Upload folder using huggingface_hub
Browse files- model.safetensors +1 -1
- optimizer.pt +1 -1
- scheduler.pt +1 -1
- trainer_state.json +703 -3
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2815117504
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5ce622b98726b39ea3525b3a60f165f4a2e245c6f303bb285dc9a30d63fae39d
|
3 |
size 2815117504
|
optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2486357050
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e7e8c9cebcc9825d94ce1a61b80717541b152fc6070b98f7b5289160d3393d52
|
3 |
size 2486357050
|
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:a8238cb53c30f0afbee1f016fc637546ca2ff4b464b4241ee444a3083fdd4a52
|
3 |
size 1064
|
trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 1.
|
5 |
"eval_steps": 500,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -4207,6 +4207,706 @@
|
|
4207 |
"learning_rate": 8.374629289717937e-06,
|
4208 |
"loss": 1.7362,
|
4209 |
"step": 600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4210 |
}
|
4211 |
],
|
4212 |
"logging_steps": 1,
|
@@ -4226,7 +4926,7 @@
|
|
4226 |
"attributes": {}
|
4227 |
}
|
4228 |
},
|
4229 |
-
"total_flos":
|
4230 |
"train_batch_size": 10,
|
4231 |
"trial_name": null,
|
4232 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.340354236476783,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 700,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
4207 |
"learning_rate": 8.374629289717937e-06,
|
4208 |
"loss": 1.7362,
|
4209 |
"step": 600
|
4210 |
+
},
|
4211 |
+
{
|
4212 |
+
"epoch": 1.1507898516036381,
|
4213 |
+
"grad_norm": 1.765625,
|
4214 |
+
"learning_rate": 8.343357820842829e-06,
|
4215 |
+
"loss": 1.6909,
|
4216 |
+
"step": 601
|
4217 |
+
},
|
4218 |
+
{
|
4219 |
+
"epoch": 1.1527046433700334,
|
4220 |
+
"grad_norm": 1.671875,
|
4221 |
+
"learning_rate": 8.31210300068534e-06,
|
4222 |
+
"loss": 1.641,
|
4223 |
+
"step": 602
|
4224 |
+
},
|
4225 |
+
{
|
4226 |
+
"epoch": 1.154619435136429,
|
4227 |
+
"grad_norm": 1.78125,
|
4228 |
+
"learning_rate": 8.280865143346301e-06,
|
4229 |
+
"loss": 1.7413,
|
4230 |
+
"step": 603
|
4231 |
+
},
|
4232 |
+
{
|
4233 |
+
"epoch": 1.1565342269028243,
|
4234 |
+
"grad_norm": 1.7265625,
|
4235 |
+
"learning_rate": 8.249644562756056e-06,
|
4236 |
+
"loss": 1.6597,
|
4237 |
+
"step": 604
|
4238 |
+
},
|
4239 |
+
{
|
4240 |
+
"epoch": 1.1584490186692198,
|
4241 |
+
"grad_norm": 1.7890625,
|
4242 |
+
"learning_rate": 8.218441572671343e-06,
|
4243 |
+
"loss": 1.6769,
|
4244 |
+
"step": 605
|
4245 |
+
},
|
4246 |
+
{
|
4247 |
+
"epoch": 1.160363810435615,
|
4248 |
+
"grad_norm": 1.8671875,
|
4249 |
+
"learning_rate": 8.187256486672106e-06,
|
4250 |
+
"loss": 1.6976,
|
4251 |
+
"step": 606
|
4252 |
+
},
|
4253 |
+
{
|
4254 |
+
"epoch": 1.1622786022020106,
|
4255 |
+
"grad_norm": 1.640625,
|
4256 |
+
"learning_rate": 8.156089618158366e-06,
|
4257 |
+
"loss": 1.6065,
|
4258 |
+
"step": 607
|
4259 |
+
},
|
4260 |
+
{
|
4261 |
+
"epoch": 1.164193393968406,
|
4262 |
+
"grad_norm": 1.75,
|
4263 |
+
"learning_rate": 8.12494128034706e-06,
|
4264 |
+
"loss": 1.7632,
|
4265 |
+
"step": 608
|
4266 |
+
},
|
4267 |
+
{
|
4268 |
+
"epoch": 1.1661081857348012,
|
4269 |
+
"grad_norm": 1.7421875,
|
4270 |
+
"learning_rate": 8.093811786268905e-06,
|
4271 |
+
"loss": 1.6849,
|
4272 |
+
"step": 609
|
4273 |
+
},
|
4274 |
+
{
|
4275 |
+
"epoch": 1.1680229775011968,
|
4276 |
+
"grad_norm": 1.796875,
|
4277 |
+
"learning_rate": 8.06270144876523e-06,
|
4278 |
+
"loss": 1.7192,
|
4279 |
+
"step": 610
|
4280 |
+
},
|
4281 |
+
{
|
4282 |
+
"epoch": 1.169937769267592,
|
4283 |
+
"grad_norm": 1.6796875,
|
4284 |
+
"learning_rate": 8.03161058048487e-06,
|
4285 |
+
"loss": 1.6864,
|
4286 |
+
"step": 611
|
4287 |
+
},
|
4288 |
+
{
|
4289 |
+
"epoch": 1.1718525610339876,
|
4290 |
+
"grad_norm": 1.703125,
|
4291 |
+
"learning_rate": 8.000539493880972e-06,
|
4292 |
+
"loss": 1.704,
|
4293 |
+
"step": 612
|
4294 |
+
},
|
4295 |
+
{
|
4296 |
+
"epoch": 1.173767352800383,
|
4297 |
+
"grad_norm": 1.765625,
|
4298 |
+
"learning_rate": 7.969488501207909e-06,
|
4299 |
+
"loss": 1.6947,
|
4300 |
+
"step": 613
|
4301 |
+
},
|
4302 |
+
{
|
4303 |
+
"epoch": 1.1756821445667784,
|
4304 |
+
"grad_norm": 1.703125,
|
4305 |
+
"learning_rate": 7.938457914518098e-06,
|
4306 |
+
"loss": 1.7126,
|
4307 |
+
"step": 614
|
4308 |
+
},
|
4309 |
+
{
|
4310 |
+
"epoch": 1.1775969363331737,
|
4311 |
+
"grad_norm": 1.765625,
|
4312 |
+
"learning_rate": 7.907448045658899e-06,
|
4313 |
+
"loss": 1.7502,
|
4314 |
+
"step": 615
|
4315 |
+
},
|
4316 |
+
{
|
4317 |
+
"epoch": 1.1795117280995693,
|
4318 |
+
"grad_norm": 1.7578125,
|
4319 |
+
"learning_rate": 7.876459206269446e-06,
|
4320 |
+
"loss": 1.7348,
|
4321 |
+
"step": 616
|
4322 |
+
},
|
4323 |
+
{
|
4324 |
+
"epoch": 1.1814265198659646,
|
4325 |
+
"grad_norm": 1.7109375,
|
4326 |
+
"learning_rate": 7.845491707777551e-06,
|
4327 |
+
"loss": 1.6578,
|
4328 |
+
"step": 617
|
4329 |
+
},
|
4330 |
+
{
|
4331 |
+
"epoch": 1.18334131163236,
|
4332 |
+
"grad_norm": 1.734375,
|
4333 |
+
"learning_rate": 7.814545861396543e-06,
|
4334 |
+
"loss": 1.7479,
|
4335 |
+
"step": 618
|
4336 |
+
},
|
4337 |
+
{
|
4338 |
+
"epoch": 1.1852561033987554,
|
4339 |
+
"grad_norm": 1.7421875,
|
4340 |
+
"learning_rate": 7.783621978122167e-06,
|
4341 |
+
"loss": 1.7027,
|
4342 |
+
"step": 619
|
4343 |
+
},
|
4344 |
+
{
|
4345 |
+
"epoch": 1.1871708951651507,
|
4346 |
+
"grad_norm": 1.7265625,
|
4347 |
+
"learning_rate": 7.752720368729436e-06,
|
4348 |
+
"loss": 1.6828,
|
4349 |
+
"step": 620
|
4350 |
+
},
|
4351 |
+
{
|
4352 |
+
"epoch": 1.1890856869315463,
|
4353 |
+
"grad_norm": 1.703125,
|
4354 |
+
"learning_rate": 7.721841343769518e-06,
|
4355 |
+
"loss": 1.689,
|
4356 |
+
"step": 621
|
4357 |
+
},
|
4358 |
+
{
|
4359 |
+
"epoch": 1.1910004786979416,
|
4360 |
+
"grad_norm": 1.6796875,
|
4361 |
+
"learning_rate": 7.69098521356662e-06,
|
4362 |
+
"loss": 1.6498,
|
4363 |
+
"step": 622
|
4364 |
+
},
|
4365 |
+
{
|
4366 |
+
"epoch": 1.192915270464337,
|
4367 |
+
"grad_norm": 1.6796875,
|
4368 |
+
"learning_rate": 7.660152288214865e-06,
|
4369 |
+
"loss": 1.6855,
|
4370 |
+
"step": 623
|
4371 |
+
},
|
4372 |
+
{
|
4373 |
+
"epoch": 1.1948300622307324,
|
4374 |
+
"grad_norm": 1.7109375,
|
4375 |
+
"learning_rate": 7.629342877575169e-06,
|
4376 |
+
"loss": 1.7234,
|
4377 |
+
"step": 624
|
4378 |
+
},
|
4379 |
+
{
|
4380 |
+
"epoch": 1.196744853997128,
|
4381 |
+
"grad_norm": 1.7734375,
|
4382 |
+
"learning_rate": 7.598557291272133e-06,
|
4383 |
+
"loss": 1.7746,
|
4384 |
+
"step": 625
|
4385 |
+
},
|
4386 |
+
{
|
4387 |
+
"epoch": 1.1986596457635232,
|
4388 |
+
"grad_norm": 1.765625,
|
4389 |
+
"learning_rate": 7.567795838690941e-06,
|
4390 |
+
"loss": 1.7283,
|
4391 |
+
"step": 626
|
4392 |
+
},
|
4393 |
+
{
|
4394 |
+
"epoch": 1.2005744375299185,
|
4395 |
+
"grad_norm": 1.6953125,
|
4396 |
+
"learning_rate": 7.537058828974226e-06,
|
4397 |
+
"loss": 1.7204,
|
4398 |
+
"step": 627
|
4399 |
+
},
|
4400 |
+
{
|
4401 |
+
"epoch": 1.202489229296314,
|
4402 |
+
"grad_norm": 1.7109375,
|
4403 |
+
"learning_rate": 7.506346571018992e-06,
|
4404 |
+
"loss": 1.7417,
|
4405 |
+
"step": 628
|
4406 |
+
},
|
4407 |
+
{
|
4408 |
+
"epoch": 1.2044040210627094,
|
4409 |
+
"grad_norm": 1.7109375,
|
4410 |
+
"learning_rate": 7.475659373473481e-06,
|
4411 |
+
"loss": 1.6638,
|
4412 |
+
"step": 629
|
4413 |
+
},
|
4414 |
+
{
|
4415 |
+
"epoch": 1.206318812829105,
|
4416 |
+
"grad_norm": 1.6796875,
|
4417 |
+
"learning_rate": 7.444997544734105e-06,
|
4418 |
+
"loss": 1.7001,
|
4419 |
+
"step": 630
|
4420 |
+
},
|
4421 |
+
{
|
4422 |
+
"epoch": 1.2082336045955002,
|
4423 |
+
"grad_norm": 1.6953125,
|
4424 |
+
"learning_rate": 7.414361392942307e-06,
|
4425 |
+
"loss": 1.7401,
|
4426 |
+
"step": 631
|
4427 |
+
},
|
4428 |
+
{
|
4429 |
+
"epoch": 1.2101483963618955,
|
4430 |
+
"grad_norm": 1.765625,
|
4431 |
+
"learning_rate": 7.383751225981503e-06,
|
4432 |
+
"loss": 1.7466,
|
4433 |
+
"step": 632
|
4434 |
+
},
|
4435 |
+
{
|
4436 |
+
"epoch": 1.212063188128291,
|
4437 |
+
"grad_norm": 1.875,
|
4438 |
+
"learning_rate": 7.353167351473955e-06,
|
4439 |
+
"loss": 1.6681,
|
4440 |
+
"step": 633
|
4441 |
+
},
|
4442 |
+
{
|
4443 |
+
"epoch": 1.2139779798946864,
|
4444 |
+
"grad_norm": 1.71875,
|
4445 |
+
"learning_rate": 7.322610076777707e-06,
|
4446 |
+
"loss": 1.7014,
|
4447 |
+
"step": 634
|
4448 |
+
},
|
4449 |
+
{
|
4450 |
+
"epoch": 1.215892771661082,
|
4451 |
+
"grad_norm": 1.734375,
|
4452 |
+
"learning_rate": 7.292079708983475e-06,
|
4453 |
+
"loss": 1.7218,
|
4454 |
+
"step": 635
|
4455 |
+
},
|
4456 |
+
{
|
4457 |
+
"epoch": 1.2178075634274772,
|
4458 |
+
"grad_norm": 1.7578125,
|
4459 |
+
"learning_rate": 7.261576554911575e-06,
|
4460 |
+
"loss": 1.7206,
|
4461 |
+
"step": 636
|
4462 |
+
},
|
4463 |
+
{
|
4464 |
+
"epoch": 1.2197223551938727,
|
4465 |
+
"grad_norm": 1.7421875,
|
4466 |
+
"learning_rate": 7.2311009211088255e-06,
|
4467 |
+
"loss": 1.6895,
|
4468 |
+
"step": 637
|
4469 |
+
},
|
4470 |
+
{
|
4471 |
+
"epoch": 1.221637146960268,
|
4472 |
+
"grad_norm": 1.734375,
|
4473 |
+
"learning_rate": 7.20065311384549e-06,
|
4474 |
+
"loss": 1.744,
|
4475 |
+
"step": 638
|
4476 |
+
},
|
4477 |
+
{
|
4478 |
+
"epoch": 1.2235519387266636,
|
4479 |
+
"grad_norm": 1.703125,
|
4480 |
+
"learning_rate": 7.170233439112164e-06,
|
4481 |
+
"loss": 1.7217,
|
4482 |
+
"step": 639
|
4483 |
+
},
|
4484 |
+
{
|
4485 |
+
"epoch": 1.2254667304930589,
|
4486 |
+
"grad_norm": 1.6875,
|
4487 |
+
"learning_rate": 7.139842202616741e-06,
|
4488 |
+
"loss": 1.6799,
|
4489 |
+
"step": 640
|
4490 |
+
},
|
4491 |
+
{
|
4492 |
+
"epoch": 1.2273815222594542,
|
4493 |
+
"grad_norm": 1.71875,
|
4494 |
+
"learning_rate": 7.109479709781302e-06,
|
4495 |
+
"loss": 1.7117,
|
4496 |
+
"step": 641
|
4497 |
+
},
|
4498 |
+
{
|
4499 |
+
"epoch": 1.2292963140258497,
|
4500 |
+
"grad_norm": 1.734375,
|
4501 |
+
"learning_rate": 7.079146265739079e-06,
|
4502 |
+
"loss": 1.6948,
|
4503 |
+
"step": 642
|
4504 |
+
},
|
4505 |
+
{
|
4506 |
+
"epoch": 1.231211105792245,
|
4507 |
+
"grad_norm": 1.78125,
|
4508 |
+
"learning_rate": 7.048842175331356e-06,
|
4509 |
+
"loss": 1.7343,
|
4510 |
+
"step": 643
|
4511 |
+
},
|
4512 |
+
{
|
4513 |
+
"epoch": 1.2331258975586405,
|
4514 |
+
"grad_norm": 1.7578125,
|
4515 |
+
"learning_rate": 7.0185677431044404e-06,
|
4516 |
+
"loss": 1.7078,
|
4517 |
+
"step": 644
|
4518 |
+
},
|
4519 |
+
{
|
4520 |
+
"epoch": 1.2350406893250359,
|
4521 |
+
"grad_norm": 1.671875,
|
4522 |
+
"learning_rate": 6.988323273306569e-06,
|
4523 |
+
"loss": 1.7168,
|
4524 |
+
"step": 645
|
4525 |
+
},
|
4526 |
+
{
|
4527 |
+
"epoch": 1.2369554810914314,
|
4528 |
+
"grad_norm": 1.6875,
|
4529 |
+
"learning_rate": 6.958109069884879e-06,
|
4530 |
+
"loss": 1.6997,
|
4531 |
+
"step": 646
|
4532 |
+
},
|
4533 |
+
{
|
4534 |
+
"epoch": 1.2388702728578267,
|
4535 |
+
"grad_norm": 1.6640625,
|
4536 |
+
"learning_rate": 6.9279254364823265e-06,
|
4537 |
+
"loss": 1.6204,
|
4538 |
+
"step": 647
|
4539 |
+
},
|
4540 |
+
{
|
4541 |
+
"epoch": 1.2407850646242222,
|
4542 |
+
"grad_norm": 1.7265625,
|
4543 |
+
"learning_rate": 6.897772676434663e-06,
|
4544 |
+
"loss": 1.727,
|
4545 |
+
"step": 648
|
4546 |
+
},
|
4547 |
+
{
|
4548 |
+
"epoch": 1.2426998563906175,
|
4549 |
+
"grad_norm": 1.71875,
|
4550 |
+
"learning_rate": 6.867651092767366e-06,
|
4551 |
+
"loss": 1.7278,
|
4552 |
+
"step": 649
|
4553 |
+
},
|
4554 |
+
{
|
4555 |
+
"epoch": 1.2446146481570128,
|
4556 |
+
"grad_norm": 1.703125,
|
4557 |
+
"learning_rate": 6.837560988192593e-06,
|
4558 |
+
"loss": 1.7087,
|
4559 |
+
"step": 650
|
4560 |
+
},
|
4561 |
+
{
|
4562 |
+
"epoch": 1.2465294399234084,
|
4563 |
+
"grad_norm": 1.65625,
|
4564 |
+
"learning_rate": 6.807502665106164e-06,
|
4565 |
+
"loss": 1.6614,
|
4566 |
+
"step": 651
|
4567 |
+
},
|
4568 |
+
{
|
4569 |
+
"epoch": 1.2484442316898037,
|
4570 |
+
"grad_norm": 1.7578125,
|
4571 |
+
"learning_rate": 6.777476425584486e-06,
|
4572 |
+
"loss": 1.7264,
|
4573 |
+
"step": 652
|
4574 |
+
},
|
4575 |
+
{
|
4576 |
+
"epoch": 1.2503590234561992,
|
4577 |
+
"grad_norm": 1.7109375,
|
4578 |
+
"learning_rate": 6.747482571381556e-06,
|
4579 |
+
"loss": 1.701,
|
4580 |
+
"step": 653
|
4581 |
+
},
|
4582 |
+
{
|
4583 |
+
"epoch": 1.2522738152225945,
|
4584 |
+
"grad_norm": 1.6875,
|
4585 |
+
"learning_rate": 6.717521403925892e-06,
|
4586 |
+
"loss": 1.7061,
|
4587 |
+
"step": 654
|
4588 |
+
},
|
4589 |
+
{
|
4590 |
+
"epoch": 1.2541886069889898,
|
4591 |
+
"grad_norm": 1.7890625,
|
4592 |
+
"learning_rate": 6.687593224317533e-06,
|
4593 |
+
"loss": 1.6551,
|
4594 |
+
"step": 655
|
4595 |
+
},
|
4596 |
+
{
|
4597 |
+
"epoch": 1.2561033987553853,
|
4598 |
+
"grad_norm": 1.6796875,
|
4599 |
+
"learning_rate": 6.657698333324991e-06,
|
4600 |
+
"loss": 1.731,
|
4601 |
+
"step": 656
|
4602 |
+
},
|
4603 |
+
{
|
4604 |
+
"epoch": 1.2580181905217809,
|
4605 |
+
"grad_norm": 1.65625,
|
4606 |
+
"learning_rate": 6.627837031382246e-06,
|
4607 |
+
"loss": 1.7075,
|
4608 |
+
"step": 657
|
4609 |
+
},
|
4610 |
+
{
|
4611 |
+
"epoch": 1.2599329822881762,
|
4612 |
+
"grad_norm": 1.65625,
|
4613 |
+
"learning_rate": 6.598009618585717e-06,
|
4614 |
+
"loss": 1.6765,
|
4615 |
+
"step": 658
|
4616 |
+
},
|
4617 |
+
{
|
4618 |
+
"epoch": 1.2618477740545715,
|
4619 |
+
"grad_norm": 1.765625,
|
4620 |
+
"learning_rate": 6.568216394691245e-06,
|
4621 |
+
"loss": 1.7378,
|
4622 |
+
"step": 659
|
4623 |
+
},
|
4624 |
+
{
|
4625 |
+
"epoch": 1.263762565820967,
|
4626 |
+
"grad_norm": 1.78125,
|
4627 |
+
"learning_rate": 6.538457659111084e-06,
|
4628 |
+
"loss": 1.7609,
|
4629 |
+
"step": 660
|
4630 |
+
},
|
4631 |
+
{
|
4632 |
+
"epoch": 1.2656773575873623,
|
4633 |
+
"grad_norm": 1.6953125,
|
4634 |
+
"learning_rate": 6.5087337109109e-06,
|
4635 |
+
"loss": 1.7036,
|
4636 |
+
"step": 661
|
4637 |
+
},
|
4638 |
+
{
|
4639 |
+
"epoch": 1.2675921493537579,
|
4640 |
+
"grad_norm": 1.6875,
|
4641 |
+
"learning_rate": 6.479044848806739e-06,
|
4642 |
+
"loss": 1.6546,
|
4643 |
+
"step": 662
|
4644 |
+
},
|
4645 |
+
{
|
4646 |
+
"epoch": 1.2695069411201532,
|
4647 |
+
"grad_norm": 1.7421875,
|
4648 |
+
"learning_rate": 6.4493913711620685e-06,
|
4649 |
+
"loss": 1.7018,
|
4650 |
+
"step": 663
|
4651 |
+
},
|
4652 |
+
{
|
4653 |
+
"epoch": 1.2714217328865485,
|
4654 |
+
"grad_norm": 1.703125,
|
4655 |
+
"learning_rate": 6.419773575984727e-06,
|
4656 |
+
"loss": 1.7357,
|
4657 |
+
"step": 664
|
4658 |
+
},
|
4659 |
+
{
|
4660 |
+
"epoch": 1.273336524652944,
|
4661 |
+
"grad_norm": 1.6875,
|
4662 |
+
"learning_rate": 6.390191760923978e-06,
|
4663 |
+
"loss": 1.6928,
|
4664 |
+
"step": 665
|
4665 |
+
},
|
4666 |
+
{
|
4667 |
+
"epoch": 1.2752513164193395,
|
4668 |
+
"grad_norm": 1.7734375,
|
4669 |
+
"learning_rate": 6.360646223267477e-06,
|
4670 |
+
"loss": 1.6623,
|
4671 |
+
"step": 666
|
4672 |
+
},
|
4673 |
+
{
|
4674 |
+
"epoch": 1.2771661081857348,
|
4675 |
+
"grad_norm": 1.84375,
|
4676 |
+
"learning_rate": 6.3311372599383245e-06,
|
4677 |
+
"loss": 1.6921,
|
4678 |
+
"step": 667
|
4679 |
+
},
|
4680 |
+
{
|
4681 |
+
"epoch": 1.2790808999521301,
|
4682 |
+
"grad_norm": 1.78125,
|
4683 |
+
"learning_rate": 6.301665167492037e-06,
|
4684 |
+
"loss": 1.7036,
|
4685 |
+
"step": 668
|
4686 |
+
},
|
4687 |
+
{
|
4688 |
+
"epoch": 1.2809956917185257,
|
4689 |
+
"grad_norm": 1.7578125,
|
4690 |
+
"learning_rate": 6.272230242113613e-06,
|
4691 |
+
"loss": 1.7099,
|
4692 |
+
"step": 669
|
4693 |
+
},
|
4694 |
+
{
|
4695 |
+
"epoch": 1.282910483484921,
|
4696 |
+
"grad_norm": 1.6875,
|
4697 |
+
"learning_rate": 6.242832779614521e-06,
|
4698 |
+
"loss": 1.6826,
|
4699 |
+
"step": 670
|
4700 |
+
},
|
4701 |
+
{
|
4702 |
+
"epoch": 1.2848252752513165,
|
4703 |
+
"grad_norm": 1.8046875,
|
4704 |
+
"learning_rate": 6.213473075429741e-06,
|
4705 |
+
"loss": 1.7058,
|
4706 |
+
"step": 671
|
4707 |
+
},
|
4708 |
+
{
|
4709 |
+
"epoch": 1.2867400670177118,
|
4710 |
+
"grad_norm": 1.8515625,
|
4711 |
+
"learning_rate": 6.184151424614795e-06,
|
4712 |
+
"loss": 1.6677,
|
4713 |
+
"step": 672
|
4714 |
+
},
|
4715 |
+
{
|
4716 |
+
"epoch": 1.2886548587841071,
|
4717 |
+
"grad_norm": 1.7578125,
|
4718 |
+
"learning_rate": 6.154868121842788e-06,
|
4719 |
+
"loss": 1.7125,
|
4720 |
+
"step": 673
|
4721 |
+
},
|
4722 |
+
{
|
4723 |
+
"epoch": 1.2905696505505027,
|
4724 |
+
"grad_norm": 1.7734375,
|
4725 |
+
"learning_rate": 6.1256234614014256e-06,
|
4726 |
+
"loss": 1.6581,
|
4727 |
+
"step": 674
|
4728 |
+
},
|
4729 |
+
{
|
4730 |
+
"epoch": 1.292484442316898,
|
4731 |
+
"grad_norm": 1.78125,
|
4732 |
+
"learning_rate": 6.096417737190085e-06,
|
4733 |
+
"loss": 1.7104,
|
4734 |
+
"step": 675
|
4735 |
+
},
|
4736 |
+
{
|
4737 |
+
"epoch": 1.2943992340832935,
|
4738 |
+
"grad_norm": 1.671875,
|
4739 |
+
"learning_rate": 6.067251242716838e-06,
|
4740 |
+
"loss": 1.6612,
|
4741 |
+
"step": 676
|
4742 |
+
},
|
4743 |
+
{
|
4744 |
+
"epoch": 1.2963140258496888,
|
4745 |
+
"grad_norm": 1.75,
|
4746 |
+
"learning_rate": 6.038124271095507e-06,
|
4747 |
+
"loss": 1.6501,
|
4748 |
+
"step": 677
|
4749 |
+
},
|
4750 |
+
{
|
4751 |
+
"epoch": 1.2982288176160843,
|
4752 |
+
"grad_norm": 1.734375,
|
4753 |
+
"learning_rate": 6.0090371150427375e-06,
|
4754 |
+
"loss": 1.7283,
|
4755 |
+
"step": 678
|
4756 |
+
},
|
4757 |
+
{
|
4758 |
+
"epoch": 1.3001436093824796,
|
4759 |
+
"grad_norm": 1.7265625,
|
4760 |
+
"learning_rate": 5.979990066875022e-06,
|
4761 |
+
"loss": 1.699,
|
4762 |
+
"step": 679
|
4763 |
+
},
|
4764 |
+
{
|
4765 |
+
"epoch": 1.3020584011488752,
|
4766 |
+
"grad_norm": 1.84375,
|
4767 |
+
"learning_rate": 5.950983418505799e-06,
|
4768 |
+
"loss": 1.7458,
|
4769 |
+
"step": 680
|
4770 |
+
},
|
4771 |
+
{
|
4772 |
+
"epoch": 1.3039731929152705,
|
4773 |
+
"grad_norm": 1.703125,
|
4774 |
+
"learning_rate": 5.922017461442492e-06,
|
4775 |
+
"loss": 1.6889,
|
4776 |
+
"step": 681
|
4777 |
+
},
|
4778 |
+
{
|
4779 |
+
"epoch": 1.3058879846816658,
|
4780 |
+
"grad_norm": 1.7265625,
|
4781 |
+
"learning_rate": 5.893092486783594e-06,
|
4782 |
+
"loss": 1.6935,
|
4783 |
+
"step": 682
|
4784 |
+
},
|
4785 |
+
{
|
4786 |
+
"epoch": 1.3078027764480613,
|
4787 |
+
"grad_norm": 1.7109375,
|
4788 |
+
"learning_rate": 5.864208785215732e-06,
|
4789 |
+
"loss": 1.6641,
|
4790 |
+
"step": 683
|
4791 |
+
},
|
4792 |
+
{
|
4793 |
+
"epoch": 1.3097175682144566,
|
4794 |
+
"grad_norm": 1.734375,
|
4795 |
+
"learning_rate": 5.835366647010767e-06,
|
4796 |
+
"loss": 1.7062,
|
4797 |
+
"step": 684
|
4798 |
+
},
|
4799 |
+
{
|
4800 |
+
"epoch": 1.3116323599808521,
|
4801 |
+
"grad_norm": 1.6796875,
|
4802 |
+
"learning_rate": 5.8065663620228404e-06,
|
4803 |
+
"loss": 1.7008,
|
4804 |
+
"step": 685
|
4805 |
+
},
|
4806 |
+
{
|
4807 |
+
"epoch": 1.3135471517472475,
|
4808 |
+
"grad_norm": 1.7265625,
|
4809 |
+
"learning_rate": 5.777808219685496e-06,
|
4810 |
+
"loss": 1.7002,
|
4811 |
+
"step": 686
|
4812 |
+
},
|
4813 |
+
{
|
4814 |
+
"epoch": 1.3154619435136428,
|
4815 |
+
"grad_norm": 1.7265625,
|
4816 |
+
"learning_rate": 5.749092509008761e-06,
|
4817 |
+
"loss": 1.6896,
|
4818 |
+
"step": 687
|
4819 |
+
},
|
4820 |
+
{
|
4821 |
+
"epoch": 1.3173767352800383,
|
4822 |
+
"grad_norm": 1.7578125,
|
4823 |
+
"learning_rate": 5.720419518576223e-06,
|
4824 |
+
"loss": 1.7014,
|
4825 |
+
"step": 688
|
4826 |
+
},
|
4827 |
+
{
|
4828 |
+
"epoch": 1.3192915270464338,
|
4829 |
+
"grad_norm": 1.75,
|
4830 |
+
"learning_rate": 5.691789536542161e-06,
|
4831 |
+
"loss": 1.6799,
|
4832 |
+
"step": 689
|
4833 |
+
},
|
4834 |
+
{
|
4835 |
+
"epoch": 1.3212063188128291,
|
4836 |
+
"grad_norm": 1.7265625,
|
4837 |
+
"learning_rate": 5.6632028506286266e-06,
|
4838 |
+
"loss": 1.6558,
|
4839 |
+
"step": 690
|
4840 |
+
},
|
4841 |
+
{
|
4842 |
+
"epoch": 1.3231211105792244,
|
4843 |
+
"grad_norm": 1.6640625,
|
4844 |
+
"learning_rate": 5.634659748122552e-06,
|
4845 |
+
"loss": 1.6286,
|
4846 |
+
"step": 691
|
4847 |
+
},
|
4848 |
+
{
|
4849 |
+
"epoch": 1.32503590234562,
|
4850 |
+
"grad_norm": 1.6796875,
|
4851 |
+
"learning_rate": 5.606160515872886e-06,
|
4852 |
+
"loss": 1.6983,
|
4853 |
+
"step": 692
|
4854 |
+
},
|
4855 |
+
{
|
4856 |
+
"epoch": 1.3269506941120153,
|
4857 |
+
"grad_norm": 1.8359375,
|
4858 |
+
"learning_rate": 5.57770544028768e-06,
|
4859 |
+
"loss": 1.7115,
|
4860 |
+
"step": 693
|
4861 |
+
},
|
4862 |
+
{
|
4863 |
+
"epoch": 1.3288654858784108,
|
4864 |
+
"grad_norm": 1.7265625,
|
4865 |
+
"learning_rate": 5.5492948073312406e-06,
|
4866 |
+
"loss": 1.719,
|
4867 |
+
"step": 694
|
4868 |
+
},
|
4869 |
+
{
|
4870 |
+
"epoch": 1.330780277644806,
|
4871 |
+
"grad_norm": 1.7421875,
|
4872 |
+
"learning_rate": 5.520928902521221e-06,
|
4873 |
+
"loss": 1.7074,
|
4874 |
+
"step": 695
|
4875 |
+
},
|
4876 |
+
{
|
4877 |
+
"epoch": 1.3326950694112014,
|
4878 |
+
"grad_norm": 1.65625,
|
4879 |
+
"learning_rate": 5.492608010925793e-06,
|
4880 |
+
"loss": 1.7135,
|
4881 |
+
"step": 696
|
4882 |
+
},
|
4883 |
+
{
|
4884 |
+
"epoch": 1.334609861177597,
|
4885 |
+
"grad_norm": 1.7109375,
|
4886 |
+
"learning_rate": 5.46433241716074e-06,
|
4887 |
+
"loss": 1.6395,
|
4888 |
+
"step": 697
|
4889 |
+
},
|
4890 |
+
{
|
4891 |
+
"epoch": 1.3365246529439925,
|
4892 |
+
"grad_norm": 1.90625,
|
4893 |
+
"learning_rate": 5.436102405386636e-06,
|
4894 |
+
"loss": 1.7543,
|
4895 |
+
"step": 698
|
4896 |
+
},
|
4897 |
+
{
|
4898 |
+
"epoch": 1.3384394447103878,
|
4899 |
+
"grad_norm": 1.78125,
|
4900 |
+
"learning_rate": 5.407918259305951e-06,
|
4901 |
+
"loss": 1.6431,
|
4902 |
+
"step": 699
|
4903 |
+
},
|
4904 |
+
{
|
4905 |
+
"epoch": 1.340354236476783,
|
4906 |
+
"grad_norm": 1.7109375,
|
4907 |
+
"learning_rate": 5.379780262160237e-06,
|
4908 |
+
"loss": 1.7222,
|
4909 |
+
"step": 700
|
4910 |
}
|
4911 |
],
|
4912 |
"logging_steps": 1,
|
|
|
4926 |
"attributes": {}
|
4927 |
}
|
4928 |
},
|
4929 |
+
"total_flos": 4.2104340592931635e+17,
|
4930 |
"train_batch_size": 10,
|
4931 |
"trial_name": null,
|
4932 |
"trial_params": null
|