Training in progress, step 500, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 578859568
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c6519a65749b3640a198541c50a2d98674828ec6c9d6a4383408f62deb368efa
|
| 3 |
size 578859568
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 294324692
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:366e0623982882f9616a1c85db3efb50f3266236a2f25d9632bdc591d388e3eb
|
| 3 |
size 294324692
|
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 sha256:5ceff29ed030645c389d768390ebf4a4817b18da01053a3a2c553da3856524de
|
| 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 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90b3c6a1abf04b992261ec3713fab3774b78fe3ab6f52bec78378c6c0a4a1110
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 50,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3237,6 +3237,364 @@
|
|
| 3237 |
"eval_samples_per_second": 11.548,
|
| 3238 |
"eval_steps_per_second": 5.784,
|
| 3239 |
"step": 450
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3240 |
}
|
| 3241 |
],
|
| 3242 |
"logging_steps": 1,
|
|
@@ -3260,12 +3618,12 @@
|
|
| 3260 |
"should_evaluate": false,
|
| 3261 |
"should_log": false,
|
| 3262 |
"should_save": true,
|
| 3263 |
-
"should_training_stop":
|
| 3264 |
},
|
| 3265 |
"attributes": {}
|
| 3266 |
}
|
| 3267 |
},
|
| 3268 |
-
"total_flos":
|
| 3269 |
"train_batch_size": 2,
|
| 3270 |
"trial_name": null,
|
| 3271 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.6910951733589172,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-500",
|
| 4 |
+
"epoch": 0.013898444069186455,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3237 |
"eval_samples_per_second": 11.548,
|
| 3238 |
"eval_steps_per_second": 5.784,
|
| 3239 |
"step": 450
|
| 3240 |
+
},
|
| 3241 |
+
{
|
| 3242 |
+
"epoch": 0.012536396550406182,
|
| 3243 |
+
"grad_norm": 0.08057525753974915,
|
| 3244 |
+
"learning_rate": 4.894348370484647e-05,
|
| 3245 |
+
"loss": 0.8512,
|
| 3246 |
+
"step": 451
|
| 3247 |
+
},
|
| 3248 |
+
{
|
| 3249 |
+
"epoch": 0.012564193438544554,
|
| 3250 |
+
"grad_norm": 0.08833472430706024,
|
| 3251 |
+
"learning_rate": 4.698180862871282e-05,
|
| 3252 |
+
"loss": 0.995,
|
| 3253 |
+
"step": 452
|
| 3254 |
+
},
|
| 3255 |
+
{
|
| 3256 |
+
"epoch": 0.012591990326682928,
|
| 3257 |
+
"grad_norm": 0.08962654322385788,
|
| 3258 |
+
"learning_rate": 4.505930839770966e-05,
|
| 3259 |
+
"loss": 0.9639,
|
| 3260 |
+
"step": 453
|
| 3261 |
+
},
|
| 3262 |
+
{
|
| 3263 |
+
"epoch": 0.0126197872148213,
|
| 3264 |
+
"grad_norm": 0.08324563503265381,
|
| 3265 |
+
"learning_rate": 4.3176062038291274e-05,
|
| 3266 |
+
"loss": 0.7867,
|
| 3267 |
+
"step": 454
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 0.012647584102959674,
|
| 3271 |
+
"grad_norm": 0.0913391187787056,
|
| 3272 |
+
"learning_rate": 4.1332146963339423e-05,
|
| 3273 |
+
"loss": 0.7687,
|
| 3274 |
+
"step": 455
|
| 3275 |
+
},
|
| 3276 |
+
{
|
| 3277 |
+
"epoch": 0.012675380991098047,
|
| 3278 |
+
"grad_norm": 0.08055524528026581,
|
| 3279 |
+
"learning_rate": 3.952763896898071e-05,
|
| 3280 |
+
"loss": 0.7876,
|
| 3281 |
+
"step": 456
|
| 3282 |
+
},
|
| 3283 |
+
{
|
| 3284 |
+
"epoch": 0.01270317787923642,
|
| 3285 |
+
"grad_norm": 0.08405828475952148,
|
| 3286 |
+
"learning_rate": 3.776261223147126e-05,
|
| 3287 |
+
"loss": 0.7312,
|
| 3288 |
+
"step": 457
|
| 3289 |
+
},
|
| 3290 |
+
{
|
| 3291 |
+
"epoch": 0.012730974767374793,
|
| 3292 |
+
"grad_norm": 0.08189340680837631,
|
| 3293 |
+
"learning_rate": 3.603713930414676e-05,
|
| 3294 |
+
"loss": 0.7638,
|
| 3295 |
+
"step": 458
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 0.012758771655513165,
|
| 3299 |
+
"grad_norm": 0.08689261227846146,
|
| 3300 |
+
"learning_rate": 3.435129111444113e-05,
|
| 3301 |
+
"loss": 0.7503,
|
| 3302 |
+
"step": 459
|
| 3303 |
+
},
|
| 3304 |
+
{
|
| 3305 |
+
"epoch": 0.012786568543651539,
|
| 3306 |
+
"grad_norm": 0.08499288558959961,
|
| 3307 |
+
"learning_rate": 3.270513696097055e-05,
|
| 3308 |
+
"loss": 0.7546,
|
| 3309 |
+
"step": 460
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 0.01281436543178991,
|
| 3313 |
+
"grad_norm": 0.08908785879611969,
|
| 3314 |
+
"learning_rate": 3.109874451068473e-05,
|
| 3315 |
+
"loss": 0.7841,
|
| 3316 |
+
"step": 461
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"epoch": 0.012842162319928284,
|
| 3320 |
+
"grad_norm": 0.08109744638204575,
|
| 3321 |
+
"learning_rate": 2.9532179796085356e-05,
|
| 3322 |
+
"loss": 0.6244,
|
| 3323 |
+
"step": 462
|
| 3324 |
+
},
|
| 3325 |
+
{
|
| 3326 |
+
"epoch": 0.012869959208066656,
|
| 3327 |
+
"grad_norm": 0.08815158903598785,
|
| 3328 |
+
"learning_rate": 2.800550721251216e-05,
|
| 3329 |
+
"loss": 0.697,
|
| 3330 |
+
"step": 463
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 0.01289775609620503,
|
| 3334 |
+
"grad_norm": 0.09392455220222473,
|
| 3335 |
+
"learning_rate": 2.6518789515495355e-05,
|
| 3336 |
+
"loss": 0.6835,
|
| 3337 |
+
"step": 464
|
| 3338 |
+
},
|
| 3339 |
+
{
|
| 3340 |
+
"epoch": 0.012925552984343404,
|
| 3341 |
+
"grad_norm": 0.1170358955860138,
|
| 3342 |
+
"learning_rate": 2.5072087818176382e-05,
|
| 3343 |
+
"loss": 0.938,
|
| 3344 |
+
"step": 465
|
| 3345 |
+
},
|
| 3346 |
+
{
|
| 3347 |
+
"epoch": 0.012953349872481775,
|
| 3348 |
+
"grad_norm": 0.09133084863424301,
|
| 3349 |
+
"learning_rate": 2.36654615887959e-05,
|
| 3350 |
+
"loss": 0.7392,
|
| 3351 |
+
"step": 466
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 0.01298114676062015,
|
| 3355 |
+
"grad_norm": 0.09215465933084488,
|
| 3356 |
+
"learning_rate": 2.2298968648248653e-05,
|
| 3357 |
+
"loss": 0.8161,
|
| 3358 |
+
"step": 467
|
| 3359 |
+
},
|
| 3360 |
+
{
|
| 3361 |
+
"epoch": 0.013008943648758521,
|
| 3362 |
+
"grad_norm": 0.08623038232326508,
|
| 3363 |
+
"learning_rate": 2.0972665167707127e-05,
|
| 3364 |
+
"loss": 0.7643,
|
| 3365 |
+
"step": 468
|
| 3366 |
+
},
|
| 3367 |
+
{
|
| 3368 |
+
"epoch": 0.013036740536896895,
|
| 3369 |
+
"grad_norm": 0.08695585280656815,
|
| 3370 |
+
"learning_rate": 1.968660566631275e-05,
|
| 3371 |
+
"loss": 0.6102,
|
| 3372 |
+
"step": 469
|
| 3373 |
+
},
|
| 3374 |
+
{
|
| 3375 |
+
"epoch": 0.013064537425035267,
|
| 3376 |
+
"grad_norm": 0.08945546299219131,
|
| 3377 |
+
"learning_rate": 1.844084300893456e-05,
|
| 3378 |
+
"loss": 0.7888,
|
| 3379 |
+
"step": 470
|
| 3380 |
+
},
|
| 3381 |
+
{
|
| 3382 |
+
"epoch": 0.01309233431317364,
|
| 3383 |
+
"grad_norm": 0.09600325673818588,
|
| 3384 |
+
"learning_rate": 1.7235428403996167e-05,
|
| 3385 |
+
"loss": 0.8451,
|
| 3386 |
+
"step": 471
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 0.013120131201312012,
|
| 3390 |
+
"grad_norm": 0.08915964514017105,
|
| 3391 |
+
"learning_rate": 1.6070411401370334e-05,
|
| 3392 |
+
"loss": 0.6192,
|
| 3393 |
+
"step": 472
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 0.013147928089450386,
|
| 3397 |
+
"grad_norm": 0.1044343113899231,
|
| 3398 |
+
"learning_rate": 1.494583989034326e-05,
|
| 3399 |
+
"loss": 0.6746,
|
| 3400 |
+
"step": 473
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"epoch": 0.01317572497758876,
|
| 3404 |
+
"grad_norm": 0.0989852100610733,
|
| 3405 |
+
"learning_rate": 1.386176009764506e-05,
|
| 3406 |
+
"loss": 0.7388,
|
| 3407 |
+
"step": 474
|
| 3408 |
+
},
|
| 3409 |
+
{
|
| 3410 |
+
"epoch": 0.013203521865727132,
|
| 3411 |
+
"grad_norm": 0.11182911694049835,
|
| 3412 |
+
"learning_rate": 1.2818216585549825e-05,
|
| 3413 |
+
"loss": 0.9093,
|
| 3414 |
+
"step": 475
|
| 3415 |
+
},
|
| 3416 |
+
{
|
| 3417 |
+
"epoch": 0.013231318753865505,
|
| 3418 |
+
"grad_norm": 0.10432388633489609,
|
| 3419 |
+
"learning_rate": 1.1815252250044316e-05,
|
| 3420 |
+
"loss": 0.6842,
|
| 3421 |
+
"step": 476
|
| 3422 |
+
},
|
| 3423 |
+
{
|
| 3424 |
+
"epoch": 0.013259115642003877,
|
| 3425 |
+
"grad_norm": 0.11141140758991241,
|
| 3426 |
+
"learning_rate": 1.0852908319063826e-05,
|
| 3427 |
+
"loss": 0.6488,
|
| 3428 |
+
"step": 477
|
| 3429 |
+
},
|
| 3430 |
+
{
|
| 3431 |
+
"epoch": 0.013286912530142251,
|
| 3432 |
+
"grad_norm": 0.10428661853075027,
|
| 3433 |
+
"learning_rate": 9.931224350798185e-06,
|
| 3434 |
+
"loss": 0.7708,
|
| 3435 |
+
"step": 478
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 0.013314709418280623,
|
| 3439 |
+
"grad_norm": 0.10303416848182678,
|
| 3440 |
+
"learning_rate": 9.0502382320653e-06,
|
| 3441 |
+
"loss": 0.6728,
|
| 3442 |
+
"step": 479
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 0.013342506306418997,
|
| 3446 |
+
"grad_norm": 0.10627992451190948,
|
| 3447 |
+
"learning_rate": 8.209986176753947e-06,
|
| 3448 |
+
"loss": 0.6125,
|
| 3449 |
+
"step": 480
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"epoch": 0.013370303194557369,
|
| 3453 |
+
"grad_norm": 0.09629444032907486,
|
| 3454 |
+
"learning_rate": 7.4105027243349665e-06,
|
| 3455 |
+
"loss": 0.6386,
|
| 3456 |
+
"step": 481
|
| 3457 |
+
},
|
| 3458 |
+
{
|
| 3459 |
+
"epoch": 0.013398100082695742,
|
| 3460 |
+
"grad_norm": 0.09442020207643509,
|
| 3461 |
+
"learning_rate": 6.65182073844195e-06,
|
| 3462 |
+
"loss": 0.5644,
|
| 3463 |
+
"step": 482
|
| 3464 |
+
},
|
| 3465 |
+
{
|
| 3466 |
+
"epoch": 0.013425896970834116,
|
| 3467 |
+
"grad_norm": 0.0994250699877739,
|
| 3468 |
+
"learning_rate": 5.933971405519656e-06,
|
| 3469 |
+
"loss": 0.6243,
|
| 3470 |
+
"step": 483
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 0.013453693858972488,
|
| 3474 |
+
"grad_norm": 0.1111208125948906,
|
| 3475 |
+
"learning_rate": 5.256984233542595e-06,
|
| 3476 |
+
"loss": 0.7222,
|
| 3477 |
+
"step": 484
|
| 3478 |
+
},
|
| 3479 |
+
{
|
| 3480 |
+
"epoch": 0.013481490747110862,
|
| 3481 |
+
"grad_norm": 0.11001411825418472,
|
| 3482 |
+
"learning_rate": 4.6208870508017695e-06,
|
| 3483 |
+
"loss": 0.693,
|
| 3484 |
+
"step": 485
|
| 3485 |
+
},
|
| 3486 |
+
{
|
| 3487 |
+
"epoch": 0.013509287635249233,
|
| 3488 |
+
"grad_norm": 0.0949028953909874,
|
| 3489 |
+
"learning_rate": 4.025706004760932e-06,
|
| 3490 |
+
"loss": 0.6378,
|
| 3491 |
+
"step": 486
|
| 3492 |
+
},
|
| 3493 |
+
{
|
| 3494 |
+
"epoch": 0.013537084523387607,
|
| 3495 |
+
"grad_norm": 0.10430373251438141,
|
| 3496 |
+
"learning_rate": 3.471465560981768e-06,
|
| 3497 |
+
"loss": 0.5924,
|
| 3498 |
+
"step": 487
|
| 3499 |
+
},
|
| 3500 |
+
{
|
| 3501 |
+
"epoch": 0.013564881411525979,
|
| 3502 |
+
"grad_norm": 0.09569145739078522,
|
| 3503 |
+
"learning_rate": 2.958188502118153e-06,
|
| 3504 |
+
"loss": 0.5534,
|
| 3505 |
+
"step": 488
|
| 3506 |
+
},
|
| 3507 |
+
{
|
| 3508 |
+
"epoch": 0.013592678299664353,
|
| 3509 |
+
"grad_norm": 0.09768345206975937,
|
| 3510 |
+
"learning_rate": 2.4858959269794535e-06,
|
| 3511 |
+
"loss": 0.4815,
|
| 3512 |
+
"step": 489
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 0.013620475187802725,
|
| 3516 |
+
"grad_norm": 0.09561553597450256,
|
| 3517 |
+
"learning_rate": 2.054607249663665e-06,
|
| 3518 |
+
"loss": 0.5637,
|
| 3519 |
+
"step": 490
|
| 3520 |
+
},
|
| 3521 |
+
{
|
| 3522 |
+
"epoch": 0.013648272075941098,
|
| 3523 |
+
"grad_norm": 0.11304374039173126,
|
| 3524 |
+
"learning_rate": 1.6643401987591622e-06,
|
| 3525 |
+
"loss": 0.5943,
|
| 3526 |
+
"step": 491
|
| 3527 |
+
},
|
| 3528 |
+
{
|
| 3529 |
+
"epoch": 0.013676068964079472,
|
| 3530 |
+
"grad_norm": 0.11450091749429703,
|
| 3531 |
+
"learning_rate": 1.3151108166156167e-06,
|
| 3532 |
+
"loss": 0.583,
|
| 3533 |
+
"step": 492
|
| 3534 |
+
},
|
| 3535 |
+
{
|
| 3536 |
+
"epoch": 0.013703865852217844,
|
| 3537 |
+
"grad_norm": 0.11334867030382156,
|
| 3538 |
+
"learning_rate": 1.0069334586854107e-06,
|
| 3539 |
+
"loss": 0.6076,
|
| 3540 |
+
"step": 493
|
| 3541 |
+
},
|
| 3542 |
+
{
|
| 3543 |
+
"epoch": 0.013731662740356218,
|
| 3544 |
+
"grad_norm": 0.1081780344247818,
|
| 3545 |
+
"learning_rate": 7.398207929323331e-07,
|
| 3546 |
+
"loss": 0.5119,
|
| 3547 |
+
"step": 494
|
| 3548 |
+
},
|
| 3549 |
+
{
|
| 3550 |
+
"epoch": 0.01375945962849459,
|
| 3551 |
+
"grad_norm": 0.11384209990501404,
|
| 3552 |
+
"learning_rate": 5.137837993121064e-07,
|
| 3553 |
+
"loss": 0.48,
|
| 3554 |
+
"step": 495
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 0.013787256516632963,
|
| 3558 |
+
"grad_norm": 0.11100795120000839,
|
| 3559 |
+
"learning_rate": 3.2883176932019256e-07,
|
| 3560 |
+
"loss": 0.53,
|
| 3561 |
+
"step": 496
|
| 3562 |
+
},
|
| 3563 |
+
{
|
| 3564 |
+
"epoch": 0.013815053404771335,
|
| 3565 |
+
"grad_norm": 0.14853325486183167,
|
| 3566 |
+
"learning_rate": 1.8497230560998722e-07,
|
| 3567 |
+
"loss": 0.5982,
|
| 3568 |
+
"step": 497
|
| 3569 |
+
},
|
| 3570 |
+
{
|
| 3571 |
+
"epoch": 0.013842850292909709,
|
| 3572 |
+
"grad_norm": 0.1417522132396698,
|
| 3573 |
+
"learning_rate": 8.221132168073631e-08,
|
| 3574 |
+
"loss": 0.5259,
|
| 3575 |
+
"step": 498
|
| 3576 |
+
},
|
| 3577 |
+
{
|
| 3578 |
+
"epoch": 0.013870647181048081,
|
| 3579 |
+
"grad_norm": 0.15358008444309235,
|
| 3580 |
+
"learning_rate": 2.0553041633952775e-08,
|
| 3581 |
+
"loss": 0.5328,
|
| 3582 |
+
"step": 499
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"epoch": 0.013898444069186455,
|
| 3586 |
+
"grad_norm": 0.21747428178787231,
|
| 3587 |
+
"learning_rate": 0.0,
|
| 3588 |
+
"loss": 0.5217,
|
| 3589 |
+
"step": 500
|
| 3590 |
+
},
|
| 3591 |
+
{
|
| 3592 |
+
"epoch": 0.013898444069186455,
|
| 3593 |
+
"eval_loss": 0.6910951733589172,
|
| 3594 |
+
"eval_runtime": 50.0791,
|
| 3595 |
+
"eval_samples_per_second": 11.522,
|
| 3596 |
+
"eval_steps_per_second": 5.771,
|
| 3597 |
+
"step": 500
|
| 3598 |
}
|
| 3599 |
],
|
| 3600 |
"logging_steps": 1,
|
|
|
|
| 3618 |
"should_evaluate": false,
|
| 3619 |
"should_log": false,
|
| 3620 |
"should_save": true,
|
| 3621 |
+
"should_training_stop": true
|
| 3622 |
},
|
| 3623 |
"attributes": {}
|
| 3624 |
}
|
| 3625 |
},
|
| 3626 |
+
"total_flos": 3.1913641871671296e+16,
|
| 3627 |
"train_batch_size": 2,
|
| 3628 |
"trial_name": null,
|
| 3629 |
"trial_params": null
|