Training in progress, step 44000, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 140637408
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f75eb1ddf21330639482807ec7deab113b52fe85d48ba8b6cab43e522ec03402
|
| 3 |
size 140637408
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 281362507
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:54dfc6127784cb5e89f86fea0834628983f8bfa533048d62d90c709d9514e029
|
| 3 |
size 281362507
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:159d486e538e30c3a9c773cf073925c8d684c569c5eafafd9a9cd4e6bbd35ab3
|
| 3 |
size 14645
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:30e7158a27c5dc8697c881f9f9e65a4d32cd650004eaef05bb9ee9e4b8561c26
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -31241,6 +31241,365 @@
|
|
| 31241 |
"eval_samples_per_second": 32.648,
|
| 31242 |
"eval_steps_per_second": 0.522,
|
| 31243 |
"step": 43500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31244 |
}
|
| 31245 |
],
|
| 31246 |
"logging_steps": 10,
|
|
@@ -31260,7 +31619,7 @@
|
|
| 31260 |
"attributes": {}
|
| 31261 |
}
|
| 31262 |
},
|
| 31263 |
-
"total_flos": 7.
|
| 31264 |
"train_batch_size": 128,
|
| 31265 |
"trial_name": null,
|
| 31266 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.576716386609694,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 44000,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 31241 |
"eval_samples_per_second": 32.648,
|
| 31242 |
"eval_steps_per_second": 0.522,
|
| 31243 |
"step": 43500
|
| 31244 |
+
},
|
| 31245 |
+
{
|
| 31246 |
+
"epoch": 0.5702938632133588,
|
| 31247 |
+
"grad_norm": 0.6366761922836304,
|
| 31248 |
+
"learning_rate": 0.00011845412627966858,
|
| 31249 |
+
"loss": 3.2273,
|
| 31250 |
+
"step": 43510
|
| 31251 |
+
},
|
| 31252 |
+
{
|
| 31253 |
+
"epoch": 0.5704249351194065,
|
| 31254 |
+
"grad_norm": 0.5727043747901917,
|
| 31255 |
+
"learning_rate": 0.00011839334582299703,
|
| 31256 |
+
"loss": 3.263,
|
| 31257 |
+
"step": 43520
|
| 31258 |
+
},
|
| 31259 |
+
{
|
| 31260 |
+
"epoch": 0.5705560070254542,
|
| 31261 |
+
"grad_norm": 0.6051214337348938,
|
| 31262 |
+
"learning_rate": 0.00011833257079643528,
|
| 31263 |
+
"loss": 3.2239,
|
| 31264 |
+
"step": 43530
|
| 31265 |
+
},
|
| 31266 |
+
{
|
| 31267 |
+
"epoch": 0.5706870789315018,
|
| 31268 |
+
"grad_norm": 0.6732583045959473,
|
| 31269 |
+
"learning_rate": 0.00011827180121042458,
|
| 31270 |
+
"loss": 3.2369,
|
| 31271 |
+
"step": 43540
|
| 31272 |
+
},
|
| 31273 |
+
{
|
| 31274 |
+
"epoch": 0.5708181508375495,
|
| 31275 |
+
"grad_norm": 0.5960179567337036,
|
| 31276 |
+
"learning_rate": 0.00011821103707540542,
|
| 31277 |
+
"loss": 3.2433,
|
| 31278 |
+
"step": 43550
|
| 31279 |
+
},
|
| 31280 |
+
{
|
| 31281 |
+
"epoch": 0.5709492227435972,
|
| 31282 |
+
"grad_norm": 0.6697182059288025,
|
| 31283 |
+
"learning_rate": 0.0001181502784018171,
|
| 31284 |
+
"loss": 3.1666,
|
| 31285 |
+
"step": 43560
|
| 31286 |
+
},
|
| 31287 |
+
{
|
| 31288 |
+
"epoch": 0.5710802946496448,
|
| 31289 |
+
"grad_norm": 0.658229410648346,
|
| 31290 |
+
"learning_rate": 0.00011808952520009821,
|
| 31291 |
+
"loss": 3.2425,
|
| 31292 |
+
"step": 43570
|
| 31293 |
+
},
|
| 31294 |
+
{
|
| 31295 |
+
"epoch": 0.5712113665556925,
|
| 31296 |
+
"grad_norm": 0.899276852607727,
|
| 31297 |
+
"learning_rate": 0.00011802877748068629,
|
| 31298 |
+
"loss": 3.2724,
|
| 31299 |
+
"step": 43580
|
| 31300 |
+
},
|
| 31301 |
+
{
|
| 31302 |
+
"epoch": 0.5713424384617402,
|
| 31303 |
+
"grad_norm": 0.6311741471290588,
|
| 31304 |
+
"learning_rate": 0.00011796803525401808,
|
| 31305 |
+
"loss": 3.2569,
|
| 31306 |
+
"step": 43590
|
| 31307 |
+
},
|
| 31308 |
+
{
|
| 31309 |
+
"epoch": 0.5714735103677878,
|
| 31310 |
+
"grad_norm": 0.6979262828826904,
|
| 31311 |
+
"learning_rate": 0.00011790729853052907,
|
| 31312 |
+
"loss": 3.2655,
|
| 31313 |
+
"step": 43600
|
| 31314 |
+
},
|
| 31315 |
+
{
|
| 31316 |
+
"epoch": 0.5716045822738354,
|
| 31317 |
+
"grad_norm": 0.6784321665763855,
|
| 31318 |
+
"learning_rate": 0.00011784656732065413,
|
| 31319 |
+
"loss": 3.2323,
|
| 31320 |
+
"step": 43610
|
| 31321 |
+
},
|
| 31322 |
+
{
|
| 31323 |
+
"epoch": 0.571735654179883,
|
| 31324 |
+
"grad_norm": 0.5697076320648193,
|
| 31325 |
+
"learning_rate": 0.00011778584163482699,
|
| 31326 |
+
"loss": 3.2807,
|
| 31327 |
+
"step": 43620
|
| 31328 |
+
},
|
| 31329 |
+
{
|
| 31330 |
+
"epoch": 0.5718667260859307,
|
| 31331 |
+
"grad_norm": 0.6218394637107849,
|
| 31332 |
+
"learning_rate": 0.00011772512148348052,
|
| 31333 |
+
"loss": 3.2307,
|
| 31334 |
+
"step": 43630
|
| 31335 |
+
},
|
| 31336 |
+
{
|
| 31337 |
+
"epoch": 0.5719977979919784,
|
| 31338 |
+
"grad_norm": 0.6381251215934753,
|
| 31339 |
+
"learning_rate": 0.00011766440687704656,
|
| 31340 |
+
"loss": 3.263,
|
| 31341 |
+
"step": 43640
|
| 31342 |
+
},
|
| 31343 |
+
{
|
| 31344 |
+
"epoch": 0.572128869898026,
|
| 31345 |
+
"grad_norm": 0.5995039343833923,
|
| 31346 |
+
"learning_rate": 0.00011760369782595607,
|
| 31347 |
+
"loss": 3.2647,
|
| 31348 |
+
"step": 43650
|
| 31349 |
+
},
|
| 31350 |
+
{
|
| 31351 |
+
"epoch": 0.5722599418040737,
|
| 31352 |
+
"grad_norm": 0.6439694166183472,
|
| 31353 |
+
"learning_rate": 0.0001175429943406391,
|
| 31354 |
+
"loss": 3.2332,
|
| 31355 |
+
"step": 43660
|
| 31356 |
+
},
|
| 31357 |
+
{
|
| 31358 |
+
"epoch": 0.5723910137101214,
|
| 31359 |
+
"grad_norm": 0.5705400109291077,
|
| 31360 |
+
"learning_rate": 0.00011748229643152455,
|
| 31361 |
+
"loss": 3.2642,
|
| 31362 |
+
"step": 43670
|
| 31363 |
+
},
|
| 31364 |
+
{
|
| 31365 |
+
"epoch": 0.572522085616169,
|
| 31366 |
+
"grad_norm": 0.618656575679779,
|
| 31367 |
+
"learning_rate": 0.00011742160410904057,
|
| 31368 |
+
"loss": 3.2545,
|
| 31369 |
+
"step": 43680
|
| 31370 |
+
},
|
| 31371 |
+
{
|
| 31372 |
+
"epoch": 0.5726531575222167,
|
| 31373 |
+
"grad_norm": 0.6085860729217529,
|
| 31374 |
+
"learning_rate": 0.00011736091738361427,
|
| 31375 |
+
"loss": 3.2777,
|
| 31376 |
+
"step": 43690
|
| 31377 |
+
},
|
| 31378 |
+
{
|
| 31379 |
+
"epoch": 0.5727842294282643,
|
| 31380 |
+
"grad_norm": 0.6229195594787598,
|
| 31381 |
+
"learning_rate": 0.00011730023626567177,
|
| 31382 |
+
"loss": 3.2389,
|
| 31383 |
+
"step": 43700
|
| 31384 |
+
},
|
| 31385 |
+
{
|
| 31386 |
+
"epoch": 0.572915301334312,
|
| 31387 |
+
"grad_norm": 0.6317970752716064,
|
| 31388 |
+
"learning_rate": 0.00011723956076563825,
|
| 31389 |
+
"loss": 3.2701,
|
| 31390 |
+
"step": 43710
|
| 31391 |
+
},
|
| 31392 |
+
{
|
| 31393 |
+
"epoch": 0.5730463732403597,
|
| 31394 |
+
"grad_norm": 0.5994996428489685,
|
| 31395 |
+
"learning_rate": 0.00011717889089393792,
|
| 31396 |
+
"loss": 3.2596,
|
| 31397 |
+
"step": 43720
|
| 31398 |
+
},
|
| 31399 |
+
{
|
| 31400 |
+
"epoch": 0.5731774451464073,
|
| 31401 |
+
"grad_norm": 0.6655870676040649,
|
| 31402 |
+
"learning_rate": 0.0001171182266609941,
|
| 31403 |
+
"loss": 3.2381,
|
| 31404 |
+
"step": 43730
|
| 31405 |
+
},
|
| 31406 |
+
{
|
| 31407 |
+
"epoch": 0.573308517052455,
|
| 31408 |
+
"grad_norm": 1.0828248262405396,
|
| 31409 |
+
"learning_rate": 0.00011705756807722896,
|
| 31410 |
+
"loss": 3.2781,
|
| 31411 |
+
"step": 43740
|
| 31412 |
+
},
|
| 31413 |
+
{
|
| 31414 |
+
"epoch": 0.5734395889585027,
|
| 31415 |
+
"grad_norm": 0.6480745673179626,
|
| 31416 |
+
"learning_rate": 0.00011699691515306386,
|
| 31417 |
+
"loss": 3.2219,
|
| 31418 |
+
"step": 43750
|
| 31419 |
+
},
|
| 31420 |
+
{
|
| 31421 |
+
"epoch": 0.5735706608645503,
|
| 31422 |
+
"grad_norm": 0.721631646156311,
|
| 31423 |
+
"learning_rate": 0.00011693626789891913,
|
| 31424 |
+
"loss": 3.267,
|
| 31425 |
+
"step": 43760
|
| 31426 |
+
},
|
| 31427 |
+
{
|
| 31428 |
+
"epoch": 0.573701732770598,
|
| 31429 |
+
"grad_norm": 0.8623172640800476,
|
| 31430 |
+
"learning_rate": 0.00011687562632521425,
|
| 31431 |
+
"loss": 3.2556,
|
| 31432 |
+
"step": 43770
|
| 31433 |
+
},
|
| 31434 |
+
{
|
| 31435 |
+
"epoch": 0.5738328046766457,
|
| 31436 |
+
"grad_norm": 0.6004065871238708,
|
| 31437 |
+
"learning_rate": 0.00011681499044236741,
|
| 31438 |
+
"loss": 3.2519,
|
| 31439 |
+
"step": 43780
|
| 31440 |
+
},
|
| 31441 |
+
{
|
| 31442 |
+
"epoch": 0.5739638765826932,
|
| 31443 |
+
"grad_norm": 0.5865307450294495,
|
| 31444 |
+
"learning_rate": 0.00011675436026079617,
|
| 31445 |
+
"loss": 3.2901,
|
| 31446 |
+
"step": 43790
|
| 31447 |
+
},
|
| 31448 |
+
{
|
| 31449 |
+
"epoch": 0.5740949484887409,
|
| 31450 |
+
"grad_norm": 0.6044079661369324,
|
| 31451 |
+
"learning_rate": 0.00011669373579091687,
|
| 31452 |
+
"loss": 3.2271,
|
| 31453 |
+
"step": 43800
|
| 31454 |
+
},
|
| 31455 |
+
{
|
| 31456 |
+
"epoch": 0.5742260203947885,
|
| 31457 |
+
"grad_norm": 0.5992538332939148,
|
| 31458 |
+
"learning_rate": 0.00011663311704314507,
|
| 31459 |
+
"loss": 3.2476,
|
| 31460 |
+
"step": 43810
|
| 31461 |
+
},
|
| 31462 |
+
{
|
| 31463 |
+
"epoch": 0.5743570923008362,
|
| 31464 |
+
"grad_norm": 0.6272064447402954,
|
| 31465 |
+
"learning_rate": 0.00011657250402789511,
|
| 31466 |
+
"loss": 3.2468,
|
| 31467 |
+
"step": 43820
|
| 31468 |
+
},
|
| 31469 |
+
{
|
| 31470 |
+
"epoch": 0.5744881642068839,
|
| 31471 |
+
"grad_norm": 0.6651163101196289,
|
| 31472 |
+
"learning_rate": 0.00011651189675558052,
|
| 31473 |
+
"loss": 3.2449,
|
| 31474 |
+
"step": 43830
|
| 31475 |
+
},
|
| 31476 |
+
{
|
| 31477 |
+
"epoch": 0.5746192361129315,
|
| 31478 |
+
"grad_norm": 0.674713134765625,
|
| 31479 |
+
"learning_rate": 0.00011645129523661379,
|
| 31480 |
+
"loss": 3.1796,
|
| 31481 |
+
"step": 43840
|
| 31482 |
+
},
|
| 31483 |
+
{
|
| 31484 |
+
"epoch": 0.5747503080189792,
|
| 31485 |
+
"grad_norm": 0.6734214425086975,
|
| 31486 |
+
"learning_rate": 0.00011639069948140653,
|
| 31487 |
+
"loss": 3.2686,
|
| 31488 |
+
"step": 43850
|
| 31489 |
+
},
|
| 31490 |
+
{
|
| 31491 |
+
"epoch": 0.5748813799250269,
|
| 31492 |
+
"grad_norm": 0.615215539932251,
|
| 31493 |
+
"learning_rate": 0.00011633010950036908,
|
| 31494 |
+
"loss": 3.2385,
|
| 31495 |
+
"step": 43860
|
| 31496 |
+
},
|
| 31497 |
+
{
|
| 31498 |
+
"epoch": 0.5750124518310745,
|
| 31499 |
+
"grad_norm": 0.6813642978668213,
|
| 31500 |
+
"learning_rate": 0.00011626952530391111,
|
| 31501 |
+
"loss": 3.2678,
|
| 31502 |
+
"step": 43870
|
| 31503 |
+
},
|
| 31504 |
+
{
|
| 31505 |
+
"epoch": 0.5751435237371222,
|
| 31506 |
+
"grad_norm": 0.6137657165527344,
|
| 31507 |
+
"learning_rate": 0.00011620894690244106,
|
| 31508 |
+
"loss": 3.2657,
|
| 31509 |
+
"step": 43880
|
| 31510 |
+
},
|
| 31511 |
+
{
|
| 31512 |
+
"epoch": 0.5752745956431699,
|
| 31513 |
+
"grad_norm": 0.5825940370559692,
|
| 31514 |
+
"learning_rate": 0.00011614837430636653,
|
| 31515 |
+
"loss": 3.1976,
|
| 31516 |
+
"step": 43890
|
| 31517 |
+
},
|
| 31518 |
+
{
|
| 31519 |
+
"epoch": 0.5754056675492175,
|
| 31520 |
+
"grad_norm": 0.8284161686897278,
|
| 31521 |
+
"learning_rate": 0.00011608780752609398,
|
| 31522 |
+
"loss": 3.2812,
|
| 31523 |
+
"step": 43900
|
| 31524 |
+
},
|
| 31525 |
+
{
|
| 31526 |
+
"epoch": 0.5755367394552652,
|
| 31527 |
+
"grad_norm": 0.7353925108909607,
|
| 31528 |
+
"learning_rate": 0.00011602724657202901,
|
| 31529 |
+
"loss": 3.2052,
|
| 31530 |
+
"step": 43910
|
| 31531 |
+
},
|
| 31532 |
+
{
|
| 31533 |
+
"epoch": 0.5756678113613128,
|
| 31534 |
+
"grad_norm": 0.6302970051765442,
|
| 31535 |
+
"learning_rate": 0.0001159666914545762,
|
| 31536 |
+
"loss": 3.2397,
|
| 31537 |
+
"step": 43920
|
| 31538 |
+
},
|
| 31539 |
+
{
|
| 31540 |
+
"epoch": 0.5757988832673605,
|
| 31541 |
+
"grad_norm": 0.6776291728019714,
|
| 31542 |
+
"learning_rate": 0.00011590614218413895,
|
| 31543 |
+
"loss": 3.2815,
|
| 31544 |
+
"step": 43930
|
| 31545 |
+
},
|
| 31546 |
+
{
|
| 31547 |
+
"epoch": 0.5759299551734082,
|
| 31548 |
+
"grad_norm": 0.6846403479576111,
|
| 31549 |
+
"learning_rate": 0.00011584559877111988,
|
| 31550 |
+
"loss": 3.1895,
|
| 31551 |
+
"step": 43940
|
| 31552 |
+
},
|
| 31553 |
+
{
|
| 31554 |
+
"epoch": 0.5760610270794558,
|
| 31555 |
+
"grad_norm": 0.5829125642776489,
|
| 31556 |
+
"learning_rate": 0.00011578506122592055,
|
| 31557 |
+
"loss": 3.2676,
|
| 31558 |
+
"step": 43950
|
| 31559 |
+
},
|
| 31560 |
+
{
|
| 31561 |
+
"epoch": 0.5761920989855035,
|
| 31562 |
+
"grad_norm": 0.6091601252555847,
|
| 31563 |
+
"learning_rate": 0.00011572452955894141,
|
| 31564 |
+
"loss": 3.1885,
|
| 31565 |
+
"step": 43960
|
| 31566 |
+
},
|
| 31567 |
+
{
|
| 31568 |
+
"epoch": 0.576323170891551,
|
| 31569 |
+
"grad_norm": 0.6871894001960754,
|
| 31570 |
+
"learning_rate": 0.000115664003780582,
|
| 31571 |
+
"loss": 3.2926,
|
| 31572 |
+
"step": 43970
|
| 31573 |
+
},
|
| 31574 |
+
{
|
| 31575 |
+
"epoch": 0.5764542427975987,
|
| 31576 |
+
"grad_norm": 0.6193121075630188,
|
| 31577 |
+
"learning_rate": 0.00011560348390124076,
|
| 31578 |
+
"loss": 3.3302,
|
| 31579 |
+
"step": 43980
|
| 31580 |
+
},
|
| 31581 |
+
{
|
| 31582 |
+
"epoch": 0.5765853147036464,
|
| 31583 |
+
"grad_norm": 0.6359044313430786,
|
| 31584 |
+
"learning_rate": 0.00011554296993131526,
|
| 31585 |
+
"loss": 3.2165,
|
| 31586 |
+
"step": 43990
|
| 31587 |
+
},
|
| 31588 |
+
{
|
| 31589 |
+
"epoch": 0.576716386609694,
|
| 31590 |
+
"grad_norm": 0.6618209481239319,
|
| 31591 |
+
"learning_rate": 0.0001154824618812019,
|
| 31592 |
+
"loss": 3.2372,
|
| 31593 |
+
"step": 44000
|
| 31594 |
+
},
|
| 31595 |
+
{
|
| 31596 |
+
"epoch": 0.576716386609694,
|
| 31597 |
+
"eval_accuracy": 0.39772482893450634,
|
| 31598 |
+
"eval_loss": 3.241494655609131,
|
| 31599 |
+
"eval_runtime": 60.0316,
|
| 31600 |
+
"eval_samples_per_second": 33.316,
|
| 31601 |
+
"eval_steps_per_second": 0.533,
|
| 31602 |
+
"step": 44000
|
| 31603 |
}
|
| 31604 |
],
|
| 31605 |
"logging_steps": 10,
|
|
|
|
| 31619 |
"attributes": {}
|
| 31620 |
}
|
| 31621 |
},
|
| 31622 |
+
"total_flos": 7.91287192092672e+17,
|
| 31623 |
"train_batch_size": 128,
|
| 31624 |
"trial_name": null,
|
| 31625 |
"trial_params": null
|