
-
"epoch": 0.3983130271790066,
|
| 729 |
-
"grad_norm": 1.4575105905532837,
|
| 730 |
-
"learning_rate": 4.751054358013121e-05,
|
| 731 |
-
"loss": 0.307,
|
| 732 |
-
"step": 850
|
| 733 |
-
},
|
| 734 |
-
{
|
| 735 |
-
"epoch": 0.4029990627928772,
|
| 736 |
-
"grad_norm": 1.638424277305603,
|
| 737 |
-
"learning_rate": 4.7481255857544524e-05,
|
| 738 |
-
"loss": 0.3504,
|
| 739 |
-
"step": 860
|
| 740 |
-
},
|
| 741 |
-
{
|
| 742 |
-
"epoch": 0.4076850984067479,
|
| 743 |
-
"grad_norm": 1.8157601356506348,
|
| 744 |
-
"learning_rate": 4.745196813495783e-05,
|
| 745 |
-
"loss": 0.3931,
|
| 746 |
-
"step": 870
|
| 747 |
-
},
|
| 748 |
-
{
|
| 749 |
-
"epoch": 0.41237113402061853,
|
| 750 |
-
"grad_norm": 1.6680104732513428,
|
| 751 |
-
"learning_rate": 4.7422680412371134e-05,
|
| 752 |
-
"loss": 0.362,
|
| 753 |
-
"step": 880
|
| 754 |
-
},
|
| 755 |
-
{
|
| 756 |
-
"epoch": 0.41705716963448924,
|
| 757 |
-
"grad_norm": 1.4331028461456299,
|
| 758 |
-
"learning_rate": 4.739339268978444e-05,
|
| 759 |
-
"loss": 0.3451,
|
| 760 |
-
"step": 890
|
| 761 |
-
},
|
| 762 |
-
{
|
| 763 |
-
"epoch": 0.4217432052483599,
|
| 764 |
-
"grad_norm": 1.3940101861953735,
|
| 765 |
-
"learning_rate": 4.736410496719775e-05,
|
| 766 |
-
"loss": 0.3161,
|
| 767 |
-
"step": 900
|
| 768 |
-
},
|
| 769 |
-
{
|
| 770 |
-
"epoch": 0.4217432052483599,
|
| 771 |
-
"eval_loss": 0.05680527910590172,
|
| 772 |
-
"eval_pearson_cosine": 0.816164907471336,
|
| 773 |
-
"eval_pearson_dot": 0.7659985241939467,
|
| 774 |
-
"eval_pearson_euclidean": 0.8198292531320703,
|
| 775 |
-
"eval_pearson_manhattan": 0.8209187797411488,
|
| 776 |
-
"eval_runtime": 6.6335,
|
| 777 |
-
"eval_samples_per_second": 226.126,
|
| 778 |
-
"eval_spearman_cosine": 0.8181742542924034,
|
| 779 |
-
"eval_spearman_dot": 0.7624851760530289,
|
| 780 |
-
"eval_spearman_euclidean": 0.8251528076462932,
|
| 781 |
-
"eval_spearman_manhattan": 0.8261936560831687,
|
| 782 |
-
"eval_steps_per_second": 14.171,
|
| 783 |
-
"step": 900
|
| 784 |
-
},
|
| 785 |
-
{
|
| 786 |
-
"epoch": 0.42642924086223055,
|
| 787 |
-
"grad_norm": 1.5849499702453613,
|
| 788 |
-
"learning_rate": 4.733481724461106e-05,
|
| 789 |
-
"loss": 0.2852,
|
| 790 |
-
"step": 910
|
| 791 |
-
},
|
| 792 |
-
{
|
| 793 |
-
"epoch": 0.4311152764761012,
|
| 794 |
-
"grad_norm": 1.8611364364624023,
|
| 795 |
-
"learning_rate": 4.7305529522024375e-05,
|
| 796 |
-
"loss": 0.3517,
|
| 797 |
-
"step": 920
|
| 798 |
-
},
|
| 799 |
-
{
|
| 800 |
-
"epoch": 0.43580131208997186,
|
| 801 |
-
"grad_norm": 1.759479284286499,
|
| 802 |
-
"learning_rate": 4.727624179943768e-05,
|
| 803 |
-
"loss": 0.3309,
|
| 804 |
-
"step": 930
|
| 805 |
-
},
|
| 806 |
-
{
|
| 807 |
-
"epoch": 0.44048734770384257,
|
| 808 |
-
"grad_norm": 1.3715683221817017,
|
| 809 |
-
"learning_rate": 4.7246954076850985e-05,
|
| 810 |
-
"loss": 0.2964,
|
| 811 |
-
"step": 940
|
| 812 |
-
},
|
| 813 |
-
{
|
| 814 |
-
"epoch": 0.4451733833177132,
|
| 815 |
-
"grad_norm": 1.6326545476913452,
|
| 816 |
-
"learning_rate": 4.7217666354264294e-05,
|
| 817 |
-
"loss": 0.3501,
|
| 818 |
-
"step": 950
|
| 819 |
-
},
|
| 820 |
-
{
|
| 821 |
-
"epoch": 0.4498594189315839,
|
| 822 |
-
"grad_norm": 1.238206148147583,
|
| 823 |
-
"learning_rate": 4.71883786316776e-05,
|
| 824 |
-
"loss": 0.3366,
|
| 825 |
-
"step": 960
|
| 826 |
-
},
|
| 827 |
-
{
|
| 828 |
-
"epoch": 0.45454545454545453,
|
| 829 |
-
"grad_norm": 1.6656396389007568,
|
| 830 |
-
"learning_rate": 4.715909090909091e-05,
|
| 831 |
-
"loss": 0.3594,
|
| 832 |
-
"step": 970
|
| 833 |
-
},
|
| 834 |
-
{
|
| 835 |
-
"epoch": 0.4592314901593252,
|
| 836 |
-
"grad_norm": 1.5264825820922852,
|
| 837 |
-
"learning_rate": 4.712980318650422e-05,
|
| 838 |
-
"loss": 0.3309,
|
| 839 |
-
"step": 980
|
| 840 |
-
},
|
| 841 |
-
{
|
| 842 |
-
"epoch": 0.4639175257731959,
|
| 843 |
-
"grad_norm": 1.4031989574432373,
|
| 844 |
-
"learning_rate": 4.710051546391753e-05,
|
| 845 |
-
"loss": 0.3616,
|
| 846 |
-
"step": 990
|
| 847 |
-
},
|
| 848 |
-
{
|
| 849 |
-
"epoch": 0.46860356138706655,
|
| 850 |
-
"grad_norm": 1.439453125,
|
| 851 |
-
"learning_rate": 4.7071227741330836e-05,
|
| 852 |
-
"loss": 0.3122,
|
| 853 |
-
"step": 1000
|
| 854 |
-
},
|
| 855 |
-
{
|
| 856 |
-
"epoch": 0.46860356138706655,
|
| 857 |
-
"eval_loss": 0.05414344370365143,
|
| 858 |
-
"eval_pearson_cosine": 0.8215390057088641,
|
| 859 |
-
"eval_pearson_dot": 0.7789934072191471,
|
| 860 |
-
"eval_pearson_euclidean": 0.8206818537339018,
|
| 861 |
-
"eval_pearson_manhattan": 0.8219733991381624,
|
| 862 |
-
"eval_runtime": 6.2607,
|
| 863 |
-
"eval_samples_per_second": 239.588,
|
| 864 |
-
"eval_spearman_cosine": 0.8235945278831797,
|
| 865 |
-
"eval_spearman_dot": 0.7745226194646113,
|
| 866 |
-
"eval_spearman_euclidean": 0.8268444005248111,
|
| 867 |
-
"eval_spearman_manhattan": 0.8284194308491212,
|
| 868 |
-
"eval_steps_per_second": 15.014,
|
| 869 |
-
"step": 1000
|
| 870 |
-
},
|
| 871 |
-
{
|
| 872 |
-
"epoch": 0.4732895970009372,
|
| 873 |
-
"grad_norm": 2.064291477203369,
|
| 874 |
-
"learning_rate": 4.7041940018744145e-05,
|
| 875 |
-
"loss": 0.3188,
|
| 876 |
-
"step": 1010
|
| 877 |
-
},
|
| 878 |
-
{
|
| 879 |
-
"epoch": 0.47797563261480785,
|
| 880 |
-
"grad_norm": 1.338891625404358,
|
| 881 |
-
"learning_rate": 4.7012652296157454e-05,
|
| 882 |
-
"loss": 0.3381,
|
| 883 |
-
"step": 1020
|
| 884 |
-
},
|
| 885 |
-
{
|
| 886 |
-
"epoch": 0.48266166822867856,
|
| 887 |
-
"grad_norm": 1.4479578733444214,
|
| 888 |
-
"learning_rate": 4.698336457357076e-05,
|
| 889 |
-
"loss": 0.3187,
|
| 890 |
-
"step": 1030
|
| 891 |
-
},
|
| 892 |
-
{
|
| 893 |
-
"epoch": 0.4873477038425492,
|
| 894 |
-
"grad_norm": 2.0868189334869385,
|
| 895 |
-
"learning_rate": 4.695407685098407e-05,
|
| 896 |
-
"loss": 0.3697,
|
| 897 |
-
"step": 1040
|
| 898 |
-
},
|
| 899 |
-
{
|
| 900 |
-
"epoch": 0.49203373945641987,
|
| 901 |
-
"grad_norm": 1.9820175170898438,
|
| 902 |
-
"learning_rate": 4.692478912839738e-05,
|
| 903 |
-
"loss": 0.3079,
|
| 904 |
-
"step": 1050
|
| 905 |
-
},
|
| 906 |
-
{
|
| 907 |
-
"epoch": 0.4967197750702905,
|
| 908 |
-
"grad_norm": 1.2479910850524902,
|
| 909 |
-
"learning_rate": 4.689550140581069e-05,
|
| 910 |
-
"loss": 0.3129,
|
| 911 |
-
"step": 1060
|
| 912 |
-
},
|
| 913 |
-
{
|
| 914 |
-
"epoch": 0.5014058106841612,
|
| 915 |
-
"grad_norm": 1.5005191564559937,
|
| 916 |
-
"learning_rate": 4.6866213683223996e-05,
|
| 917 |
-
"loss": 0.3588,
|
| 918 |
-
"step": 1070
|
| 919 |
-
},
|
| 920 |
-
{
|
| 921 |
-
"epoch": 0.5060918462980318,
|
| 922 |
-
"grad_norm": 1.730153203010559,
|
| 923 |
-
"learning_rate": 4.68369259606373e-05,
|
| 924 |
-
"loss": 0.3511,
|
| 925 |
-
"step": 1080
|
| 926 |
-
},
|
| 927 |
-
{
|
| 928 |
-
"epoch": 0.5107778819119025,
|
| 929 |
-
"grad_norm": 1.8256272077560425,
|
| 930 |
-
"learning_rate": 4.680763823805061e-05,
|
| 931 |
-
"loss": 0.3483,
|
| 932 |
-
"step": 1090
|
| 933 |
-
},
|
| 934 |
-
{
|
| 935 |
-
"epoch": 0.5154639175257731,
|
| 936 |
-
"grad_norm": 1.8275713920593262,
|
| 937 |
-
"learning_rate": 4.677835051546392e-05,
|
| 938 |
-
"loss": 0.3301,
|
| 939 |
-
"step": 1100
|
| 940 |
-
},
|
| 941 |
-
{
|
| 942 |
-
"epoch": 0.5154639175257731,
|
| 943 |
-
"eval_loss": 0.06174962595105171,
|
| 944 |
-
"eval_pearson_cosine": 0.8116036079365685,
|
| 945 |
-
"eval_pearson_dot": 0.7637833872485942,
|
| 946 |
-
"eval_pearson_euclidean": 0.8160585823410784,
|
| 947 |
-
"eval_pearson_manhattan": 0.8176746469698344,
|
| 948 |
-
"eval_runtime": 6.1383,
|
| 949 |
-
"eval_samples_per_second": 244.369,
|
| 950 |
-
"eval_spearman_cosine": 0.8149952741898824,
|
| 951 |
-
"eval_spearman_dot": 0.759784369983796,
|
| 952 |
-
"eval_spearman_euclidean": 0.8211588786730816,
|
| 953 |
-
"eval_spearman_manhattan": 0.8228445193252625,
|
| 954 |
-
"eval_steps_per_second": 15.314,
|
| 955 |
-
"step": 1100
|
| 956 |
-
},
|
| 957 |
-
{
|
| 958 |
-
"epoch": 0.5201499531396439,
|
| 959 |
-
"grad_norm": 1.7808656692504883,
|
| 960 |
-
"learning_rate": 4.674906279287723e-05,
|
| 961 |
-
"loss": 0.3378,
|
| 962 |
-
"step": 1110
|
| 963 |
-
},
|
| 964 |
-
{
|
| 965 |
-
"epoch": 0.5248359887535146,
|
| 966 |
-
"grad_norm": 1.3912303447723389,
|
| 967 |
-
"learning_rate": 4.671977507029054e-05,
|
| 968 |
-
"loss": 0.3247,
|
| 969 |
-
"step": 1120
|
| 970 |
-
},
|
| 971 |
-
{
|
| 972 |
-
"epoch": 0.5295220243673852,
|
| 973 |
-
"grad_norm": 1.619547724723816,
|
| 974 |
-
"learning_rate": 4.669048734770384e-05,
|
| 975 |
-
"loss": 0.3548,
|
| 976 |
-
"step": 1130
|
| 977 |
-
},
|
| 978 |
-
{
|
| 979 |
-
"epoch": 0.5342080599812559,
|
| 980 |
-
"grad_norm": 1.6785143613815308,
|
| 981 |
-
"learning_rate": 4.666119962511715e-05,
|
| 982 |
-
"loss": 0.4056,
|
| 983 |
-
"step": 1140
|
| 984 |
-
},
|
| 985 |
-
{
|
| 986 |
-
"epoch": 0.5388940955951266,
|
| 987 |
-
"grad_norm": 1.4282417297363281,
|
| 988 |
-
"learning_rate": 4.6631911902530465e-05,
|
| 989 |
-
"loss": 0.3136,
|
| 990 |
-
"step": 1150
|
| 991 |
-
},
|
| 992 |
-
{
|
| 993 |
-
"epoch": 0.5435801312089972,
|
| 994 |
-
"grad_norm": 1.5950373411178589,
|
| 995 |
-
"learning_rate": 4.660262417994377e-05,
|
| 996 |
-
"loss": 0.3094,
|
| 997 |
-
"step": 1160
|
| 998 |
-
},
|
| 999 |
-
{
|
| 1000 |
-
"epoch": 0.5482661668228679,
|
| 1001 |
-
"grad_norm": 1.9235565662384033,
|
| 1002 |
-
"learning_rate": 4.657333645735708e-05,
|
| 1003 |
-
"loss": 0.3409,
|
| 1004 |
-
"step": 1170
|
| 1005 |
-
},
|
| 1006 |
-
{
|
| 1007 |
-
"epoch": 0.5529522024367385,
|
| 1008 |
-
"grad_norm": 1.2192574739456177,
|
| 1009 |
-
"learning_rate": 4.6544048734770383e-05,
|
| 1010 |
-
"loss": 0.3387,
|
| 1011 |
-
"step": 1180
|
| 1012 |
-
},
|
| 1013 |
-
{
|
| 1014 |
-
"epoch": 0.5576382380506092,
|
| 1015 |
-
"grad_norm": 1.5550990104675293,
|
| 1016 |
-
"learning_rate": 4.651476101218369e-05,
|
| 1017 |
-
"loss": 0.3184,
|
| 1018 |
-
"step": 1190
|
| 1019 |
-
},
|
| 1020 |
-
{
|
| 1021 |
-
"epoch": 0.5623242736644799,
|
| 1022 |
-
"grad_norm": 1.8576079607009888,
|
| 1023 |
-
"learning_rate": 4.6485473289597e-05,
|
| 1024 |
-
"loss": 0.3637,
|
| 1025 |
-
"step": 1200
|
| 1026 |
-
},
|
| 1027 |
-
{
|
| 1028 |
-
"epoch": 0.5623242736644799,
|
| 1029 |
-
"eval_loss": 0.05324321612715721,
|
| 1030 |
-
"eval_pearson_cosine": 0.8107929850304814,
|
| 1031 |
-
"eval_pearson_dot": 0.768063847349957,
|
| 1032 |
-
"eval_pearson_euclidean": 0.8155502077488883,
|
| 1033 |
-
"eval_pearson_manhattan": 0.8174981555238503,
|
| 1034 |
-
"eval_runtime": 5.9644,
|
| 1035 |
-
"eval_samples_per_second": 251.492,
|
| 1036 |
-
"eval_spearman_cosine": 0.8145222586962418,
|
| 1037 |
-
"eval_spearman_dot": 0.7642997219390888,
|
| 1038 |
-
"eval_spearman_euclidean": 0.8201735536723759,
|
| 1039 |
-
"eval_spearman_manhattan": 0.8222186632592043,
|
| 1040 |
-
"eval_steps_per_second": 15.76,
|
| 1041 |
-
"step": 1200
|
| 1042 |
-
},
|
| 1043 |
-
{
|
| 1044 |
-
"epoch": 0.5670103092783505,
|
| 1045 |
-
"grad_norm": 1.41835618019104,
|
| 1046 |
-
"learning_rate": 4.6456185567010316e-05,
|
| 1047 |
-
"loss": 0.3039,
|
| 1048 |
-
"step": 1210
|
| 1049 |
-
},
|
| 1050 |
-
{
|
| 1051 |
-
"epoch": 0.5716963448922212,
|
| 1052 |
-
"grad_norm": 1.3850994110107422,
|
| 1053 |
-
"learning_rate": 4.6426897844423624e-05,
|
| 1054 |
-
"loss": 0.3187,
|
| 1055 |
-
"step": 1220
|
| 1056 |
-
},
|
| 1057 |
-
{
|
| 1058 |
-
"epoch": 0.5763823805060918,
|
| 1059 |
-
"grad_norm": 1.437373399734497,
|
| 1060 |
-
"learning_rate": 4.6397610121836926e-05,
|
| 1061 |
-
"loss": 0.3337,
|
| 1062 |
-
"step": 1230
|
| 1063 |
-
},
|
| 1064 |
-
{
|
| 1065 |
-
"epoch": 0.5810684161199625,
|
| 1066 |
-
"grad_norm": 1.2328146696090698,
|
| 1067 |
-
"learning_rate": 4.6368322399250235e-05,
|
| 1068 |
-
"loss": 0.2975,
|
| 1069 |
-
"step": 1240
|
| 1070 |
-
},
|
| 1071 |
-
{
|
| 1072 |
-
"epoch": 0.5857544517338332,
|
| 1073 |
-
"grad_norm": 1.6191329956054688,
|
| 1074 |
-
"learning_rate": 4.633903467666354e-05,
|
| 1075 |
-
"loss": 0.3275,
|
| 1076 |
-
"step": 1250
|
| 1077 |
-
},
|
| 1078 |
-
{
|
| 1079 |
-
"epoch": 0.5904404873477038,
|
| 1080 |
-
"grad_norm": 1.695470929145813,
|
| 1081 |
-
"learning_rate": 4.630974695407685e-05,
|
| 1082 |
-
"loss": 0.3485,
|
| 1083 |
-
"step": 1260
|
| 1084 |
-
},
|
| 1085 |
-
{
|
| 1086 |
-
"epoch": 0.5951265229615745,
|
| 1087 |
-
"grad_norm": 1.6120591163635254,
|
| 1088 |
-
"learning_rate": 4.628045923149017e-05,
|
| 1089 |
-
"loss": 0.3515,
|
| 1090 |
-
"step": 1270
|
| 1091 |
-
},
|
| 1092 |
-
{
|
| 1093 |
-
"epoch": 0.5998125585754451,
|
| 1094 |
-
"grad_norm": 1.6157792806625366,
|
| 1095 |
-
"learning_rate": 4.625117150890347e-05,
|
| 1096 |
-
"loss": 0.3043,
|
| 1097 |
-
"step": 1280
|
| 1098 |
-
},
|
| 1099 |
-
{
|
| 1100 |
-
"epoch": 0.6044985941893158,
|
| 1101 |
-
"grad_norm": 1.1465294361114502,
|
| 1102 |
-
"learning_rate": 4.622188378631678e-05,
|
| 1103 |
-
"loss": 0.2884,
|
| 1104 |
-
"step": 1290
|
| 1105 |
-
},
|
| 1106 |
-
{
|
| 1107 |
-
"epoch": 0.6091846298031866,
|
| 1108 |
-
"grad_norm": 1.583688497543335,
|
| 1109 |
-
"learning_rate": 4.6192596063730086e-05,
|
| 1110 |
-
"loss": 0.2885,
|
| 1111 |
-
"step": 1300
|
| 1112 |
-
},
|
| 1113 |
-
{
|
| 1114 |
-
"epoch": 0.6091846298031866,
|
| 1115 |
-
"eval_loss": 0.04511857405304909,
|
| 1116 |
-
"eval_pearson_cosine": 0.8272008065998051,
|
| 1117 |
-
"eval_pearson_dot": 0.7924789666171037,
|
| 1118 |
-
"eval_pearson_euclidean": 0.8268065548805623,
|
| 1119 |
-
"eval_pearson_manhattan": 0.827500349653536,
|
| 1120 |
-
"eval_runtime": 6.0244,
|
| 1121 |
-
"eval_samples_per_second": 248.988,
|
| 1122 |
-
"eval_spearman_cosine": 0.8278338594350843,
|
| 1123 |
-
"eval_spearman_dot": 0.7887595412839734,
|
| 1124 |
-
"eval_spearman_euclidean": 0.8317669408319824,
|
| 1125 |
-
"eval_spearman_manhattan": 0.8323949761116776,
|
| 1126 |
-
"eval_steps_per_second": 15.603,
|
| 1127 |
-
"step": 1300
|
| 1128 |
-
},
|
| 1129 |
-
{
|
| 1130 |
-
"epoch": 0.6138706654170571,
|
| 1131 |
-
"grad_norm": 1.081640362739563,
|
| 1132 |
-
"learning_rate": 4.6163308341143395e-05,
|
| 1133 |
-
"loss": 0.3526,
|
| 1134 |
-
"step": 1310
|
| 1135 |
-
},
|
| 1136 |
-
{
|
| 1137 |
-
"epoch": 0.6185567010309279,
|
| 1138 |
-
"grad_norm": 1.760512351989746,
|
| 1139 |
-
"learning_rate": 4.61340206185567e-05,
|
| 1140 |
-
"loss": 0.3113,
|
| 1141 |
-
"step": 1320
|
| 1142 |
-
},
|
| 1143 |
-
{
|
| 1144 |
-
"epoch": 0.6232427366447985,
|
| 1145 |
-
"grad_norm": 1.106444239616394,
|
| 1146 |
-
"learning_rate": 4.610473289597001e-05,
|
| 1147 |
-
"loss": 0.3126,
|
| 1148 |
-
"step": 1330
|
| 1149 |
-
},
|
| 1150 |
-
{
|
| 1151 |
-
"epoch": 0.6279287722586692,
|
| 1152 |
-
"grad_norm": 1.3500837087631226,
|
| 1153 |
-
"learning_rate": 4.607544517338332e-05,
|
| 1154 |
-
"loss": 0.3094,
|
| 1155 |
-
"step": 1340
|
| 1156 |
-
},
|
| 1157 |
-
{
|
| 1158 |
-
"epoch": 0.6326148078725399,
|
| 1159 |
-
"grad_norm": 1.727953553199768,
|
| 1160 |
-
"learning_rate": 4.604615745079663e-05,
|
| 1161 |
-
"loss": 0.3304,
|
| 1162 |
-
"step": 1350
|
| 1163 |
-
},
|
| 1164 |
-
{
|
| 1165 |
-
"epoch": 0.6373008434864105,
|
| 1166 |
-
"grad_norm": 1.4341022968292236,
|
| 1167 |
-
"learning_rate": 4.601686972820994e-05,
|
| 1168 |
-
"loss": 0.2804,
|
| 1169 |
-
"step": 1360
|
| 1170 |
-
},
|
| 1171 |
-
{
|
| 1172 |
-
"epoch": 0.6419868791002812,
|
| 1173 |
-
"grad_norm": 1.4479708671569824,
|
| 1174 |
-
"learning_rate": 4.5987582005623246e-05,
|
| 1175 |
-
"loss": 0.31,
|
| 1176 |
-
"step": 1370
|
| 1177 |
-
},
|
| 1178 |
-
{
|
| 1179 |
-
"epoch": 0.6466729147141518,
|
| 1180 |
-
"grad_norm": 1.5667890310287476,
|
| 1181 |
-
"learning_rate": 4.5958294283036554e-05,
|
| 1182 |
-
"loss": 0.3149,
|
| 1183 |
-
"step": 1380
|
| 1184 |
-
},
|
| 1185 |
-
{
|
| 1186 |
-
"epoch": 0.6513589503280225,
|
| 1187 |
-
"grad_norm": 1.7333146333694458,
|
| 1188 |
-
"learning_rate": 4.592900656044986e-05,
|
| 1189 |
-
"loss": 0.3247,
|
| 1190 |
-
"step": 1390
|
| 1191 |
-
},
|
| 1192 |
-
{
|
| 1193 |
-
"epoch": 0.6560449859418932,
|
| 1194 |
-
"grad_norm": 1.914392113685608,
|
| 1195 |
-
"learning_rate": 4.589971883786317e-05,
|
| 1196 |
-
"loss": 0.2852,
|
| 1197 |
-
"step": 1400
|
| 1198 |
-
},
|
| 1199 |
-
{
|
| 1200 |
-
"epoch": 0.6560449859418932,
|
| 1201 |
-
"eval_loss": 0.04731455817818642,
|
| 1202 |
-
"eval_pearson_cosine": 0.8245641713392331,
|
| 1203 |
-
"eval_pearson_dot": 0.7893189374890994,
|
| 1204 |
-
"eval_pearson_euclidean": 0.8220644314223797,
|
| 1205 |
-
"eval_pearson_manhattan": 0.8227839674683928,
|
| 1206 |
-
"eval_runtime": 6.0521,
|
| 1207 |
-
"eval_samples_per_second": 247.846,
|
| 1208 |
-
"eval_spearman_cosine": 0.8264178003782281,
|
| 1209 |
-
"eval_spearman_dot": 0.7874134051082518,
|
| 1210 |
-
"eval_spearman_euclidean": 0.8274821508565314,
|
| 1211 |
-
"eval_spearman_manhattan": 0.8280999297389011,
|
| 1212 |
-
"eval_steps_per_second": 15.532,
|
| 1213 |
-
"step": 1400
|
| 1214 |
-
},
|
| 1215 |
-
{
|
| 1216 |
-
"epoch": 0.6607310215557638,
|
| 1217 |
-
"grad_norm": 1.2458995580673218,
|
| 1218 |
-
"learning_rate": 4.587043111527648e-05,
|
| 1219 |
-
"loss": 0.3068,
|
| 1220 |
-
"step": 1410
|
| 1221 |
-
},
|
| 1222 |
-
{
|
| 1223 |
-
"epoch": 0.6654170571696345,
|
| 1224 |
-
"grad_norm": 1.6540151834487915,
|
| 1225 |
-
"learning_rate": 4.584114339268979e-05,
|
| 1226 |
-
"loss": 0.3034,
|
| 1227 |
-
"step": 1420
|
| 1228 |
-
},
|
| 1229 |
-
{
|
| 1230 |
-
"epoch": 0.6701030927835051,
|
| 1231 |
-
"grad_norm": 1.2585715055465698,
|
| 1232 |
-
"learning_rate": 4.581185567010309e-05,
|
| 1233 |
-
"loss": 0.3297,
|
| 1234 |
-
"step": 1430
|
| 1235 |
-
},
|
| 1236 |
-
{
|
| 1237 |
-
"epoch": 0.6747891283973758,
|
| 1238 |
-
"grad_norm": 1.5088609457015991,
|
| 1239 |
-
"learning_rate": 4.5782567947516406e-05,
|
| 1240 |
-
"loss": 0.3444,
|
| 1241 |
-
"step": 1440
|
| 1242 |
-
},
|
| 1243 |
-
{
|
| 1244 |
-
"epoch": 0.6794751640112465,
|
| 1245 |
-
"grad_norm": 1.3120390176773071,
|
| 1246 |
-
"learning_rate": 4.5753280224929714e-05,
|
| 1247 |
-
"loss": 0.2882,
|
| 1248 |
-
"step": 1450
|
| 1249 |
-
},
|
| 1250 |
-
{
|
| 1251 |
-
"epoch": 0.6841611996251171,
|
| 1252 |
-
"grad_norm": 1.1074262857437134,
|
| 1253 |
-
"learning_rate": 4.572399250234302e-05,
|
| 1254 |
-
"loss": 0.287,
|
| 1255 |
-
"step": 1460
|
| 1256 |
-
},
|
| 1257 |
-
{
|
| 1258 |
-
"epoch": 0.6888472352389878,
|
| 1259 |
-
"grad_norm": 1.5284086465835571,
|
| 1260 |
-
"learning_rate": 4.569470477975633e-05,
|
| 1261 |
-
"loss": 0.3175,
|
| 1262 |
-
"step": 1470
|
| 1263 |
-
},
|
| 1264 |
-
{
|
| 1265 |
-
"epoch": 0.6935332708528584,
|
| 1266 |
-
"grad_norm": 1.5610471963882446,
|
| 1267 |
-
"learning_rate": 4.566541705716963e-05,
|
| 1268 |
-
"loss": 0.3033,
|
| 1269 |
-
"step": 1480
|
| 1270 |
-
},
|
| 1271 |
-
{
|
| 1272 |
-
"epoch": 0.6982193064667291,
|
| 1273 |
-
"grad_norm": 1.1839112043380737,
|
| 1274 |
-
"learning_rate": 4.563612933458294e-05,
|
| 1275 |
-
"loss": 0.2917,
|
| 1276 |
-
"step": 1490
|
| 1277 |
-
},
|
| 1278 |
-
{
|
| 1279 |
-
"epoch": 0.7029053420805998,
|
| 1280 |
-
"grad_norm": 1.2611138820648193,
|
| 1281 |
-
"learning_rate": 4.560684161199626e-05,
|
| 1282 |
-
"loss": 0.3225,
|
| 1283 |
-
"step": 1500
|
| 1284 |
-
},
|
| 1285 |
-
{
|
| 1286 |
-
"epoch": 0.7029053420805998,
|
| 1287 |
-
"eval_loss": 0.05073302239179611,
|
| 1288 |
-
"eval_pearson_cosine": 0.8258767050389224,
|
| 1289 |
-
"eval_pearson_dot": 0.7737186817222579,
|
| 1290 |
-
"eval_pearson_euclidean": 0.826299226589029,
|
| 1291 |
-
"eval_pearson_manhattan": 0.8274116157485736,
|
| 1292 |
-
"eval_runtime": 6.1756,
|
| 1293 |
-
"eval_samples_per_second": 242.892,
|
| 1294 |
-
"eval_spearman_cosine": 0.828410027637777,
|
| 1295 |
-
"eval_spearman_dot": 0.7707599871747091,
|
| 1296 |
-
"eval_spearman_euclidean": 0.8324611004376368,
|
| 1297 |
-
"eval_spearman_manhattan": 0.8335300441487923,
|
| 1298 |
-
"eval_steps_per_second": 15.221,
|
| 1299 |
-
"step": 1500
|
| 1300 |
-
},
|
| 1301 |
-
{
|
| 1302 |
-
"epoch": 0.7075913776944704,
|
| 1303 |
-
"grad_norm": 1.3322237730026245,
|
| 1304 |
-
"learning_rate": 4.5577553889409565e-05,
|
| 1305 |
-
"loss": 0.27,
|
| 1306 |
-
"step": 1510
|
| 1307 |
-
},
|
| 1308 |
-
{
|
| 1309 |
-
"epoch": 0.7122774133083412,
|
| 1310 |
-
"grad_norm": 1.707694172859192,
|
| 1311 |
-
"learning_rate": 4.5548266166822874e-05,
|
| 1312 |
-
"loss": 0.3249,
|
| 1313 |
-
"step": 1520
|
| 1314 |
-
},
|
| 1315 |
-
{
|
| 1316 |
-
"epoch": 0.7169634489222118,
|
| 1317 |
-
"grad_norm": 1.280220627784729,
|
| 1318 |
-
"learning_rate": 4.5518978444236176e-05,
|
| 1319 |
-
"loss": 0.3021,
|
| 1320 |
-
"step": 1530
|
| 1321 |
-
},
|
| 1322 |
-
{
|
| 1323 |
-
"epoch": 0.7216494845360825,
|
| 1324 |
-
"grad_norm": 1.2478505373001099,
|
| 1325 |
-
"learning_rate": 4.5489690721649484e-05,
|
| 1326 |
-
"loss": 0.3033,
|
| 1327 |
-
"step": 1540
|
| 1328 |
-
},
|
| 1329 |
-
{
|
| 1330 |
-
"epoch": 0.7263355201499532,
|
| 1331 |
-
"grad_norm": 1.5291355848312378,
|
| 1332 |
-
"learning_rate": 4.546040299906279e-05,
|
| 1333 |
-
"loss": 0.2789,
|
| 1334 |
-
"step": 1550
|
| 1335 |
-
},
|
| 1336 |
-
{
|
| 1337 |
-
"epoch": 0.7310215557638238,
|
| 1338 |
-
"grad_norm": 1.6631042957305908,
|
| 1339 |
-
"learning_rate": 4.543111527647611e-05,
|
| 1340 |
-
"loss": 0.3271,
|
| 1341 |
-
"step": 1560
|
| 1342 |
-
},
|
| 1343 |
-
{
|
| 1344 |
-
"epoch": 0.7357075913776945,
|
| 1345 |
-
"grad_norm": 1.5178686380386353,
|
| 1346 |
-
"learning_rate": 4.540182755388942e-05,
|
| 1347 |
-
"loss": 0.3092,
|
| 1348 |
-
"step": 1570
|
| 1349 |
-
},
|
| 1350 |
-
{
|
| 1351 |
-
"epoch": 0.7403936269915652,
|
| 1352 |
-
"grad_norm": 1.043636441230774,
|
| 1353 |
-
"learning_rate": 4.537253983130272e-05,
|
| 1354 |
-
"loss": 0.2863,
|
| 1355 |
-
"step": 1580
|
| 1356 |
-
},
|
| 1357 |
-
{
|
| 1358 |
-
"epoch": 0.7450796626054358,
|
| 1359 |
-
"grad_norm": 1.2474050521850586,
|
| 1360 |
-
"learning_rate": 4.534325210871603e-05,
|
| 1361 |
-
"loss": 0.3184,
|
| 1362 |
-
"step": 1590
|
| 1363 |
-
},
|
| 1364 |
-
{
|
| 1365 |
-
"epoch": 0.7497656982193065,
|
| 1366 |
-
"grad_norm": 1.5337306261062622,
|
| 1367 |
-
"learning_rate": 4.5313964386129336e-05,
|
| 1368 |
-
"loss": 0.3201,
|
| 1369 |
-
"step": 1600
|
| 1370 |
-
},
|
| 1371 |
-
{
|
| 1372 |
-
"epoch": 0.7497656982193065,
|
| 1373 |
-
"eval_loss": 0.04666765406727791,
|
| 1374 |
-
"eval_pearson_cosine": 0.8247560159233132,
|
| 1375 |
-
"eval_pearson_dot": 0.7799855816644161,
|
| 1376 |
-
"eval_pearson_euclidean": 0.8221947952436466,
|
| 1377 |
-
"eval_pearson_manhattan": 0.823206574209081,
|
| 1378 |
-
"eval_runtime": 6.3641,
|
| 1379 |
-
"eval_samples_per_second": 235.697,
|
| 1380 |
-
"eval_spearman_cosine": 0.8268195350952193,
|
| 1381 |
-
"eval_spearman_dot": 0.7772130800409067,
|
| 1382 |
-
"eval_spearman_euclidean": 0.8274185345815266,
|
| 1383 |
-
"eval_spearman_manhattan": 0.8282217123816773,
|
| 1384 |
-
"eval_steps_per_second": 14.77,
|
| 1385 |
-
"step": 1600
|
| 1386 |
-
},
|
| 1387 |
-
{
|
| 1388 |
-
"epoch": 0.7544517338331771,
|
| 1389 |
-
"grad_norm": 1.8070961236953735,
|
| 1390 |
-
"learning_rate": 4.5284676663542644e-05,
|
| 1391 |
-
"loss": 0.2939,
|
| 1392 |
-
"step": 1610
|
| 1393 |
-
},
|
| 1394 |
-
{
|
| 1395 |
-
"epoch": 0.7591377694470478,
|
| 1396 |
-
"grad_norm": 1.3418357372283936,
|
| 1397 |
-
"learning_rate": 4.525538894095596e-05,
|
| 1398 |
-
"loss": 0.3122,
|
| 1399 |
-
"step": 1620
|
| 1400 |
-
},
|
| 1401 |
-
{
|
| 1402 |
-
"epoch": 0.7638238050609185,
|
| 1403 |
-
"grad_norm": 1.4164036512374878,
|
| 1404 |
-
"learning_rate": 4.522610121836926e-05,
|
| 1405 |
-
"loss": 0.294,
|
| 1406 |
-
"step": 1630
|
| 1407 |
-
},
|
| 1408 |
-
{
|
| 1409 |
-
"epoch": 0.7685098406747891,
|
| 1410 |
-
"grad_norm": 1.4862034320831299,
|
| 1411 |
-
"learning_rate": 4.519681349578257e-05,
|
| 1412 |
-
"loss": 0.3087,
|
| 1413 |
-
"step": 1640
|
| 1414 |
-
},
|
| 1415 |
-
{
|
| 1416 |
-
"epoch": 0.7731958762886598,
|
| 1417 |
-
"grad_norm": 1.3497341871261597,
|
| 1418 |
-
"learning_rate": 4.516752577319588e-05,
|
| 1419 |
-
"loss": 0.3135,
|
| 1420 |
-
"step": 1650
|
| 1421 |
-
},
|
| 1422 |
-
{
|
| 1423 |
-
"epoch": 0.7778819119025304,
|
| 1424 |
-
"grad_norm": 1.4912623167037964,
|
| 1425 |
-
"learning_rate": 4.513823805060919e-05,
|
| 1426 |
-
"loss": 0.3556,
|
| 1427 |
-
"step": 1660
|
| 1428 |
-
},
|
| 1429 |
-
{
|
| 1430 |
-
"epoch": 0.7825679475164011,
|
| 1431 |
-
"grad_norm": 1.4625390768051147,
|
| 1432 |
-
"learning_rate": 4.5108950328022495e-05,
|
| 1433 |
-
"loss": 0.2913,
|
| 1434 |
-
"step": 1670
|
| 1435 |
-
},
|
| 1436 |
-
{
|
| 1437 |
-
"epoch": 0.7872539831302718,
|
| 1438 |
-
"grad_norm": 1.7304317951202393,
|
| 1439 |
-
"learning_rate": 4.5079662605435804e-05,
|
| 1440 |
-
"loss": 0.3139,
|
| 1441 |
-
"step": 1680
|
| 1442 |
-
},
|
| 1443 |
-
{
|
| 1444 |
-
"epoch": 0.7919400187441424,
|
| 1445 |
-
"grad_norm": 1.4902634620666504,
|
| 1446 |
-
"learning_rate": 4.505037488284911e-05,
|
| 1447 |
-
"loss": 0.3286,
|
| 1448 |
-
"step": 1690
|
| 1449 |
-
},
|
| 1450 |
-
{
|
| 1451 |
-
"epoch": 0.7966260543580131,
|
| 1452 |
-
"grad_norm": 1.2882981300354004,
|
| 1453 |
-
"learning_rate": 4.502108716026242e-05,
|
| 1454 |
-
"loss": 0.3199,
|
| 1455 |
-
"step": 1700
|
| 1456 |
-
},
|
| 1457 |
-
{
|
| 1458 |
-
"epoch": 0.7966260543580131,
|
| 1459 |
-
"eval_loss": 0.05109428986907005,
|
| 1460 |
-
"eval_pearson_cosine": 0.821506482740503,
|
| 1461 |
-
"eval_pearson_dot": 0.7701997371934368,
|
| 1462 |
-
"eval_pearson_euclidean": 0.8256940312734855,
|
| 1463 |
-
"eval_pearson_manhattan": 0.8266471352779732,
|
| 1464 |
-
"eval_runtime": 5.9909,
|
| 1465 |
-
"eval_samples_per_second": 250.381,
|
| 1466 |
-
"eval_spearman_cosine": 0.8238836226780246,
|
| 1467 |
-
"eval_spearman_dot": 0.7658343447023538,
|
| 1468 |
-
"eval_spearman_euclidean": 0.8307515392696511,
|
| 1469 |
-
"eval_spearman_manhattan": 0.8322184136990032,
|
| 1470 |
-
"eval_steps_per_second": 15.691,
|
| 1471 |
-
"step": 1700
|
| 1472 |
-
},
|
| 1473 |
-
{
|
| 1474 |
-
"epoch": 0.8013120899718837,
|
| 1475 |
-
"grad_norm": 1.4731862545013428,
|
| 1476 |
-
"learning_rate": 4.499179943767573e-05,
|
| 1477 |
-
"loss": 0.3262,
|
| 1478 |
-
"step": 1710
|
| 1479 |
-
},
|
| 1480 |
-
{
|
| 1481 |
-
"epoch": 0.8059981255857545,
|
| 1482 |
-
"grad_norm": 1.440738320350647,
|
| 1483 |
-
"learning_rate": 4.496251171508904e-05,
|
| 1484 |
-
"loss": 0.2911,
|
| 1485 |
-
"step": 1720
|
| 1486 |
-
},
|
| 1487 |
-
{
|
| 1488 |
-
"epoch": 0.8106841611996252,
|
| 1489 |
-
"grad_norm": 1.594072699546814,
|
| 1490 |
-
"learning_rate": 4.4933223992502347e-05,
|
| 1491 |
-
"loss": 0.3067,
|
| 1492 |
-
"step": 1730
|
| 1493 |
-
},
|
| 1494 |
-
{
|
| 1495 |
-
"epoch": 0.8153701968134958,
|
| 1496 |
-
"grad_norm": 1.2838362455368042,
|
| 1497 |
-
"learning_rate": 4.4903936269915655e-05,
|
| 1498 |
-
"loss": 0.2976,
|
| 1499 |
-
"step": 1740
|
| 1500 |
-
},
|
| 1501 |
-
{
|
| 1502 |
-
"epoch": 0.8200562324273665,
|
| 1503 |
-
"grad_norm": 1.3946473598480225,
|
| 1504 |
-
"learning_rate": 4.4874648547328964e-05,
|
| 1505 |
-
"loss": 0.2945,
|
| 1506 |
-
"step": 1750
|
| 1507 |
-
},
|
| 1508 |
-
{
|
| 1509 |
-
"epoch": 0.8247422680412371,
|
| 1510 |
-
"grad_norm": 1.5584791898727417,
|
| 1511 |
-
"learning_rate": 4.484536082474227e-05,
|
| 1512 |
-
"loss": 0.3196,
|
| 1513 |
-
"step": 1760
|
| 1514 |
-
},
|
| 1515 |
-
{
|
| 1516 |
-
"epoch": 0.8294283036551078,
|
| 1517 |
-
"grad_norm": 1.5150123834609985,
|
| 1518 |
-
"learning_rate": 4.481607310215558e-05,
|
| 1519 |
-
"loss": 0.2792,
|
| 1520 |
-
"step": 1770
|
| 1521 |
-
},
|
| 1522 |
-
{
|
| 1523 |
-
"epoch": 0.8341143392689785,
|
| 1524 |
-
"grad_norm": 1.5679230690002441,
|
| 1525 |
-
"learning_rate": 4.478678537956888e-05,
|
| 1526 |
-
"loss": 0.289,
|
| 1527 |
-
"step": 1780
|
| 1528 |
-
},
|
| 1529 |
-
{
|
| 1530 |
-
"epoch": 0.8388003748828491,
|
| 1531 |
-
"grad_norm": 1.100917100906372,
|
| 1532 |
-
"learning_rate": 4.47574976569822e-05,
|
| 1533 |
-
"loss": 0.3021,
|
| 1534 |
-
"step": 1790
|
| 1535 |
-
},
|
| 1536 |
-
{
|
| 1537 |
-
"epoch": 0.8434864104967198,
|
| 1538 |
-
"grad_norm": 1.6804919242858887,
|
| 1539 |
-
"learning_rate": 4.4728209934395506e-05,
|
| 1540 |
-
"loss": 0.2431,
|
| 1541 |
-
"step": 1800
|
| 1542 |
-
},
|
| 1543 |
-
{
|
| 1544 |
-
"epoch": 0.8434864104967198,
|
| 1545 |
-
"eval_loss": 0.04816513508558273,
|
| 1546 |
-
"eval_pearson_cosine": 0.8270891703367056,
|
| 1547 |
-
"eval_pearson_dot": 0.7790513762200462,
|
| 1548 |
-
"eval_pearson_euclidean": 0.8276507097787587,
|
| 1549 |
-
"eval_pearson_manhattan": 0.8282151299272726,
|
| 1550 |
-
"eval_runtime": 6.0115,
|
| 1551 |
-
"eval_samples_per_second": 249.523,
|
| 1552 |
-
"eval_spearman_cosine": 0.8286758008396278,
|
| 1553 |
-
"eval_spearman_dot": 0.7748761234075037,
|
| 1554 |
-
"eval_spearman_euclidean": 0.8326421047981132,
|
| 1555 |
-
"eval_spearman_manhattan": 0.8332581593954519,
|
| 1556 |
-
"eval_steps_per_second": 15.637,
|
| 1557 |
-
"step": 1800
|
| 1558 |
-
},
|
| 1559 |
-
{
|
| 1560 |
-
"epoch": 0.8481724461105904,
|
| 1561 |
-
"grad_norm": 1.3295296430587769,
|
| 1562 |
-
"learning_rate": 4.4698922211808815e-05,
|
| 1563 |
-
"loss": 0.2877,
|
| 1564 |
-
"step": 1810
|
| 1565 |
-
},
|
| 1566 |
-
{
|
| 1567 |
-
"epoch": 0.8528584817244611,
|
| 1568 |
-
"grad_norm": 1.3233381509780884,
|
| 1569 |
-
"learning_rate": 4.466963448922212e-05,
|
| 1570 |
-
"loss": 0.2982,
|
| 1571 |
-
"step": 1820
|
| 1572 |
-
},
|
| 1573 |
-
{
|
| 1574 |
-
"epoch": 0.8575445173383318,
|
| 1575 |
-
"grad_norm": 1.5737247467041016,
|
| 1576 |
-
"learning_rate": 4.4640346766635425e-05,
|
| 1577 |
-
"loss": 0.2939,
|
| 1578 |
-
"step": 1830
|
| 1579 |
-
},
|
| 1580 |
-
{
|
| 1581 |
-
"epoch": 0.8622305529522024,
|
| 1582 |
-
"grad_norm": 1.4237866401672363,
|
| 1583 |
-
"learning_rate": 4.4611059044048734e-05,
|
| 1584 |
-
"loss": 0.3154,
|
| 1585 |
-
"step": 1840
|
| 1586 |
-
},
|
| 1587 |
-
{
|
| 1588 |
-
"epoch": 0.8669165885660731,
|
| 1589 |
-
"grad_norm": 1.4213505983352661,
|
| 1590 |
-
"learning_rate": 4.458177132146205e-05,
|
| 1591 |
-
"loss": 0.3085,
|
| 1592 |
-
"step": 1850
|
| 1593 |
-
},
|
| 1594 |
-
{
|
| 1595 |
-
"epoch": 0.8716026241799437,
|
| 1596 |
-
"grad_norm": 1.4691981077194214,
|
| 1597 |
-
"learning_rate": 4.455248359887536e-05,
|
| 1598 |
-
"loss": 0.3141,
|
| 1599 |
-
"step": 1860
|
| 1600 |
-
},
|
| 1601 |
-
{
|
| 1602 |
-
"epoch": 0.8762886597938144,
|
| 1603 |
-
"grad_norm": 1.2567983865737915,
|
| 1604 |
-
"learning_rate": 4.452319587628866e-05,
|
| 1605 |
-
"loss": 0.2998,
|
| 1606 |
-
"step": 1870
|
| 1607 |
-
},
|
| 1608 |
-
{
|
| 1609 |
-
"epoch": 0.8809746954076851,
|
| 1610 |
-
"grad_norm": 1.359161615371704,
|
| 1611 |
-
"learning_rate": 4.449390815370197e-05,
|
| 1612 |
-
"loss": 0.2891,
|
| 1613 |
-
"step": 1880
|
| 1614 |
-
},
|
| 1615 |
-
{
|
| 1616 |
-
"epoch": 0.8856607310215557,
|
| 1617 |
-
"grad_norm": 1.4557381868362427,
|
| 1618 |
-
"learning_rate": 4.4464620431115277e-05,
|
| 1619 |
-
"loss": 0.3068,
|
| 1620 |
-
"step": 1890
|
| 1621 |
-
},
|
| 1622 |
-
{
|
| 1623 |
-
"epoch": 0.8903467666354264,
|
| 1624 |
-
"grad_norm": 1.2976425886154175,
|
| 1625 |
-
"learning_rate": 4.4435332708528585e-05,
|
| 1626 |
-
"loss": 0.3051,
|
| 1627 |
-
"step": 1900
|
| 1628 |
-
},
|
| 1629 |
-
{
|
| 1630 |
-
"epoch": 0.8903467666354264,
|
| 1631 |
-
"eval_loss": 0.04645664617419243,
|
| 1632 |
-
"eval_pearson_cosine": 0.8277422687657179,
|
| 1633 |
-
"eval_pearson_dot": 0.7813512072475213,
|
| 1634 |
-
"eval_pearson_euclidean": 0.8249433703124964,
|
| 1635 |
-
"eval_pearson_manhattan": 0.8257048412620271,
|
| 1636 |
-
"eval_runtime": 6.5216,
|
| 1637 |
-
"eval_samples_per_second": 230.006,
|
| 1638 |
-
"eval_spearman_cosine": 0.829450894301767,
|
| 1639 |
-
"eval_spearman_dot": 0.778239364994612,
|
| 1640 |
-
"eval_spearman_euclidean": 0.8318708700161123,
|
| 1641 |
-
"eval_spearman_manhattan": 0.8323977739737524,
|
| 1642 |
-
"eval_steps_per_second": 14.414,
|
| 1643 |
-
"step": 1900
|
| 1644 |
-
},
|
| 1645 |
-
{
|
| 1646 |
-
"epoch": 0.895032802249297,
|
| 1647 |
-
"grad_norm": 2.003398895263672,
|
| 1648 |
-
"learning_rate": 4.44060449859419e-05,
|
| 1649 |
-
"loss": 0.3261,
|
| 1650 |
-
"step": 1910
|
| 1651 |
-
},
|
| 1652 |
-
{
|
| 1653 |
-
"epoch": 0.8997188378631678,
|
| 1654 |
-
"grad_norm": 1.5879777669906616,
|
| 1655 |
-
"learning_rate": 4.43767572633552e-05,
|
| 1656 |
-
"loss": 0.2905,
|
| 1657 |
-
"step": 1920
|
| 1658 |
-
},
|
| 1659 |
-
{
|
| 1660 |
-
"epoch": 0.9044048734770385,
|
| 1661 |
-
"grad_norm": 1.239495873451233,
|
| 1662 |
-
"learning_rate": 4.434746954076851e-05,
|
| 1663 |
-
"loss": 0.3005,
|
| 1664 |
-
"step": 1930
|
| 1665 |
-
},
|
| 1666 |
-
{
|
| 1667 |
-
"epoch": 0.9090909090909091,
|
| 1668 |
-
"grad_norm": 1.643388271331787,
|
| 1669 |
-
"learning_rate": 4.431818181818182e-05,
|
| 1670 |
-
"loss": 0.2901,
|
| 1671 |
-
"step": 1940
|
| 1672 |
-
},
|
| 1673 |
-
{
|
| 1674 |
-
"epoch": 0.9137769447047798,
|
| 1675 |
-
"grad_norm": 1.5589861869812012,
|
| 1676 |
-
"learning_rate": 4.428889409559513e-05,
|
| 1677 |
-
"loss": 0.269,
|
| 1678 |
-
"step": 1950
|
| 1679 |
-
},
|
| 1680 |
-
{
|
| 1681 |
-
"epoch": 0.9184629803186504,
|
| 1682 |
-
"grad_norm": 1.9895987510681152,
|
| 1683 |
-
"learning_rate": 4.4259606373008436e-05,
|
| 1684 |
-
"loss": 0.3354,
|
| 1685 |
-
"step": 1960
|
| 1686 |
-
},
|
| 1687 |
-
{
|
| 1688 |
-
"epoch": 0.9231490159325211,
|
| 1689 |
-
"grad_norm": 1.5158389806747437,
|
| 1690 |
-
"learning_rate": 4.4230318650421745e-05,
|
| 1691 |
-
"loss": 0.2987,
|
| 1692 |
-
"step": 1970
|
| 1693 |
-
},
|
| 1694 |
-
{
|
| 1695 |
-
"epoch": 0.9278350515463918,
|
| 1696 |
-
"grad_norm": 1.498703956604004,
|
| 1697 |
-
"learning_rate": 4.4201030927835053e-05,
|
| 1698 |
-
"loss": 0.2703,
|
| 1699 |
-
"step": 1980
|
| 1700 |
-
},
|
| 1701 |
-
{
|
| 1702 |
-
"epoch": 0.9325210871602624,
|
| 1703 |
-
"grad_norm": 1.609595537185669,
|
| 1704 |
-
"learning_rate": 4.417174320524836e-05,
|
| 1705 |
-
"loss": 0.3511,
|
| 1706 |
-
"step": 1990
|
| 1707 |
-
},
|
| 1708 |
-
{
|
| 1709 |
-
"epoch": 0.9372071227741331,
|
| 1710 |
-
"grad_norm": 1.5775402784347534,
|
| 1711 |
-
"learning_rate": 4.414245548266167e-05,
|
| 1712 |
-
"loss": 0.3287,
|
| 1713 |
-
"step": 2000
|
| 1714 |
-
},
|
| 1715 |
-
{
|
| 1716 |
-
"epoch": 0.9372071227741331,
|
| 1717 |
-
"eval_loss": 0.05509389936923981,
|
| 1718 |
-
"eval_pearson_cosine": 0.8207199970543648,
|
| 1719 |
-
"eval_pearson_dot": 0.7620359249952955,
|
| 1720 |
-
"eval_pearson_euclidean": 0.8229490445973937,
|
| 1721 |
-
"eval_pearson_manhattan": 0.8238150410767417,
|
| 1722 |
-
"eval_runtime": 6.0216,
|
| 1723 |
-
"eval_samples_per_second": 249.103,
|
| 1724 |
-
"eval_spearman_cosine": 0.8243786401215701,
|
| 1725 |
-
"eval_spearman_dot": 0.7569267777069086,
|
| 1726 |
-
"eval_spearman_euclidean": 0.8286694906125902,
|
| 1727 |
-
"eval_spearman_manhattan": 0.8295593456330026,
|
| 1728 |
-
"eval_steps_per_second": 15.61,
|
| 1729 |
-
"step": 2000
|
| 1730 |
-
},
|
| 1731 |
-
{
|
| 1732 |
-
"epoch": 0.9418931583880038,
|
| 1733 |
-
"grad_norm": 1.1670805215835571,
|
| 1734 |
-
"learning_rate": 4.411316776007498e-05,
|
| 1735 |
-
"loss": 0.2544,
|
| 1736 |
-
"step": 2010
|
| 1737 |
-
},
|
| 1738 |
-
{
|
| 1739 |
-
"epoch": 0.9465791940018744,
|
| 1740 |
-
"grad_norm": 1.6743440628051758,
|
| 1741 |
-
"learning_rate": 4.408388003748829e-05,
|
| 1742 |
-
"loss": 0.2866,
|
| 1743 |
-
"step": 2020
|
| 1744 |
-
},
|
| 1745 |
-
{
|
| 1746 |
-
"epoch": 0.9512652296157451,
|
| 1747 |
-
"grad_norm": 1.5439425706863403,
|
| 1748 |
-
"learning_rate": 4.4054592314901596e-05,
|
| 1749 |
-
"loss": 0.286,
|
| 1750 |
-
"step": 2030
|
| 1751 |
-
},
|
| 1752 |
-
{
|
| 1753 |
-
"epoch": 0.9559512652296157,
|
| 1754 |
-
"grad_norm": 1.317328691482544,
|
| 1755 |
-
"learning_rate": 4.4025304592314905e-05,
|
| 1756 |
-
"loss": 0.2674,
|
| 1757 |
-
"step": 2040
|
| 1758 |
-
},
|
| 1759 |
-
{
|
| 1760 |
-
"epoch": 0.9606373008434864,
|
| 1761 |
-
"grad_norm": 1.3477058410644531,
|
| 1762 |
-
"learning_rate": 4.399601686972821e-05,
|
| 1763 |
-
"loss": 0.3005,
|
| 1764 |
-
"step": 2050
|
| 1765 |
-
},
|
| 1766 |
-
{
|
| 1767 |
-
"epoch": 0.9653233364573571,
|
| 1768 |
-
"grad_norm": 1.8991141319274902,
|
| 1769 |
-
"learning_rate": 4.396672914714152e-05,
|
| 1770 |
-
"loss": 0.2924,
|
| 1771 |
-
"step": 2060
|
| 1772 |
-
},
|
| 1773 |
-
{
|
| 1774 |
-
"epoch": 0.9700093720712277,
|
| 1775 |
-
"grad_norm": 1.406160593032837,
|
| 1776 |
-
"learning_rate": 4.3937441424554824e-05,
|
| 1777 |
-
"loss": 0.2977,
|
| 1778 |
-
"step": 2070
|
| 1779 |
-
},
|
| 1780 |
-
{
|
| 1781 |
-
"epoch": 0.9746954076850984,
|
| 1782 |
-
"grad_norm": 1.7128149271011353,
|
| 1783 |
-
"learning_rate": 4.390815370196814e-05,
|
| 1784 |
-
"loss": 0.3003,
|
| 1785 |
-
"step": 2080
|
| 1786 |
-
},
|
| 1787 |
-
{
|
| 1788 |
-
"epoch": 0.979381443298969,
|
| 1789 |
-
"grad_norm": 1.2714091539382935,
|
| 1790 |
-
"learning_rate": 4.387886597938145e-05,
|
| 1791 |
-
"loss": 0.2807,
|
| 1792 |
-
"step": 2090
|
| 1793 |
-
},
|
| 1794 |
-
{
|
| 1795 |
-
"epoch": 0.9840674789128397,
|
| 1796 |
-
"grad_norm": 1.1022099256515503,
|
| 1797 |
-
"learning_rate": 4.3849578256794756e-05,
|
| 1798 |
-
"loss": 0.2889,
|
| 1799 |
-
"step": 2100
|
| 1800 |
-
},
|
| 1801 |
-
{
|
| 1802 |
-
"epoch": 0.9840674789128397,
|
| 1803 |
-
"eval_loss": 0.04628630727529526,
|
| 1804 |
-
"eval_pearson_cosine": 0.8264380630920911,
|
| 1805 |
-
"eval_pearson_dot": 0.765122231613347,
|
| 1806 |
-
"eval_pearson_euclidean": 0.8297712335935188,
|
| 1807 |
-
"eval_pearson_manhattan": 0.8303968843751832,
|
| 1808 |
-
"eval_runtime": 5.8458,
|
| 1809 |
-
"eval_samples_per_second": 256.595,
|
| 1810 |
-
"eval_spearman_cosine": 0.8280517554752953,
|
| 1811 |
-
"eval_spearman_dot": 0.7620029087019149,
|
| 1812 |
-
"eval_spearman_euclidean": 0.8362970969318975,
|
| 1813 |
-
"eval_spearman_manhattan": 0.8368418361210694,
|
| 1814 |
-
"eval_steps_per_second": 16.08,
|
| 1815 |
-
"step": 2100
|
| 1816 |
-
},
|
| 1817 |
-
{
|
| 1818 |
-
"epoch": 0.9887535145267105,
|
| 1819 |
-
"grad_norm": 1.5780407190322876,
|
| 1820 |
-
"learning_rate": 4.3820290534208064e-05,
|
| 1821 |
-
"loss": 0.3019,
|
| 1822 |
-
"step": 2110
|
| 1823 |
-
},
|
| 1824 |
-
{
|
| 1825 |
-
"epoch": 0.993439550140581,
|
| 1826 |
-
"grad_norm": 1.5804523229599,
|
| 1827 |
-
"learning_rate": 4.3791002811621366e-05,
|
| 1828 |
-
"loss": 0.254,
|
| 1829 |
-
"step": 2120
|
| 1830 |
-
},
|
| 1831 |
-
{
|
| 1832 |
-
"epoch": 0.9981255857544518,
|
| 1833 |
-
"grad_norm": 1.5754133462905884,
|
| 1834 |
-
"learning_rate": 4.3761715089034675e-05,
|
| 1835 |
-
"loss": 0.2948,
|
| 1836 |
-
"step": 2130
|
| 1837 |
}
|
| 1838 |
],
|
| 1839 |
"logging_steps": 10,
|
|
@@ -1848,7 +371,7 @@
|
|
| 1848 |
"should_evaluate": false,
|
| 1849 |
"should_log": false,
|
| 1850 |
"should_save": true,
|
| 1851 |
-
"should_training_stop":
|
| 1852 |
},
|
| 1853 |
"attributes": {}
|
| 1854 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.23430178069353327,
|
| 5 |
+
"eval_steps": 2.0,
|
| 6 |
+
"global_step": 500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.004686035613870665,
|
| 13 |
+
"grad_norm": 4.305652141571045,
|
| 14 |
"learning_rate": 4.997071227741331e-05,
|
| 15 |
+
"loss": 2.2476,
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.00937207122774133,
|
| 20 |
+
"grad_norm": 3.343132972717285,
|
| 21 |
"learning_rate": 4.994142455482662e-05,
|
| 22 |
+
"loss": 1.2208,
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.014058106841611996,
|
| 27 |
+
"grad_norm": 3.07961368560791,
|
| 28 |
"learning_rate": 4.991213683223993e-05,
|
| 29 |
+
"loss": 0.9737,
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.01874414245548266,
|
| 34 |
+
"grad_norm": 2.348374128341675,
|
| 35 |
"learning_rate": 4.9882849109653237e-05,
|
| 36 |
+
"loss": 0.8232,
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.023430178069353328,
|
| 41 |
+
"grad_norm": 2.149184465408325,
|
| 42 |
"learning_rate": 4.9853561387066545e-05,
|
| 43 |
+
"loss": 0.7357,
|
| 44 |
"step": 50
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.028116213683223992,
|
| 48 |
+
"grad_norm": 2.793274164199829,
|
| 49 |
"learning_rate": 4.9824273664479854e-05,
|
| 50 |
+
"loss": 0.6115,
|
| 51 |
"step": 60
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.03280224929709466,
|
| 55 |
+
"grad_norm": 2.3996222019195557,
|
| 56 |
"learning_rate": 4.979498594189316e-05,
|
| 57 |
+
"loss": 0.5681,
|
| 58 |
"step": 70
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.03748828491096532,
|
| 62 |
+
"grad_norm": 2.7164816856384277,
|
| 63 |
"learning_rate": 4.9765698219306464e-05,
|
| 64 |
+
"loss": 0.6132,
|
| 65 |
"step": 80
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.04217432052483599,
|
| 69 |
+
"grad_norm": 2.3677186965942383,
|
| 70 |
"learning_rate": 4.973641049671978e-05,
|
| 71 |
+
"loss": 0.5407,
|
| 72 |
"step": 90
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.046860356138706656,
|
| 76 |
+
"grad_norm": 2.1071460247039795,
|
| 77 |
"learning_rate": 4.970712277413309e-05,
|
| 78 |
+
"loss": 0.5947,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 79 |
"step": 100
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.05154639175257732,
|
| 83 |
+
"grad_norm": 2.226364850997925,
|
| 84 |
"learning_rate": 4.9677835051546396e-05,
|
| 85 |
+
"loss": 0.532,
|
| 86 |
"step": 110
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.056232427366447985,
|
| 90 |
+
"grad_norm": 1.9689487218856812,
|
| 91 |
"learning_rate": 4.9648547328959705e-05,
|
| 92 |
+
"loss": 0.4874,
|
| 93 |
"step": 120
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.06091846298031865,
|
| 97 |
+
"grad_norm": 2.2253146171569824,
|
| 98 |
"learning_rate": 4.961925960637301e-05,
|
| 99 |
+
"loss": 0.5034,
|
| 100 |
"step": 130
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.06560449859418932,
|
| 104 |
+
"grad_norm": 1.8077352046966553,
|
| 105 |
"learning_rate": 4.9589971883786315e-05,
|
| 106 |
+
"loss": 0.4774,
|
| 107 |
"step": 140
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.07029053420805999,
|
| 111 |
+
"grad_norm": 1.9207241535186768,
|
| 112 |
"learning_rate": 4.956068416119963e-05,
|
| 113 |
+
"loss": 0.5066,
|
| 114 |
"step": 150
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.07497656982193064,
|
| 118 |
+
"grad_norm": 2.000474691390991,
|
| 119 |
"learning_rate": 4.953139643861294e-05,
|
| 120 |
+
"loss": 0.5007,
|
| 121 |
"step": 160
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.07966260543580131,
|
| 125 |
+
"grad_norm": 2.3965399265289307,
|
| 126 |
"learning_rate": 4.950210871602625e-05,
|
| 127 |
+
"loss": 0.4752,
|
| 128 |
"step": 170
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.08434864104967198,
|
| 132 |
+
"grad_norm": 2.164004325866699,
|
| 133 |
"learning_rate": 4.947282099343955e-05,
|
| 134 |
+
"loss": 0.4241,
|
| 135 |
"step": 180
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.08903467666354264,
|
| 139 |
+
"grad_norm": 2.105633497238159,
|
| 140 |
"learning_rate": 4.944353327085286e-05,
|
| 141 |
+
"loss": 0.4975,
|
| 142 |
"step": 190
|
| 143 |
},
|
| 144 |
{
|
| 145 |
"epoch": 0.09372071227741331,
|
| 146 |
+
"grad_norm": 2.2528748512268066,
|
| 147 |
"learning_rate": 4.9414245548266166e-05,
|
| 148 |
+
"loss": 0.4117,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 149 |
"step": 200
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.09840674789128398,
|
| 153 |
+
"grad_norm": 2.4957821369171143,
|
| 154 |
"learning_rate": 4.938495782567948e-05,
|
| 155 |
+
"loss": 0.4842,
|
| 156 |
"step": 210
|
| 157 |
},
|
| 158 |
{
|
| 159 |
"epoch": 0.10309278350515463,
|
| 160 |
+
"grad_norm": 1.6977312564849854,
|
| 161 |
"learning_rate": 4.935567010309279e-05,
|
| 162 |
+
"loss": 0.4407,
|
| 163 |
"step": 220
|
| 164 |
},
|
| 165 |
{
|
| 166 |
"epoch": 0.1077788191190253,
|
| 167 |
+
"grad_norm": 2.498537302017212,
|
| 168 |
"learning_rate": 4.932638238050609e-05,
|
| 169 |
+
"loss": 0.4402,
|
| 170 |
"step": 230
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.11246485473289597,
|
| 174 |
+
"grad_norm": 1.9550998210906982,
|
| 175 |
"learning_rate": 4.92970946579194e-05,
|
| 176 |
+
"loss": 0.4062,
|
| 177 |
"step": 240
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.11715089034676664,
|
| 181 |
+
"grad_norm": 1.692822813987732,
|
| 182 |
"learning_rate": 4.926780693533271e-05,
|
| 183 |
+
"loss": 0.4135,
|
| 184 |
"step": 250
|
| 185 |
},
|
| 186 |
{
|
| 187 |
"epoch": 0.1218369259606373,
|
| 188 |
+
"grad_norm": 1.936856985092163,
|
| 189 |
"learning_rate": 4.923851921274602e-05,
|
| 190 |
+
"loss": 0.4518,
|
| 191 |
"step": 260
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.12652296157450796,
|
| 195 |
+
"grad_norm": 2.509472370147705,
|
| 196 |
"learning_rate": 4.920923149015933e-05,
|
| 197 |
+
"loss": 0.4065,
|
| 198 |
"step": 270
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"epoch": 0.13120899718837864,
|
| 202 |
+
"grad_norm": 1.993790864944458,
|
| 203 |
"learning_rate": 4.9179943767572635e-05,
|
| 204 |
+
"loss": 0.4252,
|
| 205 |
"step": 280
|
| 206 |
},
|
| 207 |
{
|
| 208 |
"epoch": 0.1358950328022493,
|
| 209 |
+
"grad_norm": 2.542051315307617,
|
| 210 |
"learning_rate": 4.9150656044985943e-05,
|
| 211 |
+
"loss": 0.4342,
|
| 212 |
"step": 290
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.14058106841611998,
|
| 216 |
+
"grad_norm": 2.0401690006256104,
|
| 217 |
"learning_rate": 4.912136832239925e-05,
|
| 218 |
+
"loss": 0.4005,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 219 |
"step": 300
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.14526710402999063,
|
| 223 |
+
"grad_norm": 2.234344005584717,
|
| 224 |
"learning_rate": 4.909208059981256e-05,
|
| 225 |
+
"loss": 0.4044,
|
| 226 |
"step": 310
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"epoch": 0.14995313964386128,
|
| 230 |
+
"grad_norm": 2.4048752784729004,
|
| 231 |
"learning_rate": 4.906279287722587e-05,
|
| 232 |
+
"loss": 0.3832,
|
| 233 |
"step": 320
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.15463917525773196,
|
| 237 |
+
"grad_norm": 2.027322769165039,
|
| 238 |
"learning_rate": 4.903350515463918e-05,
|
| 239 |
+
"loss": 0.4425,
|
| 240 |
"step": 330
|
| 241 |
},
|
| 242 |
{
|
| 243 |
"epoch": 0.15932521087160262,
|
| 244 |
+
"grad_norm": 1.7849469184875488,
|
| 245 |
"learning_rate": 4.9004217432052486e-05,
|
| 246 |
+
"loss": 0.4034,
|
| 247 |
"step": 340
|
| 248 |
},
|
| 249 |
{
|
| 250 |
"epoch": 0.1640112464854733,
|
| 251 |
+
"grad_norm": 1.865513563156128,
|
| 252 |
"learning_rate": 4.8974929709465795e-05,
|
| 253 |
+
"loss": 0.4256,
|
| 254 |
"step": 350
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.16869728209934395,
|
| 258 |
+
"grad_norm": 2.17820143699646,
|
| 259 |
"learning_rate": 4.89456419868791e-05,
|
| 260 |
+
"loss": 0.388,
|
| 261 |
"step": 360
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.1733833177132146,
|
| 265 |
+
"grad_norm": 2.6553549766540527,
|
| 266 |
"learning_rate": 4.891635426429241e-05,
|
| 267 |
+
"loss": 0.3645,
|
| 268 |
"step": 370
|
| 269 |
},
|
| 270 |
{
|
| 271 |
"epoch": 0.1780693533270853,
|
| 272 |
+
"grad_norm": 2.155061960220337,
|
| 273 |
"learning_rate": 4.888706654170572e-05,
|
| 274 |
+
"loss": 0.3819,
|
| 275 |
"step": 380
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 0.18275538894095594,
|
| 279 |
+
"grad_norm": 1.9706778526306152,
|
| 280 |
"learning_rate": 4.885777881911903e-05,
|
| 281 |
+
"loss": 0.3959,
|
| 282 |
"step": 390
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 0.18744142455482662,
|
| 286 |
+
"grad_norm": 2.111262321472168,
|
| 287 |
"learning_rate": 4.882849109653234e-05,
|
| 288 |
+
"loss": 0.3929,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 289 |
"step": 400
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 0.19212746016869728,
|
| 293 |
+
"grad_norm": 2.65875244140625,
|
| 294 |
"learning_rate": 4.8799203373945646e-05,
|
| 295 |
+
"loss": 0.4151,
|
| 296 |
"step": 410
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.19681349578256796,
|
| 300 |
+
"grad_norm": 1.8690752983093262,
|
| 301 |
"learning_rate": 4.8769915651358954e-05,
|
| 302 |
+
"loss": 0.3823,
|
| 303 |
"step": 420
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.2014995313964386,
|
| 307 |
+
"grad_norm": 2.35809326171875,
|
| 308 |
"learning_rate": 4.8740627928772256e-05,
|
| 309 |
+
"loss": 0.4079,
|
| 310 |
"step": 430
|
| 311 |
},
|
| 312 |
{
|
| 313 |
"epoch": 0.20618556701030927,
|
| 314 |
+
"grad_norm": 1.4293204545974731,
|
| 315 |
"learning_rate": 4.871134020618557e-05,
|
| 316 |
+
"loss": 0.3732,
|
| 317 |
"step": 440
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 0.21087160262417995,
|
| 321 |
+
"grad_norm": 2.2345097064971924,
|
| 322 |
"learning_rate": 4.868205248359888e-05,
|
| 323 |
+
"loss": 0.3513,
|
| 324 |
"step": 450
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"epoch": 0.2155576382380506,
|
| 328 |
+
"grad_norm": 1.7603412866592407,
|
| 329 |
"learning_rate": 4.865276476101219e-05,
|
| 330 |
+
"loss": 0.3872,
|
| 331 |
"step": 460
|
| 332 |
},
|
| 333 |
{
|
| 334 |
"epoch": 0.22024367385192128,
|
| 335 |
+
"grad_norm": 1.8551238775253296,
|
| 336 |
"learning_rate": 4.86234770384255e-05,
|
| 337 |
+
"loss": 0.377,
|
| 338 |
"step": 470
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.22492970946579194,
|
| 342 |
+
"grad_norm": 2.2718453407287598,
|
| 343 |
"learning_rate": 4.85941893158388e-05,
|
| 344 |
+
"loss": 0.3466,
|
| 345 |
"step": 480
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.2296157450796626,
|
| 349 |
+
"grad_norm": 2.021726608276367,
|
| 350 |
"learning_rate": 4.856490159325211e-05,
|
| 351 |
+
"loss": 0.3778,
|
| 352 |
"step": 490
|
| 353 |
},
|
| 354 |
{
|
| 355 |
"epoch": 0.23430178069353327,
|
| 356 |
+
"grad_norm": 1.4741500616073608,
|
| 357 |
"learning_rate": 4.853561387066542e-05,
|
| 358 |
+
"loss": 0.3862,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 359 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 360 |
}
|
| 361 |
],
|
| 362 |
"logging_steps": 10,
|
|
|
|
| 371 |
"should_evaluate": false,
|
| 372 |
"should_log": false,
|
| 373 |
"should_save": true,
|
| 374 |
+
"should_training_stop": false
|
| 375 |
},
|
| 376 |
"attributes": {}
|
| 377 |
}
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5368
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7be9e20c2c0889091baaa0347720d7888707c1703a6f97836e6beae35fe15fe
|
| 3 |
size 5368
|