Commit
·
45c5b3c
1
Parent(s):
873008c
Upload folder using huggingface_hub
Browse files- README.md +31 -13
- adapter_model.bin +2 -2
- checkpoint-300/README.md +7 -6
- checkpoint-300/optimizer.pt +2 -2
- checkpoint-300/rng_state.pth +1 -1
- checkpoint-300/scheduler.pt +1 -1
- checkpoint-300/trainer_state.json +888 -885
- checkpoint-300/training_args.bin +1 -1
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -1,16 +1,34 @@
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
- yahma/alpaca-cleaned
|
| 4 |
-
language:
|
| 5 |
-
- en
|
| 6 |
-
license: apache-2.0
|
| 7 |
-
tags:
|
| 8 |
-
- pytorch
|
| 9 |
-
- causal-lm
|
| 10 |
-
- llama2
|
| 11 |
-
- fine-tuning
|
| 12 |
-
- alpaca
|
| 13 |
-
|
| 14 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
-
|
|
|
|
| 1 |
---
|
| 2 |
+
library_name: peft
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
+
- quant_method: bitsandbytes
|
| 9 |
+
- load_in_8bit: False
|
| 10 |
+
- load_in_4bit: True
|
| 11 |
+
- llm_int8_threshold: 6.0
|
| 12 |
+
- llm_int8_skip_modules: None
|
| 13 |
+
- llm_int8_enable_fp32_cpu_offload: True
|
| 14 |
+
- llm_int8_has_fp16_weight: False
|
| 15 |
+
- bnb_4bit_quant_type: nf4
|
| 16 |
+
- bnb_4bit_use_double_quant: True
|
| 17 |
+
- bnb_4bit_compute_dtype: bfloat16
|
| 18 |
+
|
| 19 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 20 |
+
- quant_method: bitsandbytes
|
| 21 |
+
- load_in_8bit: False
|
| 22 |
+
- load_in_4bit: True
|
| 23 |
+
- llm_int8_threshold: 6.0
|
| 24 |
+
- llm_int8_skip_modules: None
|
| 25 |
+
- llm_int8_enable_fp32_cpu_offload: True
|
| 26 |
+
- llm_int8_has_fp16_weight: False
|
| 27 |
+
- bnb_4bit_quant_type: nf4
|
| 28 |
+
- bnb_4bit_use_double_quant: True
|
| 29 |
+
- bnb_4bit_compute_dtype: bfloat16
|
| 30 |
+
### Framework versions
|
| 31 |
+
|
| 32 |
+
- PEFT 0.5.0.dev0
|
| 33 |
|
| 34 |
+
- PEFT 0.5.0.dev0
|
adapter_model.bin
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:e5e1621f48d9ad8feb1d6d31050275f0aafd080c5c07153301fe2f48411f4406
|
| 3 |
+
size 443
|
checkpoint-300/README.md
CHANGED
|
@@ -5,15 +5,16 @@ library_name: peft
|
|
| 5 |
|
| 6 |
|
| 7 |
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
-
-
|
| 9 |
-
-
|
|
|
|
| 10 |
- llm_int8_threshold: 6.0
|
| 11 |
- llm_int8_skip_modules: None
|
| 12 |
-
- llm_int8_enable_fp32_cpu_offload:
|
| 13 |
- llm_int8_has_fp16_weight: False
|
| 14 |
-
- bnb_4bit_quant_type:
|
| 15 |
-
- bnb_4bit_use_double_quant:
|
| 16 |
-
- bnb_4bit_compute_dtype:
|
| 17 |
### Framework versions
|
| 18 |
|
| 19 |
|
|
|
|
| 5 |
|
| 6 |
|
| 7 |
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
+
- quant_method: bitsandbytes
|
| 9 |
+
- load_in_8bit: False
|
| 10 |
+
- load_in_4bit: True
|
| 11 |
- llm_int8_threshold: 6.0
|
| 12 |
- llm_int8_skip_modules: None
|
| 13 |
+
- llm_int8_enable_fp32_cpu_offload: True
|
| 14 |
- llm_int8_has_fp16_weight: False
|
| 15 |
+
- bnb_4bit_quant_type: nf4
|
| 16 |
+
- bnb_4bit_use_double_quant: True
|
| 17 |
+
- bnb_4bit_compute_dtype: bfloat16
|
| 18 |
### Framework versions
|
| 19 |
|
| 20 |
|
checkpoint-300/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:be9efd61012bf13cedff400fcbb8a2a70e1d6dfb96b4b9c96b0f2082b4421403
|
| 3 |
+
size 19991557
|
checkpoint-300/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14575
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d58a2d49c8bd88ae73de10c23b44dda16a3841706f663470cc7b0116797d3b2
|
| 3 |
size 14575
|
checkpoint-300/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 627
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:503d141bd03d55a69c94b428e328d3b060b988ace448865bb4b7d17216ac0c68
|
| 3 |
size 627
|
checkpoint-300/trainer_state.json
CHANGED
|
@@ -1,7 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
|
|
|
| 5 |
"global_step": 300,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
|
@@ -9,1808 +10,1810 @@
|
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.0,
|
| 12 |
-
"learning_rate": 3.333333333333333e-
|
| 13 |
-
"loss": 1.
|
| 14 |
"step": 1
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.0,
|
| 18 |
-
"learning_rate": 6.666666666666666e-
|
| 19 |
-
"loss": 1.
|
| 20 |
"step": 2
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.0,
|
| 24 |
-
"learning_rate":
|
| 25 |
-
"loss": 1.
|
| 26 |
"step": 3
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.0,
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss": 1.
|
| 32 |
"step": 4
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.0,
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss": 1.
|
| 38 |
"step": 5
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.0,
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 1.
|
| 44 |
"step": 6
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.0,
|
| 48 |
-
"learning_rate":
|
| 49 |
-
"loss": 1.
|
| 50 |
"step": 7
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.0,
|
| 54 |
-
"learning_rate":
|
| 55 |
-
"loss": 1.
|
| 56 |
"step": 8
|
| 57 |
},
|
| 58 |
{
|
| 59 |
-
"epoch": 0.
|
| 60 |
-
"learning_rate":
|
| 61 |
-
"loss": 1.
|
| 62 |
"step": 9
|
| 63 |
},
|
| 64 |
{
|
| 65 |
-
"epoch": 0.
|
| 66 |
-
"learning_rate":
|
| 67 |
-
"loss": 1.
|
| 68 |
"step": 10
|
| 69 |
},
|
| 70 |
{
|
| 71 |
-
"epoch": 0.
|
| 72 |
-
"learning_rate":
|
| 73 |
-
"loss": 1.
|
| 74 |
"step": 11
|
| 75 |
},
|
| 76 |
{
|
| 77 |
-
"epoch": 0.
|
| 78 |
-
"learning_rate":
|
| 79 |
-
"loss":
|
| 80 |
"step": 12
|
| 81 |
},
|
| 82 |
{
|
| 83 |
-
"epoch": 0.
|
| 84 |
-
"learning_rate":
|
| 85 |
-
"loss":
|
| 86 |
"step": 13
|
| 87 |
},
|
| 88 |
{
|
| 89 |
-
"epoch": 0.
|
| 90 |
-
"learning_rate":
|
| 91 |
-
"loss":
|
| 92 |
"step": 14
|
| 93 |
},
|
| 94 |
{
|
| 95 |
-
"epoch": 0.
|
| 96 |
-
"learning_rate":
|
| 97 |
-
"loss":
|
| 98 |
"step": 15
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 0.
|
| 102 |
-
"learning_rate":
|
| 103 |
-
"loss":
|
| 104 |
"step": 16
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.01,
|
| 108 |
-
"learning_rate":
|
| 109 |
-
"loss":
|
| 110 |
"step": 17
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.01,
|
| 114 |
-
"learning_rate":
|
| 115 |
-
"loss":
|
| 116 |
"step": 18
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.01,
|
| 120 |
-
"learning_rate":
|
| 121 |
-
"loss":
|
| 122 |
"step": 19
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.01,
|
| 126 |
-
"learning_rate":
|
| 127 |
-
"loss":
|
| 128 |
"step": 20
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.01,
|
| 132 |
-
"learning_rate":
|
| 133 |
-
"loss":
|
| 134 |
"step": 21
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.01,
|
| 138 |
-
"learning_rate":
|
| 139 |
-
"loss":
|
| 140 |
"step": 22
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.01,
|
| 144 |
-
"learning_rate":
|
| 145 |
-
"loss":
|
| 146 |
"step": 23
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.01,
|
| 150 |
-
"learning_rate":
|
| 151 |
-
"loss":
|
| 152 |
"step": 24
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch": 0.
|
| 156 |
-
"learning_rate":
|
| 157 |
-
"loss":
|
| 158 |
"step": 25
|
| 159 |
},
|
| 160 |
{
|
| 161 |
-
"epoch": 0.
|
| 162 |
-
"learning_rate":
|
| 163 |
-
"loss":
|
| 164 |
"step": 26
|
| 165 |
},
|
| 166 |
{
|
| 167 |
-
"epoch": 0.
|
| 168 |
-
"learning_rate":
|
| 169 |
-
"loss":
|
| 170 |
"step": 27
|
| 171 |
},
|
| 172 |
{
|
| 173 |
-
"epoch": 0.
|
| 174 |
-
"learning_rate":
|
| 175 |
-
"loss":
|
| 176 |
"step": 28
|
| 177 |
},
|
| 178 |
{
|
| 179 |
-
"epoch": 0.
|
| 180 |
-
"learning_rate":
|
| 181 |
-
"loss":
|
| 182 |
"step": 29
|
| 183 |
},
|
| 184 |
{
|
| 185 |
-
"epoch": 0.
|
| 186 |
-
"learning_rate":
|
| 187 |
-
"loss":
|
| 188 |
"step": 30
|
| 189 |
},
|
| 190 |
{
|
| 191 |
-
"epoch": 0.
|
| 192 |
-
"learning_rate":
|
| 193 |
-
"loss":
|
| 194 |
"step": 31
|
| 195 |
},
|
| 196 |
{
|
| 197 |
-
"epoch": 0.
|
| 198 |
-
"learning_rate":
|
| 199 |
-
"loss":
|
| 200 |
"step": 32
|
| 201 |
},
|
| 202 |
{
|
| 203 |
-
"epoch": 0.
|
| 204 |
-
"learning_rate":
|
| 205 |
-
"loss":
|
| 206 |
"step": 33
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch": 0.
|
| 210 |
-
"learning_rate":
|
| 211 |
-
"loss":
|
| 212 |
"step": 34
|
| 213 |
},
|
| 214 |
{
|
| 215 |
-
"epoch": 0.
|
| 216 |
-
"learning_rate":
|
| 217 |
-
"loss":
|
| 218 |
"step": 35
|
| 219 |
},
|
| 220 |
{
|
| 221 |
-
"epoch": 0.
|
| 222 |
-
"learning_rate":
|
| 223 |
-
"loss":
|
| 224 |
"step": 36
|
| 225 |
},
|
| 226 |
{
|
| 227 |
-
"epoch": 0.
|
| 228 |
-
"learning_rate":
|
| 229 |
-
"loss":
|
| 230 |
"step": 37
|
| 231 |
},
|
| 232 |
{
|
| 233 |
-
"epoch": 0.
|
| 234 |
-
"learning_rate":
|
| 235 |
-
"loss":
|
| 236 |
"step": 38
|
| 237 |
},
|
| 238 |
{
|
| 239 |
-
"epoch": 0.
|
| 240 |
-
"learning_rate":
|
| 241 |
-
"loss":
|
| 242 |
"step": 39
|
| 243 |
},
|
| 244 |
{
|
| 245 |
-
"epoch": 0.
|
| 246 |
-
"learning_rate":
|
| 247 |
-
"loss":
|
| 248 |
"step": 40
|
| 249 |
},
|
| 250 |
{
|
| 251 |
-
"epoch": 0.
|
| 252 |
-
"learning_rate":
|
| 253 |
-
"loss":
|
| 254 |
"step": 41
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch": 0.
|
| 258 |
-
"learning_rate":
|
| 259 |
-
"loss":
|
| 260 |
"step": 42
|
| 261 |
},
|
| 262 |
{
|
| 263 |
-
"epoch": 0.
|
| 264 |
-
"learning_rate":
|
| 265 |
-
"loss":
|
| 266 |
"step": 43
|
| 267 |
},
|
| 268 |
{
|
| 269 |
-
"epoch": 0.
|
| 270 |
-
"learning_rate":
|
| 271 |
-
"loss":
|
| 272 |
"step": 44
|
| 273 |
},
|
| 274 |
{
|
| 275 |
-
"epoch": 0.
|
| 276 |
-
"learning_rate":
|
| 277 |
-
"loss":
|
| 278 |
"step": 45
|
| 279 |
},
|
| 280 |
{
|
| 281 |
-
"epoch": 0.
|
| 282 |
-
"learning_rate":
|
| 283 |
-
"loss":
|
| 284 |
"step": 46
|
| 285 |
},
|
| 286 |
{
|
| 287 |
-
"epoch": 0.
|
| 288 |
-
"learning_rate":
|
| 289 |
-
"loss":
|
| 290 |
"step": 47
|
| 291 |
},
|
| 292 |
{
|
| 293 |
-
"epoch": 0.
|
| 294 |
-
"learning_rate":
|
| 295 |
-
"loss":
|
| 296 |
"step": 48
|
| 297 |
},
|
| 298 |
{
|
| 299 |
-
"epoch": 0.
|
| 300 |
-
"learning_rate":
|
| 301 |
-
"loss":
|
| 302 |
"step": 49
|
| 303 |
},
|
| 304 |
{
|
| 305 |
-
"epoch": 0.
|
| 306 |
-
"learning_rate":
|
| 307 |
-
"loss":
|
| 308 |
"step": 50
|
| 309 |
},
|
| 310 |
{
|
| 311 |
-
"epoch": 0.
|
| 312 |
-
"learning_rate":
|
| 313 |
-
"loss":
|
| 314 |
"step": 51
|
| 315 |
},
|
| 316 |
{
|
| 317 |
-
"epoch": 0.
|
| 318 |
-
"learning_rate":
|
| 319 |
-
"loss":
|
| 320 |
"step": 52
|
| 321 |
},
|
| 322 |
{
|
| 323 |
-
"epoch": 0.
|
| 324 |
-
"learning_rate":
|
| 325 |
-
"loss":
|
| 326 |
"step": 53
|
| 327 |
},
|
| 328 |
{
|
| 329 |
-
"epoch": 0.
|
| 330 |
-
"learning_rate":
|
| 331 |
-
"loss":
|
| 332 |
"step": 54
|
| 333 |
},
|
| 334 |
{
|
| 335 |
-
"epoch": 0.
|
| 336 |
-
"learning_rate":
|
| 337 |
-
"loss":
|
| 338 |
"step": 55
|
| 339 |
},
|
| 340 |
{
|
| 341 |
-
"epoch": 0.
|
| 342 |
-
"learning_rate":
|
| 343 |
-
"loss":
|
| 344 |
"step": 56
|
| 345 |
},
|
| 346 |
{
|
| 347 |
-
"epoch": 0.
|
| 348 |
-
"learning_rate":
|
| 349 |
-
"loss":
|
| 350 |
"step": 57
|
| 351 |
},
|
| 352 |
{
|
| 353 |
-
"epoch": 0.
|
| 354 |
-
"learning_rate":
|
| 355 |
-
"loss":
|
| 356 |
"step": 58
|
| 357 |
},
|
| 358 |
{
|
| 359 |
-
"epoch": 0.
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"loss":
|
| 362 |
"step": 59
|
| 363 |
},
|
| 364 |
{
|
| 365 |
-
"epoch": 0.
|
| 366 |
-
"learning_rate":
|
| 367 |
-
"loss":
|
| 368 |
"step": 60
|
| 369 |
},
|
| 370 |
{
|
| 371 |
-
"epoch": 0.
|
| 372 |
-
"learning_rate":
|
| 373 |
-
"loss":
|
| 374 |
"step": 61
|
| 375 |
},
|
| 376 |
{
|
| 377 |
-
"epoch": 0.
|
| 378 |
-
"learning_rate":
|
| 379 |
-
"loss":
|
| 380 |
"step": 62
|
| 381 |
},
|
| 382 |
{
|
| 383 |
-
"epoch": 0.
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss":
|
| 386 |
"step": 63
|
| 387 |
},
|
| 388 |
{
|
| 389 |
-
"epoch": 0.
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"loss":
|
| 392 |
"step": 64
|
| 393 |
},
|
| 394 |
{
|
| 395 |
-
"epoch": 0.
|
| 396 |
-
"learning_rate":
|
| 397 |
-
"loss":
|
| 398 |
"step": 65
|
| 399 |
},
|
| 400 |
{
|
| 401 |
-
"epoch": 0.
|
| 402 |
-
"learning_rate":
|
| 403 |
-
"loss":
|
| 404 |
"step": 66
|
| 405 |
},
|
| 406 |
{
|
| 407 |
-
"epoch": 0.
|
| 408 |
-
"learning_rate":
|
| 409 |
-
"loss":
|
| 410 |
"step": 67
|
| 411 |
},
|
| 412 |
{
|
| 413 |
-
"epoch": 0.
|
| 414 |
-
"learning_rate":
|
| 415 |
-
"loss":
|
| 416 |
"step": 68
|
| 417 |
},
|
| 418 |
{
|
| 419 |
-
"epoch": 0.
|
| 420 |
-
"learning_rate":
|
| 421 |
-
"loss":
|
| 422 |
"step": 69
|
| 423 |
},
|
| 424 |
{
|
| 425 |
-
"epoch": 0.
|
| 426 |
-
"learning_rate":
|
| 427 |
-
"loss":
|
| 428 |
"step": 70
|
| 429 |
},
|
| 430 |
{
|
| 431 |
-
"epoch": 0.
|
| 432 |
-
"learning_rate":
|
| 433 |
-
"loss":
|
| 434 |
"step": 71
|
| 435 |
},
|
| 436 |
{
|
| 437 |
-
"epoch": 0.
|
| 438 |
-
"learning_rate":
|
| 439 |
-
"loss":
|
| 440 |
"step": 72
|
| 441 |
},
|
| 442 |
{
|
| 443 |
-
"epoch": 0.
|
| 444 |
-
"learning_rate":
|
| 445 |
-
"loss":
|
| 446 |
"step": 73
|
| 447 |
},
|
| 448 |
{
|
| 449 |
-
"epoch": 0.
|
| 450 |
-
"learning_rate":
|
| 451 |
-
"loss":
|
| 452 |
"step": 74
|
| 453 |
},
|
| 454 |
{
|
| 455 |
-
"epoch": 0.
|
| 456 |
-
"learning_rate":
|
| 457 |
-
"loss":
|
| 458 |
"step": 75
|
| 459 |
},
|
| 460 |
{
|
| 461 |
-
"epoch": 0.
|
| 462 |
-
"learning_rate":
|
| 463 |
-
"loss":
|
| 464 |
"step": 76
|
| 465 |
},
|
| 466 |
{
|
| 467 |
-
"epoch": 0.
|
| 468 |
-
"learning_rate":
|
| 469 |
-
"loss":
|
| 470 |
"step": 77
|
| 471 |
},
|
| 472 |
{
|
| 473 |
-
"epoch": 0.
|
| 474 |
-
"learning_rate":
|
| 475 |
-
"loss":
|
| 476 |
"step": 78
|
| 477 |
},
|
| 478 |
{
|
| 479 |
-
"epoch": 0.
|
| 480 |
-
"learning_rate":
|
| 481 |
-
"loss":
|
| 482 |
"step": 79
|
| 483 |
},
|
| 484 |
{
|
| 485 |
-
"epoch": 0.
|
| 486 |
-
"learning_rate":
|
| 487 |
-
"loss":
|
| 488 |
"step": 80
|
| 489 |
},
|
| 490 |
{
|
| 491 |
-
"epoch": 0.
|
| 492 |
-
"learning_rate":
|
| 493 |
-
"loss":
|
| 494 |
"step": 81
|
| 495 |
},
|
| 496 |
{
|
| 497 |
-
"epoch": 0.
|
| 498 |
-
"learning_rate":
|
| 499 |
-
"loss":
|
| 500 |
"step": 82
|
| 501 |
},
|
| 502 |
{
|
| 503 |
-
"epoch": 0.
|
| 504 |
-
"learning_rate":
|
| 505 |
-
"loss":
|
| 506 |
"step": 83
|
| 507 |
},
|
| 508 |
{
|
| 509 |
-
"epoch": 0.
|
| 510 |
-
"learning_rate":
|
| 511 |
-
"loss":
|
| 512 |
"step": 84
|
| 513 |
},
|
| 514 |
{
|
| 515 |
-
"epoch": 0.
|
| 516 |
-
"learning_rate":
|
| 517 |
-
"loss":
|
| 518 |
"step": 85
|
| 519 |
},
|
| 520 |
{
|
| 521 |
-
"epoch": 0.
|
| 522 |
-
"learning_rate":
|
| 523 |
-
"loss":
|
| 524 |
"step": 86
|
| 525 |
},
|
| 526 |
{
|
| 527 |
-
"epoch": 0.
|
| 528 |
-
"learning_rate":
|
| 529 |
-
"loss":
|
| 530 |
"step": 87
|
| 531 |
},
|
| 532 |
{
|
| 533 |
-
"epoch": 0.
|
| 534 |
-
"learning_rate":
|
| 535 |
-
"loss":
|
| 536 |
"step": 88
|
| 537 |
},
|
| 538 |
{
|
| 539 |
-
"epoch": 0.
|
| 540 |
-
"learning_rate":
|
| 541 |
-
"loss":
|
| 542 |
"step": 89
|
| 543 |
},
|
| 544 |
{
|
| 545 |
-
"epoch": 0.
|
| 546 |
-
"learning_rate":
|
| 547 |
-
"loss":
|
| 548 |
"step": 90
|
| 549 |
},
|
| 550 |
{
|
| 551 |
-
"epoch": 0.
|
| 552 |
-
"learning_rate":
|
| 553 |
-
"loss":
|
| 554 |
"step": 91
|
| 555 |
},
|
| 556 |
{
|
| 557 |
-
"epoch": 0.
|
| 558 |
-
"learning_rate":
|
| 559 |
-
"loss":
|
| 560 |
"step": 92
|
| 561 |
},
|
| 562 |
{
|
| 563 |
-
"epoch": 0.
|
| 564 |
-
"learning_rate":
|
| 565 |
-
"loss":
|
| 566 |
"step": 93
|
| 567 |
},
|
| 568 |
{
|
| 569 |
-
"epoch": 0.
|
| 570 |
-
"learning_rate":
|
| 571 |
-
"loss":
|
| 572 |
"step": 94
|
| 573 |
},
|
| 574 |
{
|
| 575 |
-
"epoch": 0.
|
| 576 |
-
"learning_rate":
|
| 577 |
-
"loss":
|
| 578 |
"step": 95
|
| 579 |
},
|
| 580 |
{
|
| 581 |
-
"epoch": 0.
|
| 582 |
-
"learning_rate":
|
| 583 |
-
"loss":
|
| 584 |
"step": 96
|
| 585 |
},
|
| 586 |
{
|
| 587 |
-
"epoch": 0.
|
| 588 |
-
"learning_rate":
|
| 589 |
-
"loss":
|
| 590 |
"step": 97
|
| 591 |
},
|
| 592 |
{
|
| 593 |
-
"epoch": 0.
|
| 594 |
-
"learning_rate":
|
| 595 |
-
"loss":
|
| 596 |
"step": 98
|
| 597 |
},
|
| 598 |
{
|
| 599 |
-
"epoch": 0.
|
| 600 |
-
"learning_rate":
|
| 601 |
-
"loss":
|
| 602 |
"step": 99
|
| 603 |
},
|
| 604 |
{
|
| 605 |
-
"epoch": 0.
|
| 606 |
-
"learning_rate":
|
| 607 |
-
"loss":
|
| 608 |
"step": 100
|
| 609 |
},
|
| 610 |
{
|
| 611 |
-
"epoch": 0.
|
| 612 |
-
"learning_rate":
|
| 613 |
-
"loss":
|
| 614 |
"step": 101
|
| 615 |
},
|
| 616 |
{
|
| 617 |
-
"epoch": 0.
|
| 618 |
-
"learning_rate":
|
| 619 |
-
"loss":
|
| 620 |
"step": 102
|
| 621 |
},
|
| 622 |
{
|
| 623 |
-
"epoch": 0.
|
| 624 |
-
"learning_rate":
|
| 625 |
-
"loss":
|
| 626 |
"step": 103
|
| 627 |
},
|
| 628 |
{
|
| 629 |
-
"epoch": 0.
|
| 630 |
-
"learning_rate":
|
| 631 |
-
"loss":
|
| 632 |
"step": 104
|
| 633 |
},
|
| 634 |
{
|
| 635 |
-
"epoch": 0.
|
| 636 |
-
"learning_rate":
|
| 637 |
-
"loss":
|
| 638 |
"step": 105
|
| 639 |
},
|
| 640 |
{
|
| 641 |
-
"epoch": 0.
|
| 642 |
-
"learning_rate":
|
| 643 |
-
"loss":
|
| 644 |
"step": 106
|
| 645 |
},
|
| 646 |
{
|
| 647 |
-
"epoch": 0.
|
| 648 |
-
"learning_rate":
|
| 649 |
-
"loss":
|
| 650 |
"step": 107
|
| 651 |
},
|
| 652 |
{
|
| 653 |
-
"epoch": 0.
|
| 654 |
-
"learning_rate":
|
| 655 |
-
"loss":
|
| 656 |
"step": 108
|
| 657 |
},
|
| 658 |
{
|
| 659 |
-
"epoch": 0.
|
| 660 |
-
"learning_rate":
|
| 661 |
-
"loss":
|
| 662 |
"step": 109
|
| 663 |
},
|
| 664 |
{
|
| 665 |
-
"epoch": 0.
|
| 666 |
-
"learning_rate":
|
| 667 |
-
"loss":
|
| 668 |
"step": 110
|
| 669 |
},
|
| 670 |
{
|
| 671 |
-
"epoch": 0.
|
| 672 |
-
"learning_rate":
|
| 673 |
-
"loss":
|
| 674 |
"step": 111
|
| 675 |
},
|
| 676 |
{
|
| 677 |
-
"epoch": 0.
|
| 678 |
-
"learning_rate":
|
| 679 |
-
"loss":
|
| 680 |
"step": 112
|
| 681 |
},
|
| 682 |
{
|
| 683 |
-
"epoch": 0.
|
| 684 |
-
"learning_rate":
|
| 685 |
-
"loss":
|
| 686 |
"step": 113
|
| 687 |
},
|
| 688 |
{
|
| 689 |
-
"epoch": 0.
|
| 690 |
-
"learning_rate":
|
| 691 |
-
"loss":
|
| 692 |
"step": 114
|
| 693 |
},
|
| 694 |
{
|
| 695 |
-
"epoch": 0.
|
| 696 |
-
"learning_rate":
|
| 697 |
-
"loss":
|
| 698 |
"step": 115
|
| 699 |
},
|
| 700 |
{
|
| 701 |
-
"epoch": 0.
|
| 702 |
-
"learning_rate":
|
| 703 |
-
"loss":
|
| 704 |
"step": 116
|
| 705 |
},
|
| 706 |
{
|
| 707 |
-
"epoch": 0.
|
| 708 |
-
"learning_rate":
|
| 709 |
-
"loss":
|
| 710 |
"step": 117
|
| 711 |
},
|
| 712 |
{
|
| 713 |
-
"epoch": 0.
|
| 714 |
-
"learning_rate":
|
| 715 |
-
"loss":
|
| 716 |
"step": 118
|
| 717 |
},
|
| 718 |
{
|
| 719 |
-
"epoch": 0.
|
| 720 |
-
"learning_rate":
|
| 721 |
-
"loss":
|
| 722 |
"step": 119
|
| 723 |
},
|
| 724 |
{
|
| 725 |
-
"epoch": 0.
|
| 726 |
-
"learning_rate":
|
| 727 |
-
"loss":
|
| 728 |
"step": 120
|
| 729 |
},
|
| 730 |
{
|
| 731 |
-
"epoch": 0.
|
| 732 |
-
"learning_rate":
|
| 733 |
-
"loss":
|
| 734 |
"step": 121
|
| 735 |
},
|
| 736 |
{
|
| 737 |
-
"epoch": 0.
|
| 738 |
-
"learning_rate":
|
| 739 |
-
"loss":
|
| 740 |
"step": 122
|
| 741 |
},
|
| 742 |
{
|
| 743 |
-
"epoch": 0.
|
| 744 |
-
"learning_rate":
|
| 745 |
-
"loss":
|
| 746 |
"step": 123
|
| 747 |
},
|
| 748 |
{
|
| 749 |
-
"epoch": 0.
|
| 750 |
-
"learning_rate":
|
| 751 |
-
"loss":
|
| 752 |
"step": 124
|
| 753 |
},
|
| 754 |
{
|
| 755 |
-
"epoch": 0.
|
| 756 |
-
"learning_rate":
|
| 757 |
-
"loss":
|
| 758 |
"step": 125
|
| 759 |
},
|
| 760 |
{
|
| 761 |
-
"epoch": 0.
|
| 762 |
-
"learning_rate":
|
| 763 |
-
"loss":
|
| 764 |
"step": 126
|
| 765 |
},
|
| 766 |
{
|
| 767 |
-
"epoch": 0.
|
| 768 |
-
"learning_rate":
|
| 769 |
-
"loss":
|
| 770 |
"step": 127
|
| 771 |
},
|
| 772 |
{
|
| 773 |
-
"epoch": 0.
|
| 774 |
-
"learning_rate":
|
| 775 |
-
"loss":
|
| 776 |
"step": 128
|
| 777 |
},
|
| 778 |
{
|
| 779 |
-
"epoch": 0.
|
| 780 |
-
"learning_rate":
|
| 781 |
-
"loss":
|
| 782 |
"step": 129
|
| 783 |
},
|
| 784 |
{
|
| 785 |
-
"epoch": 0.
|
| 786 |
-
"learning_rate":
|
| 787 |
-
"loss":
|
| 788 |
"step": 130
|
| 789 |
},
|
| 790 |
{
|
| 791 |
-
"epoch": 0.
|
| 792 |
-
"learning_rate":
|
| 793 |
-
"loss":
|
| 794 |
"step": 131
|
| 795 |
},
|
| 796 |
{
|
| 797 |
-
"epoch": 0.
|
| 798 |
-
"learning_rate":
|
| 799 |
-
"loss":
|
| 800 |
"step": 132
|
| 801 |
},
|
| 802 |
{
|
| 803 |
-
"epoch": 0.
|
| 804 |
-
"learning_rate":
|
| 805 |
-
"loss":
|
| 806 |
"step": 133
|
| 807 |
},
|
| 808 |
{
|
| 809 |
-
"epoch": 0.
|
| 810 |
-
"learning_rate":
|
| 811 |
-
"loss":
|
| 812 |
"step": 134
|
| 813 |
},
|
| 814 |
{
|
| 815 |
-
"epoch": 0.
|
| 816 |
-
"learning_rate":
|
| 817 |
-
"loss":
|
| 818 |
"step": 135
|
| 819 |
},
|
| 820 |
{
|
| 821 |
-
"epoch": 0.
|
| 822 |
-
"learning_rate":
|
| 823 |
-
"loss":
|
| 824 |
"step": 136
|
| 825 |
},
|
| 826 |
{
|
| 827 |
-
"epoch": 0.
|
| 828 |
-
"learning_rate":
|
| 829 |
-
"loss":
|
| 830 |
"step": 137
|
| 831 |
},
|
| 832 |
{
|
| 833 |
-
"epoch": 0.
|
| 834 |
-
"learning_rate":
|
| 835 |
-
"loss":
|
| 836 |
"step": 138
|
| 837 |
},
|
| 838 |
{
|
| 839 |
-
"epoch": 0.
|
| 840 |
-
"learning_rate":
|
| 841 |
-
"loss":
|
| 842 |
"step": 139
|
| 843 |
},
|
| 844 |
{
|
| 845 |
-
"epoch": 0.
|
| 846 |
-
"learning_rate":
|
| 847 |
-
"loss":
|
| 848 |
"step": 140
|
| 849 |
},
|
| 850 |
{
|
| 851 |
-
"epoch": 0.
|
| 852 |
-
"learning_rate":
|
| 853 |
-
"loss":
|
| 854 |
"step": 141
|
| 855 |
},
|
| 856 |
{
|
| 857 |
-
"epoch": 0.
|
| 858 |
-
"learning_rate":
|
| 859 |
-
"loss":
|
| 860 |
"step": 142
|
| 861 |
},
|
| 862 |
{
|
| 863 |
-
"epoch": 0.
|
| 864 |
-
"learning_rate":
|
| 865 |
-
"loss":
|
| 866 |
"step": 143
|
| 867 |
},
|
| 868 |
{
|
| 869 |
-
"epoch": 0.
|
| 870 |
-
"learning_rate":
|
| 871 |
-
"loss":
|
| 872 |
"step": 144
|
| 873 |
},
|
| 874 |
{
|
| 875 |
-
"epoch": 0.
|
| 876 |
-
"learning_rate":
|
| 877 |
-
"loss":
|
| 878 |
"step": 145
|
| 879 |
},
|
| 880 |
{
|
| 881 |
-
"epoch": 0.
|
| 882 |
-
"learning_rate":
|
| 883 |
-
"loss":
|
| 884 |
"step": 146
|
| 885 |
},
|
| 886 |
{
|
| 887 |
-
"epoch": 0.
|
| 888 |
-
"learning_rate":
|
| 889 |
-
"loss":
|
| 890 |
"step": 147
|
| 891 |
},
|
| 892 |
{
|
| 893 |
-
"epoch": 0.
|
| 894 |
-
"learning_rate":
|
| 895 |
-
"loss":
|
| 896 |
"step": 148
|
| 897 |
},
|
| 898 |
{
|
| 899 |
-
"epoch": 0.
|
| 900 |
-
"learning_rate":
|
| 901 |
-
"loss":
|
| 902 |
"step": 149
|
| 903 |
},
|
| 904 |
{
|
| 905 |
-
"epoch": 0.
|
| 906 |
-
"learning_rate":
|
| 907 |
-
"loss":
|
| 908 |
"step": 150
|
| 909 |
},
|
| 910 |
{
|
| 911 |
-
"epoch": 0.
|
| 912 |
-
"learning_rate":
|
| 913 |
-
"loss":
|
| 914 |
"step": 151
|
| 915 |
},
|
| 916 |
{
|
| 917 |
-
"epoch": 0.
|
| 918 |
-
"learning_rate":
|
| 919 |
-
"loss":
|
| 920 |
"step": 152
|
| 921 |
},
|
| 922 |
{
|
| 923 |
-
"epoch": 0.
|
| 924 |
-
"learning_rate":
|
| 925 |
-
"loss":
|
| 926 |
"step": 153
|
| 927 |
},
|
| 928 |
{
|
| 929 |
-
"epoch": 0.
|
| 930 |
-
"learning_rate":
|
| 931 |
-
"loss":
|
| 932 |
"step": 154
|
| 933 |
},
|
| 934 |
{
|
| 935 |
-
"epoch": 0.
|
| 936 |
-
"learning_rate":
|
| 937 |
-
"loss":
|
| 938 |
"step": 155
|
| 939 |
},
|
| 940 |
{
|
| 941 |
-
"epoch": 0.
|
| 942 |
-
"learning_rate":
|
| 943 |
-
"loss":
|
| 944 |
"step": 156
|
| 945 |
},
|
| 946 |
{
|
| 947 |
-
"epoch": 0.
|
| 948 |
-
"learning_rate":
|
| 949 |
-
"loss":
|
| 950 |
"step": 157
|
| 951 |
},
|
| 952 |
{
|
| 953 |
-
"epoch": 0.
|
| 954 |
-
"learning_rate":
|
| 955 |
-
"loss":
|
| 956 |
"step": 158
|
| 957 |
},
|
| 958 |
{
|
| 959 |
-
"epoch": 0.
|
| 960 |
-
"learning_rate":
|
| 961 |
-
"loss":
|
| 962 |
"step": 159
|
| 963 |
},
|
| 964 |
{
|
| 965 |
-
"epoch": 0.
|
| 966 |
-
"learning_rate":
|
| 967 |
-
"loss":
|
| 968 |
"step": 160
|
| 969 |
},
|
| 970 |
{
|
| 971 |
-
"epoch": 0.
|
| 972 |
-
"learning_rate":
|
| 973 |
-
"loss":
|
| 974 |
"step": 161
|
| 975 |
},
|
| 976 |
{
|
| 977 |
-
"epoch": 0.
|
| 978 |
-
"learning_rate":
|
| 979 |
-
"loss":
|
| 980 |
"step": 162
|
| 981 |
},
|
| 982 |
{
|
| 983 |
-
"epoch": 0.
|
| 984 |
-
"learning_rate":
|
| 985 |
-
"loss":
|
| 986 |
"step": 163
|
| 987 |
},
|
| 988 |
{
|
| 989 |
-
"epoch": 0.
|
| 990 |
-
"learning_rate":
|
| 991 |
-
"loss":
|
| 992 |
"step": 164
|
| 993 |
},
|
| 994 |
{
|
| 995 |
-
"epoch": 0.
|
| 996 |
-
"learning_rate":
|
| 997 |
-
"loss":
|
| 998 |
"step": 165
|
| 999 |
},
|
| 1000 |
{
|
| 1001 |
-
"epoch": 0.
|
| 1002 |
-
"learning_rate":
|
| 1003 |
-
"loss":
|
| 1004 |
"step": 166
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
-
"epoch": 0.
|
| 1008 |
-
"learning_rate":
|
| 1009 |
-
"loss":
|
| 1010 |
"step": 167
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
-
"epoch": 0.
|
| 1014 |
-
"learning_rate":
|
| 1015 |
-
"loss":
|
| 1016 |
"step": 168
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
-
"epoch": 0.
|
| 1020 |
-
"learning_rate":
|
| 1021 |
-
"loss":
|
| 1022 |
"step": 169
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
-
"epoch": 0.
|
| 1026 |
-
"learning_rate":
|
| 1027 |
-
"loss":
|
| 1028 |
"step": 170
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
-
"epoch": 0.
|
| 1032 |
-
"learning_rate":
|
| 1033 |
-
"loss":
|
| 1034 |
"step": 171
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
-
"epoch": 0.
|
| 1038 |
-
"learning_rate":
|
| 1039 |
-
"loss":
|
| 1040 |
"step": 172
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
-
"epoch": 0.
|
| 1044 |
-
"learning_rate":
|
| 1045 |
-
"loss":
|
| 1046 |
"step": 173
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
-
"epoch": 0.
|
| 1050 |
-
"learning_rate":
|
| 1051 |
-
"loss":
|
| 1052 |
"step": 174
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
-
"epoch": 0.
|
| 1056 |
-
"learning_rate":
|
| 1057 |
-
"loss":
|
| 1058 |
"step": 175
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
-
"epoch": 0.
|
| 1062 |
-
"learning_rate":
|
| 1063 |
-
"loss":
|
| 1064 |
"step": 176
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
-
"epoch": 0.
|
| 1068 |
-
"learning_rate":
|
| 1069 |
-
"loss":
|
| 1070 |
"step": 177
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
-
"epoch": 0.
|
| 1074 |
-
"learning_rate":
|
| 1075 |
-
"loss":
|
| 1076 |
"step": 178
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
-
"epoch": 0.
|
| 1080 |
-
"learning_rate":
|
| 1081 |
-
"loss":
|
| 1082 |
"step": 179
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
-
"epoch": 0.
|
| 1086 |
-
"learning_rate":
|
| 1087 |
-
"loss":
|
| 1088 |
"step": 180
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
-
"epoch": 0.
|
| 1092 |
-
"learning_rate":
|
| 1093 |
-
"loss":
|
| 1094 |
"step": 181
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
-
"epoch": 0.
|
| 1098 |
-
"learning_rate":
|
| 1099 |
-
"loss":
|
| 1100 |
"step": 182
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
-
"epoch": 0.
|
| 1104 |
-
"learning_rate":
|
| 1105 |
-
"loss":
|
| 1106 |
"step": 183
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
-
"epoch": 0.
|
| 1110 |
-
"learning_rate":
|
| 1111 |
-
"loss":
|
| 1112 |
"step": 184
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
-
"epoch": 0.
|
| 1116 |
-
"learning_rate":
|
| 1117 |
-
"loss":
|
| 1118 |
"step": 185
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
-
"epoch": 0.
|
| 1122 |
-
"learning_rate": 9.996713525237694e-
|
| 1123 |
-
"loss":
|
| 1124 |
"step": 186
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
-
"epoch": 0.
|
| 1128 |
-
"learning_rate": 9.8443442146016e-
|
| 1129 |
-
"loss":
|
| 1130 |
"step": 187
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
-
"epoch": 0.
|
| 1134 |
-
"learning_rate": 9.
|
| 1135 |
-
"loss":
|
| 1136 |
"step": 188
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
-
"epoch": 0.
|
| 1140 |
-
"learning_rate": 9.
|
| 1141 |
-
"loss":
|
| 1142 |
"step": 189
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
-
"epoch": 0.
|
| 1146 |
-
"learning_rate": 9.
|
| 1147 |
-
"loss":
|
| 1148 |
"step": 190
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
-
"epoch": 0.
|
| 1152 |
-
"learning_rate": 9.
|
| 1153 |
-
"loss":
|
| 1154 |
"step": 191
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
-
"epoch": 0.
|
| 1158 |
-
"learning_rate": 9.
|
| 1159 |
-
"loss":
|
| 1160 |
"step": 192
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
-
"epoch": 0.
|
| 1164 |
-
"learning_rate": 8.
|
| 1165 |
-
"loss":
|
| 1166 |
"step": 193
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
-
"epoch": 0.
|
| 1170 |
-
"learning_rate": 8.
|
| 1171 |
-
"loss":
|
| 1172 |
"step": 194
|
| 1173 |
},
|
| 1174 |
{
|
| 1175 |
-
"epoch": 0.
|
| 1176 |
-
"learning_rate": 8.648498186137653e-
|
| 1177 |
-
"loss":
|
| 1178 |
"step": 195
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
-
"epoch": 0.
|
| 1182 |
-
"learning_rate": 8.502167339067705e-
|
| 1183 |
-
"loss":
|
| 1184 |
"step": 196
|
| 1185 |
},
|
| 1186 |
{
|
| 1187 |
-
"epoch": 0.
|
| 1188 |
-
"learning_rate": 8.356593809636371e-
|
| 1189 |
-
"loss":
|
| 1190 |
"step": 197
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
-
"epoch": 0.
|
| 1194 |
-
"learning_rate": 8.
|
| 1195 |
-
"loss":
|
| 1196 |
"step": 198
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
-
"epoch": 0.
|
| 1200 |
-
"learning_rate": 8.067786479400346e-
|
| 1201 |
-
"loss":
|
| 1202 |
"step": 199
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
-
"epoch": 0.
|
| 1206 |
-
"learning_rate": 7.
|
| 1207 |
-
"loss":
|
| 1208 |
"step": 200
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
-
"epoch": 0.
|
| 1212 |
-
"learning_rate": 7.
|
| 1213 |
-
"loss":
|
| 1214 |
"step": 201
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
-
"epoch": 0.
|
| 1218 |
-
"learning_rate": 7.640676554439594e-
|
| 1219 |
-
"loss":
|
| 1220 |
"step": 202
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
-
"epoch": 0.
|
| 1224 |
-
"learning_rate": 7.
|
| 1225 |
-
"loss":
|
| 1226 |
"step": 203
|
| 1227 |
},
|
| 1228 |
{
|
| 1229 |
-
"epoch": 0.
|
| 1230 |
-
"learning_rate": 7.
|
| 1231 |
-
"loss":
|
| 1232 |
"step": 204
|
| 1233 |
},
|
| 1234 |
{
|
| 1235 |
-
"epoch": 0.
|
| 1236 |
-
"learning_rate": 7.22128554351668e-
|
| 1237 |
-
"loss":
|
| 1238 |
"step": 205
|
| 1239 |
},
|
| 1240 |
{
|
| 1241 |
-
"epoch": 0.
|
| 1242 |
-
"learning_rate": 7.083280125438766e-
|
| 1243 |
-
"loss":
|
| 1244 |
"step": 206
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
-
"epoch": 0.
|
| 1248 |
-
"learning_rate": 6.
|
| 1249 |
-
"loss":
|
| 1250 |
"step": 207
|
| 1251 |
},
|
| 1252 |
{
|
| 1253 |
-
"epoch": 0.
|
| 1254 |
-
"learning_rate": 6.
|
| 1255 |
-
"loss":
|
| 1256 |
"step": 208
|
| 1257 |
},
|
| 1258 |
{
|
| 1259 |
-
"epoch": 0.
|
| 1260 |
-
"learning_rate": 6.
|
| 1261 |
-
"loss":
|
| 1262 |
"step": 209
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
-
"epoch": 0.
|
| 1266 |
-
"learning_rate": 6.540644552236401e-
|
| 1267 |
-
"loss":
|
| 1268 |
"step": 210
|
| 1269 |
},
|
| 1270 |
{
|
| 1271 |
-
"epoch": 0.
|
| 1272 |
-
"learning_rate": 6.
|
| 1273 |
-
"loss":
|
| 1274 |
"step": 211
|
| 1275 |
},
|
| 1276 |
{
|
| 1277 |
-
"epoch": 0.
|
| 1278 |
-
"learning_rate": 6.275179386407663e-
|
| 1279 |
-
"loss":
|
| 1280 |
"step": 212
|
| 1281 |
},
|
| 1282 |
{
|
| 1283 |
-
"epoch": 0.
|
| 1284 |
-
"learning_rate": 6.
|
| 1285 |
-
"loss":
|
| 1286 |
"step": 213
|
| 1287 |
},
|
| 1288 |
{
|
| 1289 |
-
"epoch": 0.
|
| 1290 |
-
"learning_rate": 6.
|
| 1291 |
-
"loss":
|
| 1292 |
"step": 214
|
| 1293 |
},
|
| 1294 |
{
|
| 1295 |
-
"epoch": 0.
|
| 1296 |
-
"learning_rate": 5.
|
| 1297 |
-
"loss":
|
| 1298 |
"step": 215
|
| 1299 |
},
|
| 1300 |
{
|
| 1301 |
-
"epoch": 0.
|
| 1302 |
-
"learning_rate": 5.756573015986089e-
|
| 1303 |
-
"loss":
|
| 1304 |
"step": 216
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
-
"epoch": 0.
|
| 1308 |
-
"learning_rate": 5.629577238597132e-
|
| 1309 |
-
"loss":
|
| 1310 |
"step": 217
|
| 1311 |
},
|
| 1312 |
{
|
| 1313 |
-
"epoch": 0.
|
| 1314 |
-
"learning_rate": 5.503673577119552e-
|
| 1315 |
-
"loss":
|
| 1316 |
"step": 218
|
| 1317 |
},
|
| 1318 |
{
|
| 1319 |
-
"epoch": 0.
|
| 1320 |
-
"learning_rate": 5.378876705532904e-
|
| 1321 |
-
"loss":
|
| 1322 |
"step": 219
|
| 1323 |
},
|
| 1324 |
{
|
| 1325 |
-
"epoch": 0.
|
| 1326 |
-
"learning_rate": 5.
|
| 1327 |
-
"loss":
|
| 1328 |
"step": 220
|
| 1329 |
},
|
| 1330 |
{
|
| 1331 |
-
"epoch": 0.
|
| 1332 |
-
"learning_rate": 5.
|
| 1333 |
-
"loss":
|
| 1334 |
"step": 221
|
| 1335 |
},
|
| 1336 |
{
|
| 1337 |
-
"epoch": 0.
|
| 1338 |
-
"learning_rate": 5.
|
| 1339 |
-
"loss":
|
| 1340 |
"step": 222
|
| 1341 |
},
|
| 1342 |
{
|
| 1343 |
-
"epoch": 0.
|
| 1344 |
-
"learning_rate": 4.891046047349837e-
|
| 1345 |
-
"loss":
|
| 1346 |
"step": 223
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
-
"epoch": 0.
|
| 1350 |
-
"learning_rate": 4.
|
| 1351 |
-
"loss":
|
| 1352 |
"step": 224
|
| 1353 |
},
|
| 1354 |
{
|
| 1355 |
-
"epoch": 0.
|
| 1356 |
-
"learning_rate": 4.
|
| 1357 |
-
"loss":
|
| 1358 |
"step": 225
|
| 1359 |
},
|
| 1360 |
{
|
| 1361 |
-
"epoch": 0.
|
| 1362 |
-
"learning_rate": 4.
|
| 1363 |
-
"loss":
|
| 1364 |
"step": 226
|
| 1365 |
},
|
| 1366 |
{
|
| 1367 |
-
"epoch": 0.
|
| 1368 |
-
"learning_rate": 4.
|
| 1369 |
-
"loss":
|
| 1370 |
"step": 227
|
| 1371 |
},
|
| 1372 |
{
|
| 1373 |
-
"epoch": 0.
|
| 1374 |
-
"learning_rate": 4.
|
| 1375 |
-
"loss":
|
| 1376 |
"step": 228
|
| 1377 |
},
|
| 1378 |
{
|
| 1379 |
-
"epoch": 0.
|
| 1380 |
-
"learning_rate": 4.
|
| 1381 |
-
"loss":
|
| 1382 |
"step": 229
|
| 1383 |
},
|
| 1384 |
{
|
| 1385 |
-
"epoch": 0.
|
| 1386 |
-
"learning_rate": 4.
|
| 1387 |
-
"loss":
|
| 1388 |
"step": 230
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
-
"epoch": 0.
|
| 1392 |
-
"learning_rate": 3.972804094391998e-
|
| 1393 |
-
"loss":
|
| 1394 |
"step": 231
|
| 1395 |
},
|
| 1396 |
{
|
| 1397 |
-
"epoch": 0.
|
| 1398 |
-
"learning_rate": 3.
|
| 1399 |
-
"loss":
|
| 1400 |
"step": 232
|
| 1401 |
},
|
| 1402 |
{
|
| 1403 |
-
"epoch": 0.
|
| 1404 |
-
"learning_rate": 3.755833563336293e-
|
| 1405 |
-
"loss":
|
| 1406 |
"step": 233
|
| 1407 |
},
|
| 1408 |
{
|
| 1409 |
-
"epoch": 0.
|
| 1410 |
-
"learning_rate": 3.64930776273457e-
|
| 1411 |
-
"loss":
|
| 1412 |
"step": 234
|
| 1413 |
},
|
| 1414 |
{
|
| 1415 |
-
"epoch": 0.
|
| 1416 |
-
"learning_rate": 3.
|
| 1417 |
-
"loss":
|
| 1418 |
"step": 235
|
| 1419 |
},
|
| 1420 |
{
|
| 1421 |
-
"epoch": 0.
|
| 1422 |
-
"learning_rate": 3.
|
| 1423 |
-
"loss":
|
| 1424 |
"step": 236
|
| 1425 |
},
|
| 1426 |
{
|
| 1427 |
-
"epoch": 0.
|
| 1428 |
-
"learning_rate": 3.
|
| 1429 |
-
"loss":
|
| 1430 |
"step": 237
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
-
"epoch": 0.
|
| 1434 |
-
"learning_rate": 3.2365555427281634e-
|
| 1435 |
-
"loss":
|
| 1436 |
"step": 238
|
| 1437 |
},
|
| 1438 |
{
|
| 1439 |
-
"epoch": 0.
|
| 1440 |
-
"learning_rate": 3.
|
| 1441 |
-
"loss":
|
| 1442 |
"step": 239
|
| 1443 |
},
|
| 1444 |
{
|
| 1445 |
-
"epoch": 0.
|
| 1446 |
-
"learning_rate": 3.
|
| 1447 |
-
"loss":
|
| 1448 |
"step": 240
|
| 1449 |
},
|
| 1450 |
{
|
| 1451 |
-
"epoch": 0.
|
| 1452 |
-
"learning_rate": 2.
|
| 1453 |
-
"loss":
|
| 1454 |
"step": 241
|
| 1455 |
},
|
| 1456 |
{
|
| 1457 |
-
"epoch": 0.
|
| 1458 |
-
"learning_rate": 2.
|
| 1459 |
-
"loss":
|
| 1460 |
"step": 242
|
| 1461 |
},
|
| 1462 |
{
|
| 1463 |
-
"epoch": 0.
|
| 1464 |
-
"learning_rate": 2.
|
| 1465 |
-
"loss":
|
| 1466 |
"step": 243
|
| 1467 |
},
|
| 1468 |
{
|
| 1469 |
-
"epoch": 0.
|
| 1470 |
-
"learning_rate": 2.
|
| 1471 |
-
"loss":
|
| 1472 |
"step": 244
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
-
"epoch": 0.
|
| 1476 |
-
"learning_rate": 2.
|
| 1477 |
-
"loss":
|
| 1478 |
"step": 245
|
| 1479 |
},
|
| 1480 |
{
|
| 1481 |
-
"epoch": 0.
|
| 1482 |
-
"learning_rate": 2.
|
| 1483 |
-
"loss":
|
| 1484 |
"step": 246
|
| 1485 |
},
|
| 1486 |
{
|
| 1487 |
-
"epoch": 0.
|
| 1488 |
-
"learning_rate": 2.389162189636188e-
|
| 1489 |
-
"loss":
|
| 1490 |
"step": 247
|
| 1491 |
},
|
| 1492 |
{
|
| 1493 |
-
"epoch": 0.
|
| 1494 |
-
"learning_rate": 2.
|
| 1495 |
-
"loss":
|
| 1496 |
"step": 248
|
| 1497 |
},
|
| 1498 |
{
|
| 1499 |
-
"epoch": 0.
|
| 1500 |
-
"learning_rate": 2.
|
| 1501 |
-
"loss":
|
| 1502 |
"step": 249
|
| 1503 |
},
|
| 1504 |
{
|
| 1505 |
-
"epoch": 0.
|
| 1506 |
-
"learning_rate": 2.
|
| 1507 |
-
"loss":
|
| 1508 |
"step": 250
|
| 1509 |
},
|
| 1510 |
{
|
| 1511 |
-
"epoch": 0.
|
| 1512 |
-
"learning_rate": 2.
|
| 1513 |
-
"loss":
|
| 1514 |
"step": 251
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
-
"epoch": 0.
|
| 1518 |
-
"learning_rate": 1.9693239522277327e-
|
| 1519 |
-
"loss":
|
| 1520 |
"step": 252
|
| 1521 |
},
|
| 1522 |
{
|
| 1523 |
-
"epoch": 0.
|
| 1524 |
-
"learning_rate": 1.
|
| 1525 |
-
"loss":
|
| 1526 |
"step": 253
|
| 1527 |
},
|
| 1528 |
{
|
| 1529 |
-
"epoch": 0.
|
| 1530 |
-
"learning_rate": 1.
|
| 1531 |
-
"loss":
|
| 1532 |
"step": 254
|
| 1533 |
},
|
| 1534 |
{
|
| 1535 |
-
"epoch": 0.
|
| 1536 |
-
"learning_rate": 1.
|
| 1537 |
-
"loss":
|
| 1538 |
"step": 255
|
| 1539 |
},
|
| 1540 |
{
|
| 1541 |
-
"epoch": 0.
|
| 1542 |
-
"learning_rate": 1.
|
| 1543 |
-
"loss":
|
| 1544 |
"step": 256
|
| 1545 |
},
|
| 1546 |
{
|
| 1547 |
-
"epoch": 0.
|
| 1548 |
-
"learning_rate": 1.
|
| 1549 |
-
"loss":
|
| 1550 |
"step": 257
|
| 1551 |
},
|
| 1552 |
{
|
| 1553 |
-
"epoch": 0.
|
| 1554 |
-
"learning_rate": 1.
|
| 1555 |
-
"loss":
|
| 1556 |
"step": 258
|
| 1557 |
},
|
| 1558 |
{
|
| 1559 |
-
"epoch": 0.
|
| 1560 |
-
"learning_rate": 1.4455745908724226e-
|
| 1561 |
-
"loss":
|
| 1562 |
"step": 259
|
| 1563 |
},
|
| 1564 |
{
|
| 1565 |
-
"epoch": 0.
|
| 1566 |
-
"learning_rate": 1.
|
| 1567 |
-
"loss":
|
| 1568 |
"step": 260
|
| 1569 |
},
|
| 1570 |
{
|
| 1571 |
-
"epoch": 0.
|
| 1572 |
-
"learning_rate": 1.
|
| 1573 |
-
"loss":
|
| 1574 |
"step": 261
|
| 1575 |
},
|
| 1576 |
{
|
| 1577 |
-
"epoch": 0.
|
| 1578 |
-
"learning_rate": 1.2446370443349863e-
|
| 1579 |
-
"loss":
|
| 1580 |
"step": 262
|
| 1581 |
},
|
| 1582 |
{
|
| 1583 |
-
"epoch": 0.
|
| 1584 |
-
"learning_rate": 1.180853994173236e-
|
| 1585 |
-
"loss":
|
| 1586 |
"step": 263
|
| 1587 |
},
|
| 1588 |
{
|
| 1589 |
-
"epoch": 0.
|
| 1590 |
-
"learning_rate": 1.
|
| 1591 |
-
"loss":
|
| 1592 |
"step": 264
|
| 1593 |
},
|
| 1594 |
{
|
| 1595 |
-
"epoch": 0.
|
| 1596 |
-
"learning_rate": 1.058126974045811e-
|
| 1597 |
-
"loss":
|
| 1598 |
"step": 265
|
| 1599 |
},
|
| 1600 |
{
|
| 1601 |
-
"epoch": 0.
|
| 1602 |
-
"learning_rate": 9.
|
| 1603 |
-
"loss":
|
| 1604 |
"step": 266
|
| 1605 |
},
|
| 1606 |
{
|
| 1607 |
-
"epoch": 0.
|
| 1608 |
-
"learning_rate": 9.
|
| 1609 |
-
"loss":
|
| 1610 |
"step": 267
|
| 1611 |
},
|
| 1612 |
{
|
| 1613 |
-
"epoch": 0.
|
| 1614 |
-
"learning_rate": 8.
|
| 1615 |
-
"loss":
|
| 1616 |
"step": 268
|
| 1617 |
},
|
| 1618 |
{
|
| 1619 |
-
"epoch": 0.
|
| 1620 |
-
"learning_rate": 8.322255301145204e-
|
| 1621 |
-
"loss":
|
| 1622 |
"step": 269
|
| 1623 |
},
|
| 1624 |
{
|
| 1625 |
-
"epoch": 0.
|
| 1626 |
-
"learning_rate": 7.
|
| 1627 |
-
"loss":
|
| 1628 |
"step": 270
|
| 1629 |
},
|
| 1630 |
{
|
| 1631 |
-
"epoch": 0.
|
| 1632 |
-
"learning_rate": 7.
|
| 1633 |
-
"loss":
|
| 1634 |
"step": 271
|
| 1635 |
},
|
| 1636 |
{
|
| 1637 |
-
"epoch": 0.
|
| 1638 |
-
"learning_rate": 6.
|
| 1639 |
-
"loss":
|
| 1640 |
"step": 272
|
| 1641 |
},
|
| 1642 |
{
|
| 1643 |
-
"epoch": 0.
|
| 1644 |
-
"learning_rate": 6.
|
| 1645 |
-
"loss":
|
| 1646 |
"step": 273
|
| 1647 |
},
|
| 1648 |
{
|
| 1649 |
-
"epoch": 0.
|
| 1650 |
-
"learning_rate": 5.
|
| 1651 |
-
"loss":
|
| 1652 |
"step": 274
|
| 1653 |
},
|
| 1654 |
{
|
| 1655 |
-
"epoch": 0.
|
| 1656 |
-
"learning_rate": 5.430216358323309e-
|
| 1657 |
-
"loss":
|
| 1658 |
"step": 275
|
| 1659 |
},
|
| 1660 |
{
|
| 1661 |
-
"epoch": 0.
|
| 1662 |
-
"learning_rate": 5.
|
| 1663 |
-
"loss":
|
| 1664 |
"step": 276
|
| 1665 |
},
|
| 1666 |
{
|
| 1667 |
-
"epoch": 0.
|
| 1668 |
-
"learning_rate": 4.
|
| 1669 |
-
"loss":
|
| 1670 |
"step": 277
|
| 1671 |
},
|
| 1672 |
{
|
| 1673 |
-
"epoch": 0.
|
| 1674 |
-
"learning_rate": 4.
|
| 1675 |
-
"loss":
|
| 1676 |
"step": 278
|
| 1677 |
},
|
| 1678 |
{
|
| 1679 |
-
"epoch": 0.
|
| 1680 |
-
"learning_rate": 3.838421302961098e-
|
| 1681 |
-
"loss":
|
| 1682 |
"step": 279
|
| 1683 |
},
|
| 1684 |
{
|
| 1685 |
-
"epoch": 0.
|
| 1686 |
-
"learning_rate": 3.
|
| 1687 |
-
"loss":
|
| 1688 |
"step": 280
|
| 1689 |
},
|
| 1690 |
{
|
| 1691 |
-
"epoch": 0.
|
| 1692 |
-
"learning_rate": 3.
|
| 1693 |
-
"loss":
|
| 1694 |
"step": 281
|
| 1695 |
},
|
| 1696 |
{
|
| 1697 |
-
"epoch": 0.
|
| 1698 |
-
"learning_rate": 2.
|
| 1699 |
-
"loss":
|
| 1700 |
"step": 282
|
| 1701 |
},
|
| 1702 |
{
|
| 1703 |
-
"epoch": 0.
|
| 1704 |
-
"learning_rate": 2.
|
| 1705 |
-
"loss":
|
| 1706 |
"step": 283
|
| 1707 |
},
|
| 1708 |
{
|
| 1709 |
-
"epoch": 0.
|
| 1710 |
-
"learning_rate": 2.
|
| 1711 |
-
"loss":
|
| 1712 |
"step": 284
|
| 1713 |
},
|
| 1714 |
{
|
| 1715 |
-
"epoch": 0.
|
| 1716 |
-
"learning_rate": 1.
|
| 1717 |
-
"loss":
|
| 1718 |
"step": 285
|
| 1719 |
},
|
| 1720 |
{
|
| 1721 |
-
"epoch": 0.
|
| 1722 |
-
"learning_rate": 1.
|
| 1723 |
-
"loss":
|
| 1724 |
"step": 286
|
| 1725 |
},
|
| 1726 |
{
|
| 1727 |
-
"epoch": 0.
|
| 1728 |
-
"learning_rate": 1.
|
| 1729 |
-
"loss":
|
| 1730 |
"step": 287
|
| 1731 |
},
|
| 1732 |
{
|
| 1733 |
-
"epoch": 0.
|
| 1734 |
-
"learning_rate": 1.2569848160362384e-
|
| 1735 |
-
"loss":
|
| 1736 |
"step": 288
|
| 1737 |
},
|
| 1738 |
{
|
| 1739 |
-
"epoch": 0.
|
| 1740 |
-
"learning_rate": 1.
|
| 1741 |
-
"loss":
|
| 1742 |
"step": 289
|
| 1743 |
},
|
| 1744 |
{
|
| 1745 |
-
"epoch": 0.
|
| 1746 |
-
"learning_rate": 8.
|
| 1747 |
-
"loss":
|
| 1748 |
"step": 290
|
| 1749 |
},
|
| 1750 |
{
|
| 1751 |
-
"epoch": 0.
|
| 1752 |
-
"learning_rate": 7.
|
| 1753 |
-
"loss":
|
| 1754 |
"step": 291
|
| 1755 |
},
|
| 1756 |
{
|
| 1757 |
-
"epoch": 0.
|
| 1758 |
-
"learning_rate": 5.
|
| 1759 |
-
"loss":
|
| 1760 |
"step": 292
|
| 1761 |
},
|
| 1762 |
{
|
| 1763 |
-
"epoch": 0.
|
| 1764 |
-
"learning_rate": 4.
|
| 1765 |
-
"loss":
|
| 1766 |
"step": 293
|
| 1767 |
},
|
| 1768 |
{
|
| 1769 |
-
"epoch": 0.
|
| 1770 |
-
"learning_rate": 3.
|
| 1771 |
-
"loss":
|
| 1772 |
"step": 294
|
| 1773 |
},
|
| 1774 |
{
|
| 1775 |
-
"epoch": 0.
|
| 1776 |
-
"learning_rate": 2.
|
| 1777 |
-
"loss":
|
| 1778 |
"step": 295
|
| 1779 |
},
|
| 1780 |
{
|
| 1781 |
-
"epoch": 0.
|
| 1782 |
-
"learning_rate": 1.3983874665589035e-
|
| 1783 |
-
"loss":
|
| 1784 |
"step": 296
|
| 1785 |
},
|
| 1786 |
{
|
| 1787 |
-
"epoch": 0.
|
| 1788 |
-
"learning_rate": 7.
|
| 1789 |
-
"loss":
|
| 1790 |
"step": 297
|
| 1791 |
},
|
| 1792 |
{
|
| 1793 |
-
"epoch": 0.
|
| 1794 |
-
"learning_rate": 3.
|
| 1795 |
-
"loss":
|
| 1796 |
"step": 298
|
| 1797 |
},
|
| 1798 |
{
|
| 1799 |
-
"epoch": 0.
|
| 1800 |
-
"learning_rate": 8.741195081479747e-
|
| 1801 |
-
"loss":
|
| 1802 |
"step": 299
|
| 1803 |
},
|
| 1804 |
{
|
| 1805 |
-
"epoch": 0.
|
| 1806 |
"learning_rate": 0.0,
|
| 1807 |
-
"loss":
|
| 1808 |
"step": 300
|
| 1809 |
}
|
| 1810 |
],
|
|
|
|
| 1811 |
"max_steps": 300,
|
| 1812 |
"num_train_epochs": 1,
|
| 1813 |
-
"
|
|
|
|
| 1814 |
"trial_name": null,
|
| 1815 |
"trial_params": null
|
| 1816 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.18547140649149924,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
"global_step": 300,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
|
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 3.333333333333333e-05,
|
| 14 |
+
"loss": 1.1619,
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.0,
|
| 19 |
+
"learning_rate": 6.666666666666666e-05,
|
| 20 |
+
"loss": 1.1173,
|
| 21 |
"step": 2
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.0,
|
| 25 |
+
"learning_rate": 9.999999999999999e-05,
|
| 26 |
+
"loss": 1.1862,
|
| 27 |
"step": 3
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.0,
|
| 31 |
+
"learning_rate": 0.0001333333333333333,
|
| 32 |
+
"loss": 1.243,
|
| 33 |
"step": 4
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.0,
|
| 37 |
+
"learning_rate": 0.00016666666666666666,
|
| 38 |
+
"loss": 1.2164,
|
| 39 |
"step": 5
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.0,
|
| 43 |
+
"learning_rate": 0.00019999999999999998,
|
| 44 |
+
"loss": 1.1963,
|
| 45 |
"step": 6
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.0,
|
| 49 |
+
"learning_rate": 0.0002333333333333333,
|
| 50 |
+
"loss": 1.19,
|
| 51 |
"step": 7
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.0,
|
| 55 |
+
"learning_rate": 0.0002666666666666666,
|
| 56 |
+
"loss": 1.1195,
|
| 57 |
"step": 8
|
| 58 |
},
|
| 59 |
{
|
| 60 |
+
"epoch": 0.01,
|
| 61 |
+
"learning_rate": 0.0003,
|
| 62 |
+
"loss": 1.1389,
|
| 63 |
"step": 9
|
| 64 |
},
|
| 65 |
{
|
| 66 |
+
"epoch": 0.01,
|
| 67 |
+
"learning_rate": 0.00029999125880491846,
|
| 68 |
+
"loss": 1.0748,
|
| 69 |
"step": 10
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 0.01,
|
| 73 |
+
"learning_rate": 0.00029996503623845393,
|
| 74 |
+
"loss": 1.0508,
|
| 75 |
"step": 11
|
| 76 |
},
|
| 77 |
{
|
| 78 |
+
"epoch": 0.01,
|
| 79 |
+
"learning_rate": 0.00029992133535682725,
|
| 80 |
+
"loss": 0.9739,
|
| 81 |
"step": 12
|
| 82 |
},
|
| 83 |
{
|
| 84 |
+
"epoch": 0.01,
|
| 85 |
+
"learning_rate": 0.00029986016125334406,
|
| 86 |
+
"loss": 0.9708,
|
| 87 |
"step": 13
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 0.01,
|
| 91 |
+
"learning_rate": 0.0002997815210578015,
|
| 92 |
+
"loss": 0.9641,
|
| 93 |
"step": 14
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.01,
|
| 97 |
+
"learning_rate": 0.0002996854239356567,
|
| 98 |
+
"loss": 0.8698,
|
| 99 |
"step": 15
|
| 100 |
},
|
| 101 |
{
|
| 102 |
+
"epoch": 0.01,
|
| 103 |
+
"learning_rate": 0.0002995718810869589,
|
| 104 |
+
"loss": 0.9405,
|
| 105 |
"step": 16
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.01,
|
| 109 |
+
"learning_rate": 0.00029944090574504395,
|
| 110 |
+
"loss": 0.9377,
|
| 111 |
"step": 17
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.01,
|
| 115 |
+
"learning_rate": 0.0002992925131749921,
|
| 116 |
+
"loss": 0.8775,
|
| 117 |
"step": 18
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.01,
|
| 121 |
+
"learning_rate": 0.0002991267206718486,
|
| 122 |
+
"loss": 0.8682,
|
| 123 |
"step": 19
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.01,
|
| 127 |
+
"learning_rate": 0.00029894354755860845,
|
| 128 |
+
"loss": 0.8541,
|
| 129 |
"step": 20
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.01,
|
| 133 |
+
"learning_rate": 0.00029874301518396376,
|
| 134 |
+
"loss": 0.8653,
|
| 135 |
"step": 21
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.01,
|
| 139 |
+
"learning_rate": 0.000298525146919816,
|
| 140 |
+
"loss": 0.9036,
|
| 141 |
"step": 22
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.01,
|
| 145 |
+
"learning_rate": 0.0002982899681585518,
|
| 146 |
+
"loss": 0.8277,
|
| 147 |
"step": 23
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 0.01,
|
| 151 |
+
"learning_rate": 0.00029803750631008356,
|
| 152 |
+
"loss": 0.8784,
|
| 153 |
"step": 24
|
| 154 |
},
|
| 155 |
{
|
| 156 |
+
"epoch": 0.02,
|
| 157 |
+
"learning_rate": 0.00029776779079865496,
|
| 158 |
+
"loss": 0.7871,
|
| 159 |
"step": 25
|
| 160 |
},
|
| 161 |
{
|
| 162 |
+
"epoch": 0.02,
|
| 163 |
+
"learning_rate": 0.00029748085305941123,
|
| 164 |
+
"loss": 0.8347,
|
| 165 |
"step": 26
|
| 166 |
},
|
| 167 |
{
|
| 168 |
+
"epoch": 0.02,
|
| 169 |
+
"learning_rate": 0.0002971767265347358,
|
| 170 |
+
"loss": 0.8515,
|
| 171 |
"step": 27
|
| 172 |
},
|
| 173 |
{
|
| 174 |
+
"epoch": 0.02,
|
| 175 |
+
"learning_rate": 0.0002968554466703525,
|
| 176 |
+
"loss": 0.7445,
|
| 177 |
"step": 28
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 0.02,
|
| 181 |
+
"learning_rate": 0.0002965170509111942,
|
| 182 |
+
"loss": 0.7724,
|
| 183 |
"step": 29
|
| 184 |
},
|
| 185 |
{
|
| 186 |
+
"epoch": 0.02,
|
| 187 |
+
"learning_rate": 0.0002961615786970389,
|
| 188 |
+
"loss": 0.8284,
|
| 189 |
"step": 30
|
| 190 |
},
|
| 191 |
{
|
| 192 |
+
"epoch": 0.02,
|
| 193 |
+
"learning_rate": 0.00029578907145791274,
|
| 194 |
+
"loss": 0.8293,
|
| 195 |
"step": 31
|
| 196 |
},
|
| 197 |
{
|
| 198 |
+
"epoch": 0.02,
|
| 199 |
+
"learning_rate": 0.00029539957260926183,
|
| 200 |
+
"loss": 0.7876,
|
| 201 |
"step": 32
|
| 202 |
},
|
| 203 |
{
|
| 204 |
+
"epoch": 0.02,
|
| 205 |
+
"learning_rate": 0.0002949931275468917,
|
| 206 |
+
"loss": 0.7748,
|
| 207 |
"step": 33
|
| 208 |
},
|
| 209 |
{
|
| 210 |
+
"epoch": 0.02,
|
| 211 |
+
"learning_rate": 0.0002945697836416767,
|
| 212 |
+
"loss": 0.7591,
|
| 213 |
"step": 34
|
| 214 |
},
|
| 215 |
{
|
| 216 |
+
"epoch": 0.02,
|
| 217 |
+
"learning_rate": 0.000294129590234039,
|
| 218 |
+
"loss": 0.73,
|
| 219 |
"step": 35
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 0.02,
|
| 223 |
+
"learning_rate": 0.00029367259862819804,
|
| 224 |
+
"loss": 0.7324,
|
| 225 |
"step": 36
|
| 226 |
},
|
| 227 |
{
|
| 228 |
+
"epoch": 0.02,
|
| 229 |
+
"learning_rate": 0.00029319886208619073,
|
| 230 |
+
"loss": 0.7659,
|
| 231 |
"step": 37
|
| 232 |
},
|
| 233 |
{
|
| 234 |
+
"epoch": 0.02,
|
| 235 |
+
"learning_rate": 0.00029270843582166427,
|
| 236 |
+
"loss": 0.7069,
|
| 237 |
"step": 38
|
| 238 |
},
|
| 239 |
{
|
| 240 |
+
"epoch": 0.02,
|
| 241 |
+
"learning_rate": 0.00029220137699344055,
|
| 242 |
+
"loss": 0.7087,
|
| 243 |
"step": 39
|
| 244 |
},
|
| 245 |
{
|
| 246 |
+
"epoch": 0.02,
|
| 247 |
+
"learning_rate": 0.0002916777446988548,
|
| 248 |
+
"loss": 0.6674,
|
| 249 |
"step": 40
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 0.03,
|
| 253 |
+
"learning_rate": 0.00029113759996686743,
|
| 254 |
+
"loss": 0.5521,
|
| 255 |
"step": 41
|
| 256 |
},
|
| 257 |
{
|
| 258 |
+
"epoch": 0.03,
|
| 259 |
+
"learning_rate": 0.0002905810057509515,
|
| 260 |
+
"loss": 0.6268,
|
| 261 |
"step": 42
|
| 262 |
},
|
| 263 |
{
|
| 264 |
+
"epoch": 0.03,
|
| 265 |
+
"learning_rate": 0.00029000802692175537,
|
| 266 |
+
"loss": 0.666,
|
| 267 |
"step": 43
|
| 268 |
},
|
| 269 |
{
|
| 270 |
+
"epoch": 0.03,
|
| 271 |
+
"learning_rate": 0.0002894187302595419,
|
| 272 |
+
"loss": 0.6509,
|
| 273 |
"step": 44
|
| 274 |
},
|
| 275 |
{
|
| 276 |
+
"epoch": 0.03,
|
| 277 |
+
"learning_rate": 0.0002888131844464056,
|
| 278 |
+
"loss": 0.5752,
|
| 279 |
"step": 45
|
| 280 |
},
|
| 281 |
{
|
| 282 |
+
"epoch": 0.03,
|
| 283 |
+
"learning_rate": 0.0002881914600582676,
|
| 284 |
+
"loss": 0.5702,
|
| 285 |
"step": 46
|
| 286 |
},
|
| 287 |
{
|
| 288 |
+
"epoch": 0.03,
|
| 289 |
+
"learning_rate": 0.0002875536295566501,
|
| 290 |
+
"loss": 0.5882,
|
| 291 |
"step": 47
|
| 292 |
},
|
| 293 |
{
|
| 294 |
+
"epoch": 0.03,
|
| 295 |
+
"learning_rate": 0.000286899767280231,
|
| 296 |
+
"loss": 0.5374,
|
| 297 |
"step": 48
|
| 298 |
},
|
| 299 |
{
|
| 300 |
+
"epoch": 0.03,
|
| 301 |
+
"learning_rate": 0.0002862299494361798,
|
| 302 |
+
"loss": 0.4725,
|
| 303 |
"step": 49
|
| 304 |
},
|
| 305 |
{
|
| 306 |
+
"epoch": 0.03,
|
| 307 |
+
"learning_rate": 0.0002855442540912758,
|
| 308 |
+
"loss": 0.4424,
|
| 309 |
"step": 50
|
| 310 |
},
|
| 311 |
{
|
| 312 |
+
"epoch": 0.03,
|
| 313 |
+
"learning_rate": 0.00028484276116280926,
|
| 314 |
+
"loss": 0.9098,
|
| 315 |
"step": 51
|
| 316 |
},
|
| 317 |
{
|
| 318 |
+
"epoch": 0.03,
|
| 319 |
+
"learning_rate": 0.0002841255524092674,
|
| 320 |
+
"loss": 0.8235,
|
| 321 |
"step": 52
|
| 322 |
},
|
| 323 |
{
|
| 324 |
+
"epoch": 0.03,
|
| 325 |
+
"learning_rate": 0.00028339271142080534,
|
| 326 |
+
"loss": 0.8614,
|
| 327 |
"step": 53
|
| 328 |
},
|
| 329 |
{
|
| 330 |
+
"epoch": 0.03,
|
| 331 |
+
"learning_rate": 0.00028264432360950353,
|
| 332 |
+
"loss": 0.8358,
|
| 333 |
"step": 54
|
| 334 |
},
|
| 335 |
{
|
| 336 |
+
"epoch": 0.03,
|
| 337 |
+
"learning_rate": 0.00028188047619941343,
|
| 338 |
+
"loss": 0.8151,
|
| 339 |
"step": 55
|
| 340 |
},
|
| 341 |
{
|
| 342 |
+
"epoch": 0.03,
|
| 343 |
+
"learning_rate": 0.0002811012582163913,
|
| 344 |
+
"loss": 0.8758,
|
| 345 |
"step": 56
|
| 346 |
},
|
| 347 |
{
|
| 348 |
+
"epoch": 0.04,
|
| 349 |
+
"learning_rate": 0.00028030676047772265,
|
| 350 |
+
"loss": 0.8098,
|
| 351 |
"step": 57
|
| 352 |
},
|
| 353 |
{
|
| 354 |
+
"epoch": 0.04,
|
| 355 |
+
"learning_rate": 0.000279497075581537,
|
| 356 |
+
"loss": 0.8003,
|
| 357 |
"step": 58
|
| 358 |
},
|
| 359 |
{
|
| 360 |
+
"epoch": 0.04,
|
| 361 |
+
"learning_rate": 0.0002786722978960161,
|
| 362 |
+
"loss": 0.7646,
|
| 363 |
"step": 59
|
| 364 |
},
|
| 365 |
{
|
| 366 |
+
"epoch": 0.04,
|
| 367 |
+
"learning_rate": 0.0002778325235483954,
|
| 368 |
+
"loss": 0.7734,
|
| 369 |
"step": 60
|
| 370 |
},
|
| 371 |
{
|
| 372 |
+
"epoch": 0.04,
|
| 373 |
+
"learning_rate": 0.00027697785041376006,
|
| 374 |
+
"loss": 0.7807,
|
| 375 |
"step": 61
|
| 376 |
},
|
| 377 |
{
|
| 378 |
+
"epoch": 0.04,
|
| 379 |
+
"learning_rate": 0.0002761083781036381,
|
| 380 |
+
"loss": 0.8398,
|
| 381 |
"step": 62
|
| 382 |
},
|
| 383 |
{
|
| 384 |
+
"epoch": 0.04,
|
| 385 |
+
"learning_rate": 0.00027522420795439065,
|
| 386 |
+
"loss": 0.7767,
|
| 387 |
"step": 63
|
| 388 |
},
|
| 389 |
{
|
| 390 |
+
"epoch": 0.04,
|
| 391 |
+
"learning_rate": 0.0002743254430154012,
|
| 392 |
+
"loss": 0.8138,
|
| 393 |
"step": 64
|
| 394 |
},
|
| 395 |
{
|
| 396 |
+
"epoch": 0.04,
|
| 397 |
+
"learning_rate": 0.0002734121880370652,
|
| 398 |
+
"loss": 0.7638,
|
| 399 |
"step": 65
|
| 400 |
},
|
| 401 |
{
|
| 402 |
+
"epoch": 0.04,
|
| 403 |
+
"learning_rate": 0.0002724845494585816,
|
| 404 |
+
"loss": 0.8572,
|
| 405 |
"step": 66
|
| 406 |
},
|
| 407 |
{
|
| 408 |
+
"epoch": 0.04,
|
| 409 |
+
"learning_rate": 0.0002715426353955476,
|
| 410 |
+
"loss": 0.8552,
|
| 411 |
"step": 67
|
| 412 |
},
|
| 413 |
{
|
| 414 |
+
"epoch": 0.04,
|
| 415 |
+
"learning_rate": 0.0002705865556273575,
|
| 416 |
+
"loss": 0.7779,
|
| 417 |
"step": 68
|
| 418 |
},
|
| 419 |
{
|
| 420 |
+
"epoch": 0.04,
|
| 421 |
+
"learning_rate": 0.0002696164215844081,
|
| 422 |
+
"loss": 0.8717,
|
| 423 |
"step": 69
|
| 424 |
},
|
| 425 |
{
|
| 426 |
+
"epoch": 0.04,
|
| 427 |
+
"learning_rate": 0.00026863234633511183,
|
| 428 |
+
"loss": 0.8209,
|
| 429 |
"step": 70
|
| 430 |
},
|
| 431 |
{
|
| 432 |
+
"epoch": 0.04,
|
| 433 |
+
"learning_rate": 0.00026763444457271837,
|
| 434 |
+
"loss": 0.8173,
|
| 435 |
"step": 71
|
| 436 |
},
|
| 437 |
{
|
| 438 |
+
"epoch": 0.04,
|
| 439 |
+
"learning_rate": 0.0002666228326019474,
|
| 440 |
+
"loss": 0.8009,
|
| 441 |
"step": 72
|
| 442 |
},
|
| 443 |
{
|
| 444 |
+
"epoch": 0.05,
|
| 445 |
+
"learning_rate": 0.00026559762832543336,
|
| 446 |
+
"loss": 0.8182,
|
| 447 |
"step": 73
|
| 448 |
},
|
| 449 |
{
|
| 450 |
+
"epoch": 0.05,
|
| 451 |
+
"learning_rate": 0.000264558951229984,
|
| 452 |
+
"loss": 0.8317,
|
| 453 |
"step": 74
|
| 454 |
},
|
| 455 |
{
|
| 456 |
+
"epoch": 0.05,
|
| 457 |
+
"learning_rate": 0.00026350692237265427,
|
| 458 |
+
"loss": 0.7987,
|
| 459 |
"step": 75
|
| 460 |
},
|
| 461 |
{
|
| 462 |
+
"epoch": 0.05,
|
| 463 |
+
"learning_rate": 0.0002624416643666371,
|
| 464 |
+
"loss": 0.7929,
|
| 465 |
"step": 76
|
| 466 |
},
|
| 467 |
{
|
| 468 |
+
"epoch": 0.05,
|
| 469 |
+
"learning_rate": 0.000261363301366973,
|
| 470 |
+
"loss": 0.7416,
|
| 471 |
"step": 77
|
| 472 |
},
|
| 473 |
{
|
| 474 |
+
"epoch": 0.05,
|
| 475 |
+
"learning_rate": 0.00026027195905608006,
|
| 476 |
+
"loss": 0.799,
|
| 477 |
"step": 78
|
| 478 |
},
|
| 479 |
{
|
| 480 |
+
"epoch": 0.05,
|
| 481 |
+
"learning_rate": 0.0002591677646291054,
|
| 482 |
+
"loss": 0.7643,
|
| 483 |
"step": 79
|
| 484 |
},
|
| 485 |
{
|
| 486 |
+
"epoch": 0.05,
|
| 487 |
+
"learning_rate": 0.00025805084677910095,
|
| 488 |
+
"loss": 0.7045,
|
| 489 |
"step": 80
|
| 490 |
},
|
| 491 |
{
|
| 492 |
+
"epoch": 0.05,
|
| 493 |
+
"learning_rate": 0.0002569213356820244,
|
| 494 |
+
"loss": 0.7651,
|
| 495 |
"step": 81
|
| 496 |
},
|
| 497 |
{
|
| 498 |
+
"epoch": 0.05,
|
| 499 |
+
"learning_rate": 0.0002557793629815669,
|
| 500 |
+
"loss": 0.727,
|
| 501 |
"step": 82
|
| 502 |
},
|
| 503 |
{
|
| 504 |
+
"epoch": 0.05,
|
| 505 |
+
"learning_rate": 0.00025462506177381043,
|
| 506 |
+
"loss": 0.7623,
|
| 507 |
"step": 83
|
| 508 |
},
|
| 509 |
{
|
| 510 |
+
"epoch": 0.05,
|
| 511 |
+
"learning_rate": 0.00025345856659171563,
|
| 512 |
+
"loss": 0.7209,
|
| 513 |
"step": 84
|
| 514 |
},
|
| 515 |
{
|
| 516 |
+
"epoch": 0.05,
|
| 517 |
+
"learning_rate": 0.00025228001338944175,
|
| 518 |
+
"loss": 0.7104,
|
| 519 |
"step": 85
|
| 520 |
},
|
| 521 |
{
|
| 522 |
+
"epoch": 0.05,
|
| 523 |
+
"learning_rate": 0.0002510895395265016,
|
| 524 |
+
"loss": 0.6989,
|
| 525 |
"step": 86
|
| 526 |
},
|
| 527 |
{
|
| 528 |
+
"epoch": 0.05,
|
| 529 |
+
"learning_rate": 0.00024988728375175214,
|
| 530 |
+
"loss": 0.7347,
|
| 531 |
"step": 87
|
| 532 |
},
|
| 533 |
{
|
| 534 |
+
"epoch": 0.05,
|
| 535 |
+
"learning_rate": 0.00024867338618722357,
|
| 536 |
+
"loss": 0.6649,
|
| 537 |
"step": 88
|
| 538 |
},
|
| 539 |
{
|
| 540 |
+
"epoch": 0.06,
|
| 541 |
+
"learning_rate": 0.0002474479883117882,
|
| 542 |
+
"loss": 0.6432,
|
| 543 |
"step": 89
|
| 544 |
},
|
| 545 |
{
|
| 546 |
+
"epoch": 0.06,
|
| 547 |
+
"learning_rate": 0.00024621123294467096,
|
| 548 |
+
"loss": 0.6257,
|
| 549 |
"step": 90
|
| 550 |
},
|
| 551 |
{
|
| 552 |
+
"epoch": 0.06,
|
| 553 |
+
"learning_rate": 0.0002449632642288045,
|
| 554 |
+
"loss": 0.5743,
|
| 555 |
"step": 91
|
| 556 |
},
|
| 557 |
{
|
| 558 |
+
"epoch": 0.06,
|
| 559 |
+
"learning_rate": 0.00024370422761402867,
|
| 560 |
+
"loss": 0.6499,
|
| 561 |
"step": 92
|
| 562 |
},
|
| 563 |
{
|
| 564 |
+
"epoch": 0.06,
|
| 565 |
+
"learning_rate": 0.0002424342698401391,
|
| 566 |
+
"loss": 0.5454,
|
| 567 |
"step": 93
|
| 568 |
},
|
| 569 |
{
|
| 570 |
+
"epoch": 0.06,
|
| 571 |
+
"learning_rate": 0.00024115353891978431,
|
| 572 |
+
"loss": 0.561,
|
| 573 |
"step": 94
|
| 574 |
},
|
| 575 |
{
|
| 576 |
+
"epoch": 0.06,
|
| 577 |
+
"learning_rate": 0.00023986218412121537,
|
| 578 |
+
"loss": 0.6056,
|
| 579 |
"step": 95
|
| 580 |
},
|
| 581 |
{
|
| 582 |
+
"epoch": 0.06,
|
| 583 |
+
"learning_rate": 0.00023856035595088839,
|
| 584 |
+
"loss": 0.518,
|
| 585 |
"step": 96
|
| 586 |
},
|
| 587 |
{
|
| 588 |
+
"epoch": 0.06,
|
| 589 |
+
"learning_rate": 0.00023724820613592337,
|
| 590 |
+
"loss": 0.5195,
|
| 591 |
"step": 97
|
| 592 |
},
|
| 593 |
{
|
| 594 |
+
"epoch": 0.06,
|
| 595 |
+
"learning_rate": 0.00023592588760642044,
|
| 596 |
+
"loss": 0.5386,
|
| 597 |
"step": 98
|
| 598 |
},
|
| 599 |
{
|
| 600 |
+
"epoch": 0.06,
|
| 601 |
+
"learning_rate": 0.00023459355447763596,
|
| 602 |
+
"loss": 0.4816,
|
| 603 |
"step": 99
|
| 604 |
},
|
| 605 |
{
|
| 606 |
+
"epoch": 0.06,
|
| 607 |
+
"learning_rate": 0.00023325136203202049,
|
| 608 |
+
"loss": 0.3958,
|
| 609 |
"step": 100
|
| 610 |
},
|
| 611 |
{
|
| 612 |
+
"epoch": 0.06,
|
| 613 |
+
"learning_rate": 0.00023189946670112069,
|
| 614 |
+
"loss": 0.7837,
|
| 615 |
"step": 101
|
| 616 |
},
|
| 617 |
{
|
| 618 |
+
"epoch": 0.06,
|
| 619 |
+
"learning_rate": 0.00023053802604734757,
|
| 620 |
+
"loss": 0.8046,
|
| 621 |
"step": 102
|
| 622 |
},
|
| 623 |
{
|
| 624 |
+
"epoch": 0.06,
|
| 625 |
+
"learning_rate": 0.00022916719874561226,
|
| 626 |
+
"loss": 0.8202,
|
| 627 |
"step": 103
|
| 628 |
},
|
| 629 |
{
|
| 630 |
+
"epoch": 0.06,
|
| 631 |
+
"learning_rate": 0.0002277871445648332,
|
| 632 |
+
"loss": 0.8035,
|
| 633 |
"step": 104
|
| 634 |
},
|
| 635 |
{
|
| 636 |
+
"epoch": 0.06,
|
| 637 |
+
"learning_rate": 0.00022639802434931444,
|
| 638 |
+
"loss": 0.8827,
|
| 639 |
"step": 105
|
| 640 |
},
|
| 641 |
{
|
| 642 |
+
"epoch": 0.07,
|
| 643 |
+
"learning_rate": 0.000225,
|
| 644 |
+
"loss": 0.7623,
|
| 645 |
"step": 106
|
| 646 |
},
|
| 647 |
{
|
| 648 |
+
"epoch": 0.07,
|
| 649 |
+
"learning_rate": 0.00022359323445560406,
|
| 650 |
+
"loss": 0.8376,
|
| 651 |
"step": 107
|
| 652 |
},
|
| 653 |
{
|
| 654 |
+
"epoch": 0.07,
|
| 655 |
+
"learning_rate": 0.00022217789167362073,
|
| 656 |
+
"loss": 0.8254,
|
| 657 |
"step": 108
|
| 658 |
},
|
| 659 |
{
|
| 660 |
+
"epoch": 0.07,
|
| 661 |
+
"learning_rate": 0.00022075413661121492,
|
| 662 |
+
"loss": 0.8591,
|
| 663 |
"step": 109
|
| 664 |
},
|
| 665 |
{
|
| 666 |
+
"epoch": 0.07,
|
| 667 |
+
"learning_rate": 0.00021932213520599653,
|
| 668 |
+
"loss": 0.77,
|
| 669 |
"step": 110
|
| 670 |
},
|
| 671 |
{
|
| 672 |
+
"epoch": 0.07,
|
| 673 |
+
"learning_rate": 0.00021788205435668083,
|
| 674 |
+
"loss": 0.8327,
|
| 675 |
"step": 111
|
| 676 |
},
|
| 677 |
{
|
| 678 |
+
"epoch": 0.07,
|
| 679 |
+
"learning_rate": 0.00021643406190363624,
|
| 680 |
+
"loss": 0.8196,
|
| 681 |
"step": 112
|
| 682 |
},
|
| 683 |
{
|
| 684 |
+
"epoch": 0.07,
|
| 685 |
+
"learning_rate": 0.00021497832660932295,
|
| 686 |
+
"loss": 0.7589,
|
| 687 |
"step": 113
|
| 688 |
},
|
| 689 |
{
|
| 690 |
+
"epoch": 0.07,
|
| 691 |
+
"learning_rate": 0.00021351501813862356,
|
| 692 |
+
"loss": 0.8008,
|
| 693 |
"step": 114
|
| 694 |
},
|
| 695 |
{
|
| 696 |
+
"epoch": 0.07,
|
| 697 |
+
"learning_rate": 0.0002120443070390687,
|
| 698 |
+
"loss": 0.7663,
|
| 699 |
"step": 115
|
| 700 |
},
|
| 701 |
{
|
| 702 |
+
"epoch": 0.07,
|
| 703 |
+
"learning_rate": 0.00021056636472096025,
|
| 704 |
+
"loss": 0.8268,
|
| 705 |
"step": 116
|
| 706 |
},
|
| 707 |
{
|
| 708 |
+
"epoch": 0.07,
|
| 709 |
+
"learning_rate": 0.00020908136343739307,
|
| 710 |
+
"loss": 0.7696,
|
| 711 |
"step": 117
|
| 712 |
},
|
| 713 |
{
|
| 714 |
+
"epoch": 0.07,
|
| 715 |
+
"learning_rate": 0.00020758947626417943,
|
| 716 |
+
"loss": 0.8556,
|
| 717 |
"step": 118
|
| 718 |
},
|
| 719 |
{
|
| 720 |
+
"epoch": 0.07,
|
| 721 |
+
"learning_rate": 0.0002060908770796769,
|
| 722 |
+
"loss": 0.8164,
|
| 723 |
"step": 119
|
| 724 |
},
|
| 725 |
{
|
| 726 |
+
"epoch": 0.07,
|
| 727 |
+
"learning_rate": 0.00020458574054452313,
|
| 728 |
+
"loss": 0.8059,
|
| 729 |
"step": 120
|
| 730 |
},
|
| 731 |
{
|
| 732 |
+
"epoch": 0.07,
|
| 733 |
+
"learning_rate": 0.00020307424208127912,
|
| 734 |
+
"loss": 0.8276,
|
| 735 |
"step": 121
|
| 736 |
},
|
| 737 |
{
|
| 738 |
+
"epoch": 0.08,
|
| 739 |
+
"learning_rate": 0.00020155655785398393,
|
| 740 |
+
"loss": 0.7899,
|
| 741 |
"step": 122
|
| 742 |
},
|
| 743 |
{
|
| 744 |
+
"epoch": 0.08,
|
| 745 |
+
"learning_rate": 0.0002000328647476231,
|
| 746 |
+
"loss": 0.7504,
|
| 747 |
"step": 123
|
| 748 |
},
|
| 749 |
{
|
| 750 |
+
"epoch": 0.08,
|
| 751 |
+
"learning_rate": 0.00019850334034751226,
|
| 752 |
+
"loss": 0.8724,
|
| 753 |
"step": 124
|
| 754 |
},
|
| 755 |
{
|
| 756 |
+
"epoch": 0.08,
|
| 757 |
+
"learning_rate": 0.00019696816291860038,
|
| 758 |
+
"loss": 0.8076,
|
| 759 |
"step": 125
|
| 760 |
},
|
| 761 |
{
|
| 762 |
+
"epoch": 0.08,
|
| 763 |
+
"learning_rate": 0.0001954275113846926,
|
| 764 |
+
"loss": 0.7999,
|
| 765 |
"step": 126
|
| 766 |
},
|
| 767 |
{
|
| 768 |
+
"epoch": 0.08,
|
| 769 |
+
"learning_rate": 0.00019388156530759712,
|
| 770 |
+
"loss": 0.7523,
|
| 771 |
"step": 127
|
| 772 |
},
|
| 773 |
{
|
| 774 |
+
"epoch": 0.08,
|
| 775 |
+
"learning_rate": 0.00019233050486619713,
|
| 776 |
+
"loss": 0.7789,
|
| 777 |
"step": 128
|
| 778 |
},
|
| 779 |
{
|
| 780 |
+
"epoch": 0.08,
|
| 781 |
+
"learning_rate": 0.0001907745108354514,
|
| 782 |
+
"loss": 0.7342,
|
| 783 |
"step": 129
|
| 784 |
},
|
| 785 |
{
|
| 786 |
+
"epoch": 0.08,
|
| 787 |
+
"learning_rate": 0.00018921376456532482,
|
| 788 |
+
"loss": 0.7801,
|
| 789 |
"step": 130
|
| 790 |
},
|
| 791 |
{
|
| 792 |
+
"epoch": 0.08,
|
| 793 |
+
"learning_rate": 0.00018764844795965229,
|
| 794 |
+
"loss": 0.7579,
|
| 795 |
"step": 131
|
| 796 |
},
|
| 797 |
{
|
| 798 |
+
"epoch": 0.08,
|
| 799 |
+
"learning_rate": 0.00018607874345493805,
|
| 800 |
+
"loss": 0.6844,
|
| 801 |
"step": 132
|
| 802 |
},
|
| 803 |
{
|
| 804 |
+
"epoch": 0.08,
|
| 805 |
+
"learning_rate": 0.00018450483399909263,
|
| 806 |
+
"loss": 0.7052,
|
| 807 |
"step": 133
|
| 808 |
},
|
| 809 |
{
|
| 810 |
+
"epoch": 0.08,
|
| 811 |
+
"learning_rate": 0.00018292690303011076,
|
| 812 |
+
"loss": 0.6515,
|
| 813 |
"step": 134
|
| 814 |
},
|
| 815 |
{
|
| 816 |
+
"epoch": 0.08,
|
| 817 |
+
"learning_rate": 0.00018134513445469127,
|
| 818 |
+
"loss": 0.6846,
|
| 819 |
"step": 135
|
| 820 |
},
|
| 821 |
{
|
| 822 |
+
"epoch": 0.08,
|
| 823 |
+
"learning_rate": 0.00017975971262680347,
|
| 824 |
+
"loss": 0.6469,
|
| 825 |
"step": 136
|
| 826 |
},
|
| 827 |
{
|
| 828 |
+
"epoch": 0.08,
|
| 829 |
+
"learning_rate": 0.00017817082232620052,
|
| 830 |
+
"loss": 0.6703,
|
| 831 |
"step": 137
|
| 832 |
},
|
| 833 |
{
|
| 834 |
+
"epoch": 0.09,
|
| 835 |
+
"learning_rate": 0.00017657864873688343,
|
| 836 |
+
"loss": 0.6966,
|
| 837 |
"step": 138
|
| 838 |
},
|
| 839 |
{
|
| 840 |
+
"epoch": 0.09,
|
| 841 |
+
"learning_rate": 0.00017498337742551817,
|
| 842 |
+
"loss": 0.6274,
|
| 843 |
"step": 139
|
| 844 |
},
|
| 845 |
{
|
| 846 |
+
"epoch": 0.09,
|
| 847 |
+
"learning_rate": 0.00017338519431980796,
|
| 848 |
+
"loss": 0.6189,
|
| 849 |
"step": 140
|
| 850 |
},
|
| 851 |
{
|
| 852 |
+
"epoch": 0.09,
|
| 853 |
+
"learning_rate": 0.00017178428568682353,
|
| 854 |
+
"loss": 0.6881,
|
| 855 |
"step": 141
|
| 856 |
},
|
| 857 |
{
|
| 858 |
+
"epoch": 0.09,
|
| 859 |
+
"learning_rate": 0.0001701808381112938,
|
| 860 |
+
"loss": 0.6194,
|
| 861 |
"step": 142
|
| 862 |
},
|
| 863 |
{
|
| 864 |
+
"epoch": 0.09,
|
| 865 |
+
"learning_rate": 0.00016857503847385953,
|
| 866 |
+
"loss": 0.5674,
|
| 867 |
"step": 143
|
| 868 |
},
|
| 869 |
{
|
| 870 |
+
"epoch": 0.09,
|
| 871 |
+
"learning_rate": 0.00016696707392929266,
|
| 872 |
+
"loss": 0.5878,
|
| 873 |
"step": 144
|
| 874 |
},
|
| 875 |
{
|
| 876 |
+
"epoch": 0.09,
|
| 877 |
+
"learning_rate": 0.0001653571318846834,
|
| 878 |
+
"loss": 0.5229,
|
| 879 |
"step": 145
|
| 880 |
},
|
| 881 |
{
|
| 882 |
+
"epoch": 0.09,
|
| 883 |
+
"learning_rate": 0.00016374539997759821,
|
| 884 |
+
"loss": 0.535,
|
| 885 |
"step": 146
|
| 886 |
},
|
| 887 |
{
|
| 888 |
+
"epoch": 0.09,
|
| 889 |
+
"learning_rate": 0.00016213206605421063,
|
| 890 |
+
"loss": 0.4889,
|
| 891 |
"step": 147
|
| 892 |
},
|
| 893 |
{
|
| 894 |
+
"epoch": 0.09,
|
| 895 |
+
"learning_rate": 0.0001605173181474081,
|
| 896 |
+
"loss": 0.5009,
|
| 897 |
"step": 148
|
| 898 |
},
|
| 899 |
{
|
| 900 |
+
"epoch": 0.09,
|
| 901 |
+
"learning_rate": 0.00015890134445487676,
|
| 902 |
+
"loss": 0.4467,
|
| 903 |
"step": 149
|
| 904 |
},
|
| 905 |
{
|
| 906 |
+
"epoch": 0.09,
|
| 907 |
+
"learning_rate": 0.00015728433331716724,
|
| 908 |
+
"loss": 0.4275,
|
| 909 |
"step": 150
|
| 910 |
},
|
| 911 |
{
|
| 912 |
+
"epoch": 0.09,
|
| 913 |
+
"learning_rate": 0.0001556664731957435,
|
| 914 |
+
"loss": 0.7781,
|
| 915 |
"step": 151
|
| 916 |
},
|
| 917 |
{
|
| 918 |
+
"epoch": 0.09,
|
| 919 |
+
"learning_rate": 0.00015404795265101806,
|
| 920 |
+
"loss": 0.8081,
|
| 921 |
"step": 152
|
| 922 |
},
|
| 923 |
{
|
| 924 |
+
"epoch": 0.09,
|
| 925 |
+
"learning_rate": 0.00015242896032037522,
|
| 926 |
+
"loss": 0.8014,
|
| 927 |
"step": 153
|
| 928 |
},
|
| 929 |
{
|
| 930 |
+
"epoch": 0.1,
|
| 931 |
+
"learning_rate": 0.00015080968489618565,
|
| 932 |
+
"loss": 0.8031,
|
| 933 |
"step": 154
|
| 934 |
},
|
| 935 |
{
|
| 936 |
+
"epoch": 0.1,
|
| 937 |
+
"learning_rate": 0.00014919031510381435,
|
| 938 |
+
"loss": 0.8138,
|
| 939 |
"step": 155
|
| 940 |
},
|
| 941 |
{
|
| 942 |
+
"epoch": 0.1,
|
| 943 |
+
"learning_rate": 0.00014757103967962475,
|
| 944 |
+
"loss": 0.7578,
|
| 945 |
"step": 156
|
| 946 |
},
|
| 947 |
{
|
| 948 |
+
"epoch": 0.1,
|
| 949 |
+
"learning_rate": 0.00014595204734898197,
|
| 950 |
+
"loss": 0.8272,
|
| 951 |
"step": 157
|
| 952 |
},
|
| 953 |
{
|
| 954 |
+
"epoch": 0.1,
|
| 955 |
+
"learning_rate": 0.0001443335268042565,
|
| 956 |
+
"loss": 0.809,
|
| 957 |
"step": 158
|
| 958 |
},
|
| 959 |
{
|
| 960 |
+
"epoch": 0.1,
|
| 961 |
+
"learning_rate": 0.0001427156666828328,
|
| 962 |
+
"loss": 0.8103,
|
| 963 |
"step": 159
|
| 964 |
},
|
| 965 |
{
|
| 966 |
+
"epoch": 0.1,
|
| 967 |
+
"learning_rate": 0.00014109865554512319,
|
| 968 |
+
"loss": 0.7904,
|
| 969 |
"step": 160
|
| 970 |
},
|
| 971 |
{
|
| 972 |
+
"epoch": 0.1,
|
| 973 |
+
"learning_rate": 0.00013948268185259188,
|
| 974 |
+
"loss": 0.7967,
|
| 975 |
"step": 161
|
| 976 |
},
|
| 977 |
{
|
| 978 |
+
"epoch": 0.1,
|
| 979 |
+
"learning_rate": 0.00013786793394578937,
|
| 980 |
+
"loss": 0.8111,
|
| 981 |
"step": 162
|
| 982 |
},
|
| 983 |
{
|
| 984 |
+
"epoch": 0.1,
|
| 985 |
+
"learning_rate": 0.0001362546000224018,
|
| 986 |
+
"loss": 0.787,
|
| 987 |
"step": 163
|
| 988 |
},
|
| 989 |
{
|
| 990 |
+
"epoch": 0.1,
|
| 991 |
+
"learning_rate": 0.00013464286811531661,
|
| 992 |
+
"loss": 0.8282,
|
| 993 |
"step": 164
|
| 994 |
},
|
| 995 |
{
|
| 996 |
+
"epoch": 0.1,
|
| 997 |
+
"learning_rate": 0.00013303292607070737,
|
| 998 |
+
"loss": 0.7792,
|
| 999 |
"step": 165
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
+
"epoch": 0.1,
|
| 1003 |
+
"learning_rate": 0.0001314249615261405,
|
| 1004 |
+
"loss": 0.7955,
|
| 1005 |
"step": 166
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
+
"epoch": 0.1,
|
| 1009 |
+
"learning_rate": 0.0001298191618887062,
|
| 1010 |
+
"loss": 0.7874,
|
| 1011 |
"step": 167
|
| 1012 |
},
|
| 1013 |
{
|
| 1014 |
+
"epoch": 0.1,
|
| 1015 |
+
"learning_rate": 0.00012821571431317647,
|
| 1016 |
+
"loss": 0.7637,
|
| 1017 |
"step": 168
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
+
"epoch": 0.1,
|
| 1021 |
+
"learning_rate": 0.00012661480568019201,
|
| 1022 |
+
"loss": 0.7641,
|
| 1023 |
"step": 169
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
+
"epoch": 0.11,
|
| 1027 |
+
"learning_rate": 0.0001250166225744818,
|
| 1028 |
+
"loss": 0.8431,
|
| 1029 |
"step": 170
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
+
"epoch": 0.11,
|
| 1033 |
+
"learning_rate": 0.0001234213512631166,
|
| 1034 |
+
"loss": 0.8353,
|
| 1035 |
"step": 171
|
| 1036 |
},
|
| 1037 |
{
|
| 1038 |
+
"epoch": 0.11,
|
| 1039 |
+
"learning_rate": 0.00012182917767379948,
|
| 1040 |
+
"loss": 0.7732,
|
| 1041 |
"step": 172
|
| 1042 |
},
|
| 1043 |
{
|
| 1044 |
+
"epoch": 0.11,
|
| 1045 |
+
"learning_rate": 0.00012024028737319652,
|
| 1046 |
+
"loss": 0.789,
|
| 1047 |
"step": 173
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
+
"epoch": 0.11,
|
| 1051 |
+
"learning_rate": 0.00011865486554530873,
|
| 1052 |
+
"loss": 0.7473,
|
| 1053 |
"step": 174
|
| 1054 |
},
|
| 1055 |
{
|
| 1056 |
+
"epoch": 0.11,
|
| 1057 |
+
"learning_rate": 0.0001170730969698893,
|
| 1058 |
+
"loss": 0.8313,
|
| 1059 |
"step": 175
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
+
"epoch": 0.11,
|
| 1063 |
+
"learning_rate": 0.00011549516600090737,
|
| 1064 |
+
"loss": 0.855,
|
| 1065 |
"step": 176
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
+
"epoch": 0.11,
|
| 1069 |
+
"learning_rate": 0.00011392125654506198,
|
| 1070 |
+
"loss": 0.7981,
|
| 1071 |
"step": 177
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
+
"epoch": 0.11,
|
| 1075 |
+
"learning_rate": 0.00011235155204034767,
|
| 1076 |
+
"loss": 0.7148,
|
| 1077 |
"step": 178
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
+
"epoch": 0.11,
|
| 1081 |
+
"learning_rate": 0.00011078623543467518,
|
| 1082 |
+
"loss": 0.7012,
|
| 1083 |
"step": 179
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
+
"epoch": 0.11,
|
| 1087 |
+
"learning_rate": 0.00010922548916454855,
|
| 1088 |
+
"loss": 0.7313,
|
| 1089 |
"step": 180
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
+
"epoch": 0.11,
|
| 1093 |
+
"learning_rate": 0.00010766949513380284,
|
| 1094 |
+
"loss": 0.6691,
|
| 1095 |
"step": 181
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
+
"epoch": 0.11,
|
| 1099 |
+
"learning_rate": 0.00010611843469240288,
|
| 1100 |
+
"loss": 0.658,
|
| 1101 |
"step": 182
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
+
"epoch": 0.11,
|
| 1105 |
+
"learning_rate": 0.00010457248861530741,
|
| 1106 |
+
"loss": 0.7276,
|
| 1107 |
"step": 183
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
+
"epoch": 0.11,
|
| 1111 |
+
"learning_rate": 0.00010303183708139964,
|
| 1112 |
+
"loss": 0.732,
|
| 1113 |
"step": 184
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
+
"epoch": 0.11,
|
| 1117 |
+
"learning_rate": 0.00010149665965248775,
|
| 1118 |
+
"loss": 0.6751,
|
| 1119 |
"step": 185
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
+
"epoch": 0.11,
|
| 1123 |
+
"learning_rate": 9.996713525237694e-05,
|
| 1124 |
+
"loss": 0.6968,
|
| 1125 |
"step": 186
|
| 1126 |
},
|
| 1127 |
{
|
| 1128 |
+
"epoch": 0.12,
|
| 1129 |
+
"learning_rate": 9.8443442146016e-05,
|
| 1130 |
+
"loss": 0.6131,
|
| 1131 |
"step": 187
|
| 1132 |
},
|
| 1133 |
{
|
| 1134 |
+
"epoch": 0.12,
|
| 1135 |
+
"learning_rate": 9.692575791872089e-05,
|
| 1136 |
+
"loss": 0.6497,
|
| 1137 |
"step": 188
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
+
"epoch": 0.12,
|
| 1141 |
+
"learning_rate": 9.541425945547687e-05,
|
| 1142 |
+
"loss": 0.5895,
|
| 1143 |
"step": 189
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
+
"epoch": 0.12,
|
| 1147 |
+
"learning_rate": 9.390912292032309e-05,
|
| 1148 |
+
"loss": 0.6346,
|
| 1149 |
"step": 190
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
+
"epoch": 0.12,
|
| 1153 |
+
"learning_rate": 9.241052373582057e-05,
|
| 1154 |
+
"loss": 0.5706,
|
| 1155 |
"step": 191
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
+
"epoch": 0.12,
|
| 1159 |
+
"learning_rate": 9.091863656260695e-05,
|
| 1160 |
+
"loss": 0.5917,
|
| 1161 |
"step": 192
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
+
"epoch": 0.12,
|
| 1165 |
+
"learning_rate": 8.943363527903976e-05,
|
| 1166 |
+
"loss": 0.5934,
|
| 1167 |
"step": 193
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
+
"epoch": 0.12,
|
| 1171 |
+
"learning_rate": 8.795569296093132e-05,
|
| 1172 |
+
"loss": 0.5427,
|
| 1173 |
"step": 194
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
+
"epoch": 0.12,
|
| 1177 |
+
"learning_rate": 8.648498186137653e-05,
|
| 1178 |
+
"loss": 0.5419,
|
| 1179 |
"step": 195
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
+
"epoch": 0.12,
|
| 1183 |
+
"learning_rate": 8.502167339067705e-05,
|
| 1184 |
+
"loss": 0.5431,
|
| 1185 |
"step": 196
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
+
"epoch": 0.12,
|
| 1189 |
+
"learning_rate": 8.356593809636371e-05,
|
| 1190 |
+
"loss": 0.4831,
|
| 1191 |
"step": 197
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
+
"epoch": 0.12,
|
| 1195 |
+
"learning_rate": 8.211794564331917e-05,
|
| 1196 |
+
"loss": 0.5207,
|
| 1197 |
"step": 198
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
+
"epoch": 0.12,
|
| 1201 |
+
"learning_rate": 8.067786479400346e-05,
|
| 1202 |
+
"loss": 0.4644,
|
| 1203 |
"step": 199
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
+
"epoch": 0.12,
|
| 1207 |
+
"learning_rate": 7.924586338878511e-05,
|
| 1208 |
+
"loss": 0.5012,
|
| 1209 |
"step": 200
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
+
"epoch": 0.12,
|
| 1213 |
+
"learning_rate": 7.782210832637923e-05,
|
| 1214 |
+
"loss": 0.7555,
|
| 1215 |
"step": 201
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
+
"epoch": 0.12,
|
| 1219 |
+
"learning_rate": 7.640676554439594e-05,
|
| 1220 |
+
"loss": 0.7989,
|
| 1221 |
"step": 202
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
+
"epoch": 0.13,
|
| 1225 |
+
"learning_rate": 7.500000000000002e-05,
|
| 1226 |
+
"loss": 0.7734,
|
| 1227 |
"step": 203
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
+
"epoch": 0.13,
|
| 1231 |
+
"learning_rate": 7.36019756506856e-05,
|
| 1232 |
+
"loss": 0.7744,
|
| 1233 |
"step": 204
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
+
"epoch": 0.13,
|
| 1237 |
+
"learning_rate": 7.22128554351668e-05,
|
| 1238 |
+
"loss": 0.768,
|
| 1239 |
"step": 205
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
+
"epoch": 0.13,
|
| 1243 |
+
"learning_rate": 7.083280125438766e-05,
|
| 1244 |
+
"loss": 0.775,
|
| 1245 |
"step": 206
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
+
"epoch": 0.13,
|
| 1249 |
+
"learning_rate": 6.946197395265242e-05,
|
| 1250 |
+
"loss": 0.8156,
|
| 1251 |
"step": 207
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
+
"epoch": 0.13,
|
| 1255 |
+
"learning_rate": 6.810053329887928e-05,
|
| 1256 |
+
"loss": 0.7672,
|
| 1257 |
"step": 208
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
+
"epoch": 0.13,
|
| 1261 |
+
"learning_rate": 6.674863796797953e-05,
|
| 1262 |
+
"loss": 0.8119,
|
| 1263 |
"step": 209
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
+
"epoch": 0.13,
|
| 1267 |
+
"learning_rate": 6.540644552236401e-05,
|
| 1268 |
+
"loss": 0.8254,
|
| 1269 |
"step": 210
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
+
"epoch": 0.13,
|
| 1273 |
+
"learning_rate": 6.407411239357953e-05,
|
| 1274 |
+
"loss": 0.7838,
|
| 1275 |
"step": 211
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
+
"epoch": 0.13,
|
| 1279 |
+
"learning_rate": 6.275179386407663e-05,
|
| 1280 |
+
"loss": 0.7809,
|
| 1281 |
"step": 212
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
+
"epoch": 0.13,
|
| 1285 |
+
"learning_rate": 6.143964404911164e-05,
|
| 1286 |
+
"loss": 0.7387,
|
| 1287 |
"step": 213
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
+
"epoch": 0.13,
|
| 1291 |
+
"learning_rate": 6.013781587878463e-05,
|
| 1292 |
+
"loss": 0.7791,
|
| 1293 |
"step": 214
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
+
"epoch": 0.13,
|
| 1297 |
+
"learning_rate": 5.8846461080215626e-05,
|
| 1298 |
+
"loss": 0.8441,
|
| 1299 |
"step": 215
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
+
"epoch": 0.13,
|
| 1303 |
+
"learning_rate": 5.756573015986089e-05,
|
| 1304 |
+
"loss": 0.8518,
|
| 1305 |
"step": 216
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
+
"epoch": 0.13,
|
| 1309 |
+
"learning_rate": 5.629577238597132e-05,
|
| 1310 |
+
"loss": 0.7721,
|
| 1311 |
"step": 217
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
+
"epoch": 0.13,
|
| 1315 |
+
"learning_rate": 5.503673577119552e-05,
|
| 1316 |
+
"loss": 0.7412,
|
| 1317 |
"step": 218
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
+
"epoch": 0.14,
|
| 1321 |
+
"learning_rate": 5.378876705532904e-05,
|
| 1322 |
+
"loss": 0.8937,
|
| 1323 |
"step": 219
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
+
"epoch": 0.14,
|
| 1327 |
+
"learning_rate": 5.2552011688211835e-05,
|
| 1328 |
+
"loss": 0.823,
|
| 1329 |
"step": 220
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
+
"epoch": 0.14,
|
| 1333 |
+
"learning_rate": 5.1326613812776434e-05,
|
| 1334 |
+
"loss": 0.7921,
|
| 1335 |
"step": 221
|
| 1336 |
},
|
| 1337 |
{
|
| 1338 |
+
"epoch": 0.14,
|
| 1339 |
+
"learning_rate": 5.011271624824786e-05,
|
| 1340 |
+
"loss": 0.7972,
|
| 1341 |
"step": 222
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
+
"epoch": 0.14,
|
| 1345 |
+
"learning_rate": 4.891046047349837e-05,
|
| 1346 |
+
"loss": 0.8309,
|
| 1347 |
"step": 223
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
+
"epoch": 0.14,
|
| 1351 |
+
"learning_rate": 4.7719986610558234e-05,
|
| 1352 |
+
"loss": 0.7832,
|
| 1353 |
"step": 224
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
+
"epoch": 0.14,
|
| 1357 |
+
"learning_rate": 4.654143340828435e-05,
|
| 1358 |
+
"loss": 0.7451,
|
| 1359 |
"step": 225
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
+
"epoch": 0.14,
|
| 1363 |
+
"learning_rate": 4.537493822618958e-05,
|
| 1364 |
+
"loss": 0.8234,
|
| 1365 |
"step": 226
|
| 1366 |
},
|
| 1367 |
{
|
| 1368 |
+
"epoch": 0.14,
|
| 1369 |
+
"learning_rate": 4.422063701843316e-05,
|
| 1370 |
+
"loss": 0.7882,
|
| 1371 |
"step": 227
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
+
"epoch": 0.14,
|
| 1375 |
+
"learning_rate": 4.3078664317975646e-05,
|
| 1376 |
+
"loss": 0.8251,
|
| 1377 |
"step": 228
|
| 1378 |
},
|
| 1379 |
{
|
| 1380 |
+
"epoch": 0.14,
|
| 1381 |
+
"learning_rate": 4.194915322089898e-05,
|
| 1382 |
+
"loss": 0.8103,
|
| 1383 |
"step": 229
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
+
"epoch": 0.14,
|
| 1387 |
+
"learning_rate": 4.08322353708946e-05,
|
| 1388 |
+
"loss": 0.7449,
|
| 1389 |
"step": 230
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
+
"epoch": 0.14,
|
| 1393 |
+
"learning_rate": 3.972804094391998e-05,
|
| 1394 |
+
"loss": 0.75,
|
| 1395 |
"step": 231
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
+
"epoch": 0.14,
|
| 1399 |
+
"learning_rate": 3.863669863302697e-05,
|
| 1400 |
+
"loss": 0.7377,
|
| 1401 |
"step": 232
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
+
"epoch": 0.14,
|
| 1405 |
+
"learning_rate": 3.755833563336293e-05,
|
| 1406 |
+
"loss": 0.7428,
|
| 1407 |
"step": 233
|
| 1408 |
},
|
| 1409 |
{
|
| 1410 |
+
"epoch": 0.14,
|
| 1411 |
+
"learning_rate": 3.64930776273457e-05,
|
| 1412 |
+
"loss": 0.7281,
|
| 1413 |
"step": 234
|
| 1414 |
},
|
| 1415 |
{
|
| 1416 |
+
"epoch": 0.15,
|
| 1417 |
+
"learning_rate": 3.5441048770015954e-05,
|
| 1418 |
+
"loss": 0.676,
|
| 1419 |
"step": 235
|
| 1420 |
},
|
| 1421 |
{
|
| 1422 |
+
"epoch": 0.15,
|
| 1423 |
+
"learning_rate": 3.4402371674566626e-05,
|
| 1424 |
+
"loss": 0.6793,
|
| 1425 |
"step": 236
|
| 1426 |
},
|
| 1427 |
{
|
| 1428 |
+
"epoch": 0.15,
|
| 1429 |
+
"learning_rate": 3.3377167398052636e-05,
|
| 1430 |
+
"loss": 0.6211,
|
| 1431 |
"step": 237
|
| 1432 |
},
|
| 1433 |
{
|
| 1434 |
+
"epoch": 0.15,
|
| 1435 |
+
"learning_rate": 3.2365555427281634e-05,
|
| 1436 |
+
"loss": 0.5825,
|
| 1437 |
"step": 238
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
+
"epoch": 0.15,
|
| 1441 |
+
"learning_rate": 3.136765366488817e-05,
|
| 1442 |
+
"loss": 0.6659,
|
| 1443 |
"step": 239
|
| 1444 |
},
|
| 1445 |
{
|
| 1446 |
+
"epoch": 0.15,
|
| 1447 |
+
"learning_rate": 3.038357841559191e-05,
|
| 1448 |
+
"loss": 0.6726,
|
| 1449 |
"step": 240
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
+
"epoch": 0.15,
|
| 1453 |
+
"learning_rate": 2.941344437264249e-05,
|
| 1454 |
+
"loss": 0.595,
|
| 1455 |
"step": 241
|
| 1456 |
},
|
| 1457 |
{
|
| 1458 |
+
"epoch": 0.15,
|
| 1459 |
+
"learning_rate": 2.8457364604452372e-05,
|
| 1460 |
+
"loss": 0.5769,
|
| 1461 |
"step": 242
|
| 1462 |
},
|
| 1463 |
{
|
| 1464 |
+
"epoch": 0.15,
|
| 1465 |
+
"learning_rate": 2.7515450541418338e-05,
|
| 1466 |
+
"loss": 0.5748,
|
| 1467 |
"step": 243
|
| 1468 |
},
|
| 1469 |
{
|
| 1470 |
+
"epoch": 0.15,
|
| 1471 |
+
"learning_rate": 2.658781196293482e-05,
|
| 1472 |
+
"loss": 0.5505,
|
| 1473 |
"step": 244
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
+
"epoch": 0.15,
|
| 1477 |
+
"learning_rate": 2.5674556984598822e-05,
|
| 1478 |
+
"loss": 0.5472,
|
| 1479 |
"step": 245
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
+
"epoch": 0.15,
|
| 1483 |
+
"learning_rate": 2.477579204560935e-05,
|
| 1484 |
+
"loss": 0.518,
|
| 1485 |
"step": 246
|
| 1486 |
},
|
| 1487 |
{
|
| 1488 |
+
"epoch": 0.15,
|
| 1489 |
+
"learning_rate": 2.389162189636188e-05,
|
| 1490 |
+
"loss": 0.5002,
|
| 1491 |
"step": 247
|
| 1492 |
},
|
| 1493 |
{
|
| 1494 |
+
"epoch": 0.15,
|
| 1495 |
+
"learning_rate": 2.3022149586239968e-05,
|
| 1496 |
+
"loss": 0.4444,
|
| 1497 |
"step": 248
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
+
"epoch": 0.15,
|
| 1501 |
+
"learning_rate": 2.216747645160462e-05,
|
| 1502 |
+
"loss": 0.5107,
|
| 1503 |
"step": 249
|
| 1504 |
},
|
| 1505 |
{
|
| 1506 |
+
"epoch": 0.15,
|
| 1507 |
+
"learning_rate": 2.1327702103983863e-05,
|
| 1508 |
+
"loss": 0.4729,
|
| 1509 |
"step": 250
|
| 1510 |
},
|
| 1511 |
{
|
| 1512 |
+
"epoch": 0.16,
|
| 1513 |
+
"learning_rate": 2.0502924418463013e-05,
|
| 1514 |
+
"loss": 0.7721,
|
| 1515 |
"step": 251
|
| 1516 |
},
|
| 1517 |
{
|
| 1518 |
+
"epoch": 0.16,
|
| 1519 |
+
"learning_rate": 1.9693239522277327e-05,
|
| 1520 |
+
"loss": 0.774,
|
| 1521 |
"step": 252
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
+
"epoch": 0.16,
|
| 1525 |
+
"learning_rate": 1.889874178360864e-05,
|
| 1526 |
+
"loss": 0.8274,
|
| 1527 |
"step": 253
|
| 1528 |
},
|
| 1529 |
{
|
| 1530 |
+
"epoch": 0.16,
|
| 1531 |
+
"learning_rate": 1.8119523800586568e-05,
|
| 1532 |
+
"loss": 0.7989,
|
| 1533 |
"step": 254
|
| 1534 |
},
|
| 1535 |
{
|
| 1536 |
+
"epoch": 0.16,
|
| 1537 |
+
"learning_rate": 1.735567639049648e-05,
|
| 1538 |
+
"loss": 0.7691,
|
| 1539 |
"step": 255
|
| 1540 |
},
|
| 1541 |
{
|
| 1542 |
+
"epoch": 0.16,
|
| 1543 |
+
"learning_rate": 1.6607288579194638e-05,
|
| 1544 |
+
"loss": 0.7938,
|
| 1545 |
"step": 256
|
| 1546 |
},
|
| 1547 |
{
|
| 1548 |
+
"epoch": 0.16,
|
| 1549 |
+
"learning_rate": 1.5874447590732538e-05,
|
| 1550 |
+
"loss": 0.8506,
|
| 1551 |
"step": 257
|
| 1552 |
},
|
| 1553 |
{
|
| 1554 |
+
"epoch": 0.16,
|
| 1555 |
+
"learning_rate": 1.5157238837190716e-05,
|
| 1556 |
+
"loss": 0.7695,
|
| 1557 |
"step": 258
|
| 1558 |
},
|
| 1559 |
{
|
| 1560 |
+
"epoch": 0.16,
|
| 1561 |
+
"learning_rate": 1.4455745908724226e-05,
|
| 1562 |
+
"loss": 0.819,
|
| 1563 |
"step": 259
|
| 1564 |
},
|
| 1565 |
{
|
| 1566 |
+
"epoch": 0.16,
|
| 1567 |
+
"learning_rate": 1.3770050563820179e-05,
|
| 1568 |
+
"loss": 0.7904,
|
| 1569 |
"step": 260
|
| 1570 |
},
|
| 1571 |
{
|
| 1572 |
+
"epoch": 0.16,
|
| 1573 |
+
"learning_rate": 1.3100232719768994e-05,
|
| 1574 |
+
"loss": 0.8647,
|
| 1575 |
"step": 261
|
| 1576 |
},
|
| 1577 |
{
|
| 1578 |
+
"epoch": 0.16,
|
| 1579 |
+
"learning_rate": 1.2446370443349863e-05,
|
| 1580 |
+
"loss": 0.777,
|
| 1581 |
"step": 262
|
| 1582 |
},
|
| 1583 |
{
|
| 1584 |
+
"epoch": 0.16,
|
| 1585 |
+
"learning_rate": 1.180853994173236e-05,
|
| 1586 |
+
"loss": 0.7898,
|
| 1587 |
"step": 263
|
| 1588 |
},
|
| 1589 |
{
|
| 1590 |
+
"epoch": 0.16,
|
| 1591 |
+
"learning_rate": 1.118681555359438e-05,
|
| 1592 |
+
"loss": 0.7466,
|
| 1593 |
"step": 264
|
| 1594 |
},
|
| 1595 |
{
|
| 1596 |
+
"epoch": 0.16,
|
| 1597 |
+
"learning_rate": 1.058126974045811e-05,
|
| 1598 |
+
"loss": 0.8193,
|
| 1599 |
"step": 265
|
| 1600 |
},
|
| 1601 |
{
|
| 1602 |
+
"epoch": 0.16,
|
| 1603 |
+
"learning_rate": 9.991973078244636e-06,
|
| 1604 |
+
"loss": 0.8072,
|
| 1605 |
"step": 266
|
| 1606 |
},
|
| 1607 |
{
|
| 1608 |
+
"epoch": 0.17,
|
| 1609 |
+
"learning_rate": 9.418994249048472e-06,
|
| 1610 |
+
"loss": 0.8395,
|
| 1611 |
"step": 267
|
| 1612 |
},
|
| 1613 |
{
|
| 1614 |
+
"epoch": 0.17,
|
| 1615 |
+
"learning_rate": 8.862400033132571e-06,
|
| 1616 |
+
"loss": 0.8109,
|
| 1617 |
"step": 268
|
| 1618 |
},
|
| 1619 |
{
|
| 1620 |
+
"epoch": 0.17,
|
| 1621 |
+
"learning_rate": 8.322255301145204e-06,
|
| 1622 |
+
"loss": 0.8379,
|
| 1623 |
"step": 269
|
| 1624 |
},
|
| 1625 |
{
|
| 1626 |
+
"epoch": 0.17,
|
| 1627 |
+
"learning_rate": 7.798623006559435e-06,
|
| 1628 |
+
"loss": 0.7609,
|
| 1629 |
"step": 270
|
| 1630 |
},
|
| 1631 |
{
|
| 1632 |
+
"epoch": 0.17,
|
| 1633 |
+
"learning_rate": 7.291564178335718e-06,
|
| 1634 |
+
"loss": 0.7881,
|
| 1635 |
"step": 271
|
| 1636 |
},
|
| 1637 |
{
|
| 1638 |
+
"epoch": 0.17,
|
| 1639 |
+
"learning_rate": 6.801137913809213e-06,
|
| 1640 |
+
"loss": 0.8359,
|
| 1641 |
"step": 272
|
| 1642 |
},
|
| 1643 |
{
|
| 1644 |
+
"epoch": 0.17,
|
| 1645 |
+
"learning_rate": 6.3274013718019434e-06,
|
| 1646 |
+
"loss": 0.8019,
|
| 1647 |
"step": 273
|
| 1648 |
},
|
| 1649 |
{
|
| 1650 |
+
"epoch": 0.17,
|
| 1651 |
+
"learning_rate": 5.870409765960965e-06,
|
| 1652 |
+
"loss": 0.7971,
|
| 1653 |
"step": 274
|
| 1654 |
},
|
| 1655 |
{
|
| 1656 |
+
"epoch": 0.17,
|
| 1657 |
+
"learning_rate": 5.430216358323309e-06,
|
| 1658 |
+
"loss": 0.7786,
|
| 1659 |
"step": 275
|
| 1660 |
},
|
| 1661 |
{
|
| 1662 |
+
"epoch": 0.17,
|
| 1663 |
+
"learning_rate": 5.006872453108329e-06,
|
| 1664 |
+
"loss": 0.7921,
|
| 1665 |
"step": 276
|
| 1666 |
},
|
| 1667 |
{
|
| 1668 |
+
"epoch": 0.17,
|
| 1669 |
+
"learning_rate": 4.600427390738159e-06,
|
| 1670 |
+
"loss": 0.7785,
|
| 1671 |
"step": 277
|
| 1672 |
},
|
| 1673 |
{
|
| 1674 |
+
"epoch": 0.17,
|
| 1675 |
+
"learning_rate": 4.210928542087206e-06,
|
| 1676 |
+
"loss": 0.7637,
|
| 1677 |
"step": 278
|
| 1678 |
},
|
| 1679 |
{
|
| 1680 |
+
"epoch": 0.17,
|
| 1681 |
+
"learning_rate": 3.838421302961098e-06,
|
| 1682 |
+
"loss": 0.7673,
|
| 1683 |
"step": 279
|
| 1684 |
},
|
| 1685 |
{
|
| 1686 |
+
"epoch": 0.17,
|
| 1687 |
+
"learning_rate": 3.482949088805742e-06,
|
| 1688 |
+
"loss": 0.7417,
|
| 1689 |
"step": 280
|
| 1690 |
},
|
| 1691 |
{
|
| 1692 |
+
"epoch": 0.17,
|
| 1693 |
+
"learning_rate": 3.1445533296474478e-06,
|
| 1694 |
+
"loss": 0.7175,
|
| 1695 |
"step": 281
|
| 1696 |
},
|
| 1697 |
{
|
| 1698 |
+
"epoch": 0.17,
|
| 1699 |
+
"learning_rate": 2.823273465264142e-06,
|
| 1700 |
+
"loss": 0.6789,
|
| 1701 |
"step": 282
|
| 1702 |
},
|
| 1703 |
{
|
| 1704 |
+
"epoch": 0.17,
|
| 1705 |
+
"learning_rate": 2.519146940588762e-06,
|
| 1706 |
+
"loss": 0.6716,
|
| 1707 |
"step": 283
|
| 1708 |
},
|
| 1709 |
{
|
| 1710 |
+
"epoch": 0.18,
|
| 1711 |
+
"learning_rate": 2.232209201345031e-06,
|
| 1712 |
+
"loss": 0.6743,
|
| 1713 |
"step": 284
|
| 1714 |
},
|
| 1715 |
{
|
| 1716 |
+
"epoch": 0.18,
|
| 1717 |
+
"learning_rate": 1.9624936899163945e-06,
|
| 1718 |
+
"loss": 0.673,
|
| 1719 |
"step": 285
|
| 1720 |
},
|
| 1721 |
{
|
| 1722 |
+
"epoch": 0.18,
|
| 1723 |
+
"learning_rate": 1.7100318414482061e-06,
|
| 1724 |
+
"loss": 0.6362,
|
| 1725 |
"step": 286
|
| 1726 |
},
|
| 1727 |
{
|
| 1728 |
+
"epoch": 0.18,
|
| 1729 |
+
"learning_rate": 1.4748530801840074e-06,
|
| 1730 |
+
"loss": 0.673,
|
| 1731 |
"step": 287
|
| 1732 |
},
|
| 1733 |
{
|
| 1734 |
+
"epoch": 0.18,
|
| 1735 |
+
"learning_rate": 1.2569848160362384e-06,
|
| 1736 |
+
"loss": 0.6348,
|
| 1737 |
"step": 288
|
| 1738 |
},
|
| 1739 |
{
|
| 1740 |
+
"epoch": 0.18,
|
| 1741 |
+
"learning_rate": 1.056452441391542e-06,
|
| 1742 |
+
"loss": 0.6603,
|
| 1743 |
"step": 289
|
| 1744 |
},
|
| 1745 |
{
|
| 1746 |
+
"epoch": 0.18,
|
| 1747 |
+
"learning_rate": 8.732793281513661e-07,
|
| 1748 |
+
"loss": 0.5877,
|
| 1749 |
"step": 290
|
| 1750 |
},
|
| 1751 |
{
|
| 1752 |
+
"epoch": 0.18,
|
| 1753 |
+
"learning_rate": 7.07486825007908e-07,
|
| 1754 |
+
"loss": 0.6369,
|
| 1755 |
"step": 291
|
| 1756 |
},
|
| 1757 |
{
|
| 1758 |
+
"epoch": 0.18,
|
| 1759 |
+
"learning_rate": 5.590942549560051e-07,
|
| 1760 |
+
"loss": 0.5664,
|
| 1761 |
"step": 292
|
| 1762 |
},
|
| 1763 |
{
|
| 1764 |
+
"epoch": 0.18,
|
| 1765 |
+
"learning_rate": 4.281189130410534e-07,
|
| 1766 |
+
"loss": 0.5721,
|
| 1767 |
"step": 293
|
| 1768 |
},
|
| 1769 |
{
|
| 1770 |
+
"epoch": 0.18,
|
| 1771 |
+
"learning_rate": 3.1457606434325266e-07,
|
| 1772 |
+
"loss": 0.5755,
|
| 1773 |
"step": 294
|
| 1774 |
},
|
| 1775 |
{
|
| 1776 |
+
"epoch": 0.18,
|
| 1777 |
+
"learning_rate": 2.184789421984634e-07,
|
| 1778 |
+
"loss": 0.5584,
|
| 1779 |
"step": 295
|
| 1780 |
},
|
| 1781 |
{
|
| 1782 |
+
"epoch": 0.18,
|
| 1783 |
+
"learning_rate": 1.3983874665589035e-07,
|
| 1784 |
+
"loss": 0.5746,
|
| 1785 |
"step": 296
|
| 1786 |
},
|
| 1787 |
{
|
| 1788 |
+
"epoch": 0.18,
|
| 1789 |
+
"learning_rate": 7.866464317276e-08,
|
| 1790 |
+
"loss": 0.4979,
|
| 1791 |
"step": 297
|
| 1792 |
},
|
| 1793 |
{
|
| 1794 |
+
"epoch": 0.18,
|
| 1795 |
+
"learning_rate": 3.4963761546041855e-08,
|
| 1796 |
+
"loss": 0.5017,
|
| 1797 |
"step": 298
|
| 1798 |
},
|
| 1799 |
{
|
| 1800 |
+
"epoch": 0.18,
|
| 1801 |
+
"learning_rate": 8.741195081479747e-09,
|
| 1802 |
+
"loss": 0.5112,
|
| 1803 |
"step": 299
|
| 1804 |
},
|
| 1805 |
{
|
| 1806 |
+
"epoch": 0.19,
|
| 1807 |
"learning_rate": 0.0,
|
| 1808 |
+
"loss": 0.453,
|
| 1809 |
"step": 300
|
| 1810 |
}
|
| 1811 |
],
|
| 1812 |
+
"logging_steps": 1,
|
| 1813 |
"max_steps": 300,
|
| 1814 |
"num_train_epochs": 1,
|
| 1815 |
+
"save_steps": 50,
|
| 1816 |
+
"total_flos": 8.93034880303104e+16,
|
| 1817 |
"trial_name": null,
|
| 1818 |
"trial_params": null
|
| 1819 |
}
|
checkpoint-300/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4027
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d1c02534982789891a108fd8845b01f09dd0f60e3bcbbfb171714be6598e3a93
|
| 3 |
size 4027
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4027
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d1c02534982789891a108fd8845b01f09dd0f60e3bcbbfb171714be6598e3a93
|
| 3 |
size 4027
|