Training in progress, step 750, 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 671149168
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:20576ea29e8be0c4360c63fcfd6599fe6ed1ccaf070aacb265c8263331c04b3a
|
3 |
size 671149168
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 341314644
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4f510575b1aa3e1648b5bfc57047308e06b6702537cb3135095ab24c37a88337
|
3 |
size 341314644
|
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 sha256:5308048dd0c014787972e4c3767cd80f6a3368dfd7ebb6f842008ecf91ddc070
|
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 sha256:6a197a1a4d059a9ce0958158bb28df1470f0554d236703a402b461e53e77ae60
|
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": 150,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -4247,6 +4247,1064 @@
|
|
4247 |
"eval_samples_per_second": 21.26,
|
4248 |
"eval_steps_per_second": 2.659,
|
4249 |
"step": 600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4250 |
}
|
4251 |
],
|
4252 |
"logging_steps": 1,
|
@@ -4275,7 +5333,7 @@
|
|
4275 |
"attributes": {}
|
4276 |
}
|
4277 |
},
|
4278 |
-
"total_flos":
|
4279 |
"train_batch_size": 12,
|
4280 |
"trial_name": null,
|
4281 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": 0.9616905450820923,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-750",
|
4 |
+
"epoch": 0.3765375282403146,
|
5 |
"eval_steps": 150,
|
6 |
+
"global_step": 750,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
4247 |
"eval_samples_per_second": 21.26,
|
4248 |
"eval_steps_per_second": 2.659,
|
4249 |
"step": 600
|
4250 |
+
},
|
4251 |
+
{
|
4252 |
+
"epoch": 0.3017320726299054,
|
4253 |
+
"grad_norm": 0.445065438747406,
|
4254 |
+
"learning_rate": 6.655688009427832e-05,
|
4255 |
+
"loss": 1.2529,
|
4256 |
+
"step": 601
|
4257 |
+
},
|
4258 |
+
{
|
4259 |
+
"epoch": 0.3022341226675592,
|
4260 |
+
"grad_norm": 0.39952167868614197,
|
4261 |
+
"learning_rate": 6.645669584873494e-05,
|
4262 |
+
"loss": 1.2194,
|
4263 |
+
"step": 602
|
4264 |
+
},
|
4265 |
+
{
|
4266 |
+
"epoch": 0.30273617270521297,
|
4267 |
+
"grad_norm": 0.403266042470932,
|
4268 |
+
"learning_rate": 6.635643745188734e-05,
|
4269 |
+
"loss": 1.2289,
|
4270 |
+
"step": 603
|
4271 |
+
},
|
4272 |
+
{
|
4273 |
+
"epoch": 0.3032382227428667,
|
4274 |
+
"grad_norm": 0.38917073607444763,
|
4275 |
+
"learning_rate": 6.625610535548418e-05,
|
4276 |
+
"loss": 1.1336,
|
4277 |
+
"step": 604
|
4278 |
+
},
|
4279 |
+
{
|
4280 |
+
"epoch": 0.30374027278052046,
|
4281 |
+
"grad_norm": 0.4072120785713196,
|
4282 |
+
"learning_rate": 6.615570001160626e-05,
|
4283 |
+
"loss": 1.0642,
|
4284 |
+
"step": 605
|
4285 |
+
},
|
4286 |
+
{
|
4287 |
+
"epoch": 0.3042423228181742,
|
4288 |
+
"grad_norm": 0.4204983711242676,
|
4289 |
+
"learning_rate": 6.605522187266441e-05,
|
4290 |
+
"loss": 1.0719,
|
4291 |
+
"step": 606
|
4292 |
+
},
|
4293 |
+
{
|
4294 |
+
"epoch": 0.30474437285582795,
|
4295 |
+
"grad_norm": 0.39132463932037354,
|
4296 |
+
"learning_rate": 6.595467139139743e-05,
|
4297 |
+
"loss": 1.0398,
|
4298 |
+
"step": 607
|
4299 |
+
},
|
4300 |
+
{
|
4301 |
+
"epoch": 0.3052464228934817,
|
4302 |
+
"grad_norm": 0.35773175954818726,
|
4303 |
+
"learning_rate": 6.585404902087011e-05,
|
4304 |
+
"loss": 1.0631,
|
4305 |
+
"step": 608
|
4306 |
+
},
|
4307 |
+
{
|
4308 |
+
"epoch": 0.3057484729311355,
|
4309 |
+
"grad_norm": 0.36051151156425476,
|
4310 |
+
"learning_rate": 6.575335521447114e-05,
|
4311 |
+
"loss": 1.04,
|
4312 |
+
"step": 609
|
4313 |
+
},
|
4314 |
+
{
|
4315 |
+
"epoch": 0.30625052296878924,
|
4316 |
+
"grad_norm": 0.36739856004714966,
|
4317 |
+
"learning_rate": 6.565259042591113e-05,
|
4318 |
+
"loss": 1.0239,
|
4319 |
+
"step": 610
|
4320 |
+
},
|
4321 |
+
{
|
4322 |
+
"epoch": 0.306752573006443,
|
4323 |
+
"grad_norm": 0.3616657853126526,
|
4324 |
+
"learning_rate": 6.555175510922047e-05,
|
4325 |
+
"loss": 1.0545,
|
4326 |
+
"step": 611
|
4327 |
+
},
|
4328 |
+
{
|
4329 |
+
"epoch": 0.30725462304409673,
|
4330 |
+
"grad_norm": 0.3667794167995453,
|
4331 |
+
"learning_rate": 6.545084971874738e-05,
|
4332 |
+
"loss": 0.9624,
|
4333 |
+
"step": 612
|
4334 |
+
},
|
4335 |
+
{
|
4336 |
+
"epoch": 0.3077566730817505,
|
4337 |
+
"grad_norm": 0.3631950318813324,
|
4338 |
+
"learning_rate": 6.53498747091558e-05,
|
4339 |
+
"loss": 1.0004,
|
4340 |
+
"step": 613
|
4341 |
+
},
|
4342 |
+
{
|
4343 |
+
"epoch": 0.3082587231194042,
|
4344 |
+
"grad_norm": 0.35089895129203796,
|
4345 |
+
"learning_rate": 6.524883053542339e-05,
|
4346 |
+
"loss": 1.0094,
|
4347 |
+
"step": 614
|
4348 |
+
},
|
4349 |
+
{
|
4350 |
+
"epoch": 0.30876077315705797,
|
4351 |
+
"grad_norm": 0.38375306129455566,
|
4352 |
+
"learning_rate": 6.514771765283942e-05,
|
4353 |
+
"loss": 1.018,
|
4354 |
+
"step": 615
|
4355 |
+
},
|
4356 |
+
{
|
4357 |
+
"epoch": 0.30926282319471177,
|
4358 |
+
"grad_norm": 0.3634318709373474,
|
4359 |
+
"learning_rate": 6.504653651700278e-05,
|
4360 |
+
"loss": 1.0375,
|
4361 |
+
"step": 616
|
4362 |
+
},
|
4363 |
+
{
|
4364 |
+
"epoch": 0.3097648732323655,
|
4365 |
+
"grad_norm": 0.3617091774940491,
|
4366 |
+
"learning_rate": 6.494528758381984e-05,
|
4367 |
+
"loss": 1.0412,
|
4368 |
+
"step": 617
|
4369 |
+
},
|
4370 |
+
{
|
4371 |
+
"epoch": 0.31026692327001926,
|
4372 |
+
"grad_norm": 0.3729401230812073,
|
4373 |
+
"learning_rate": 6.484397130950254e-05,
|
4374 |
+
"loss": 1.0327,
|
4375 |
+
"step": 618
|
4376 |
+
},
|
4377 |
+
{
|
4378 |
+
"epoch": 0.310768973307673,
|
4379 |
+
"grad_norm": 0.3525683581829071,
|
4380 |
+
"learning_rate": 6.474258815056622e-05,
|
4381 |
+
"loss": 1.0164,
|
4382 |
+
"step": 619
|
4383 |
+
},
|
4384 |
+
{
|
4385 |
+
"epoch": 0.31127102334532675,
|
4386 |
+
"grad_norm": 0.3672581911087036,
|
4387 |
+
"learning_rate": 6.464113856382752e-05,
|
4388 |
+
"loss": 1.0148,
|
4389 |
+
"step": 620
|
4390 |
+
},
|
4391 |
+
{
|
4392 |
+
"epoch": 0.3117730733829805,
|
4393 |
+
"grad_norm": 0.3790574371814728,
|
4394 |
+
"learning_rate": 6.453962300640249e-05,
|
4395 |
+
"loss": 0.9997,
|
4396 |
+
"step": 621
|
4397 |
+
},
|
4398 |
+
{
|
4399 |
+
"epoch": 0.31227512342063424,
|
4400 |
+
"grad_norm": 0.36040011048316956,
|
4401 |
+
"learning_rate": 6.44380419357044e-05,
|
4402 |
+
"loss": 0.9505,
|
4403 |
+
"step": 622
|
4404 |
+
},
|
4405 |
+
{
|
4406 |
+
"epoch": 0.312777173458288,
|
4407 |
+
"grad_norm": 0.3569061756134033,
|
4408 |
+
"learning_rate": 6.43363958094417e-05,
|
4409 |
+
"loss": 0.9429,
|
4410 |
+
"step": 623
|
4411 |
+
},
|
4412 |
+
{
|
4413 |
+
"epoch": 0.3132792234959418,
|
4414 |
+
"grad_norm": 0.36146458983421326,
|
4415 |
+
"learning_rate": 6.423468508561599e-05,
|
4416 |
+
"loss": 0.9924,
|
4417 |
+
"step": 624
|
4418 |
+
},
|
4419 |
+
{
|
4420 |
+
"epoch": 0.31378127353359553,
|
4421 |
+
"grad_norm": 0.37957096099853516,
|
4422 |
+
"learning_rate": 6.413291022251989e-05,
|
4423 |
+
"loss": 0.9934,
|
4424 |
+
"step": 625
|
4425 |
+
},
|
4426 |
+
{
|
4427 |
+
"epoch": 0.3142833235712493,
|
4428 |
+
"grad_norm": 0.37144365906715393,
|
4429 |
+
"learning_rate": 6.403107167873509e-05,
|
4430 |
+
"loss": 0.9251,
|
4431 |
+
"step": 626
|
4432 |
+
},
|
4433 |
+
{
|
4434 |
+
"epoch": 0.314785373608903,
|
4435 |
+
"grad_norm": 0.3828261196613312,
|
4436 |
+
"learning_rate": 6.392916991313016e-05,
|
4437 |
+
"loss": 0.9649,
|
4438 |
+
"step": 627
|
4439 |
+
},
|
4440 |
+
{
|
4441 |
+
"epoch": 0.31528742364655676,
|
4442 |
+
"grad_norm": 0.3864898681640625,
|
4443 |
+
"learning_rate": 6.382720538485856e-05,
|
4444 |
+
"loss": 0.9834,
|
4445 |
+
"step": 628
|
4446 |
+
},
|
4447 |
+
{
|
4448 |
+
"epoch": 0.3157894736842105,
|
4449 |
+
"grad_norm": 0.3928738832473755,
|
4450 |
+
"learning_rate": 6.372517855335655e-05,
|
4451 |
+
"loss": 0.9759,
|
4452 |
+
"step": 629
|
4453 |
+
},
|
4454 |
+
{
|
4455 |
+
"epoch": 0.31629152372186425,
|
4456 |
+
"grad_norm": 0.42996037006378174,
|
4457 |
+
"learning_rate": 6.362308987834115e-05,
|
4458 |
+
"loss": 0.9628,
|
4459 |
+
"step": 630
|
4460 |
+
},
|
4461 |
+
{
|
4462 |
+
"epoch": 0.31679357375951805,
|
4463 |
+
"grad_norm": 0.3807196319103241,
|
4464 |
+
"learning_rate": 6.352093981980796e-05,
|
4465 |
+
"loss": 0.9842,
|
4466 |
+
"step": 631
|
4467 |
+
},
|
4468 |
+
{
|
4469 |
+
"epoch": 0.3172956237971718,
|
4470 |
+
"grad_norm": 0.39248624444007874,
|
4471 |
+
"learning_rate": 6.341872883802923e-05,
|
4472 |
+
"loss": 0.9539,
|
4473 |
+
"step": 632
|
4474 |
+
},
|
4475 |
+
{
|
4476 |
+
"epoch": 0.31779767383482554,
|
4477 |
+
"grad_norm": 0.4059353470802307,
|
4478 |
+
"learning_rate": 6.331645739355168e-05,
|
4479 |
+
"loss": 0.9635,
|
4480 |
+
"step": 633
|
4481 |
+
},
|
4482 |
+
{
|
4483 |
+
"epoch": 0.3182997238724793,
|
4484 |
+
"grad_norm": 0.4235178828239441,
|
4485 |
+
"learning_rate": 6.321412594719451e-05,
|
4486 |
+
"loss": 0.9473,
|
4487 |
+
"step": 634
|
4488 |
+
},
|
4489 |
+
{
|
4490 |
+
"epoch": 0.31880177391013304,
|
4491 |
+
"grad_norm": 0.45633211731910706,
|
4492 |
+
"learning_rate": 6.311173496004723e-05,
|
4493 |
+
"loss": 0.9836,
|
4494 |
+
"step": 635
|
4495 |
+
},
|
4496 |
+
{
|
4497 |
+
"epoch": 0.3193038239477868,
|
4498 |
+
"grad_norm": 0.4051073491573334,
|
4499 |
+
"learning_rate": 6.300928489346766e-05,
|
4500 |
+
"loss": 0.9482,
|
4501 |
+
"step": 636
|
4502 |
+
},
|
4503 |
+
{
|
4504 |
+
"epoch": 0.3198058739854405,
|
4505 |
+
"grad_norm": 0.4133238196372986,
|
4506 |
+
"learning_rate": 6.290677620907982e-05,
|
4507 |
+
"loss": 0.9009,
|
4508 |
+
"step": 637
|
4509 |
+
},
|
4510 |
+
{
|
4511 |
+
"epoch": 0.3203079240230943,
|
4512 |
+
"grad_norm": 0.4294078052043915,
|
4513 |
+
"learning_rate": 6.280420936877188e-05,
|
4514 |
+
"loss": 0.9389,
|
4515 |
+
"step": 638
|
4516 |
+
},
|
4517 |
+
{
|
4518 |
+
"epoch": 0.32080997406074807,
|
4519 |
+
"grad_norm": 0.4092111885547638,
|
4520 |
+
"learning_rate": 6.270158483469397e-05,
|
4521 |
+
"loss": 0.8397,
|
4522 |
+
"step": 639
|
4523 |
+
},
|
4524 |
+
{
|
4525 |
+
"epoch": 0.3213120240984018,
|
4526 |
+
"grad_norm": 0.42124441266059875,
|
4527 |
+
"learning_rate": 6.259890306925627e-05,
|
4528 |
+
"loss": 0.8405,
|
4529 |
+
"step": 640
|
4530 |
+
},
|
4531 |
+
{
|
4532 |
+
"epoch": 0.32181407413605556,
|
4533 |
+
"grad_norm": 0.4422035217285156,
|
4534 |
+
"learning_rate": 6.249616453512677e-05,
|
4535 |
+
"loss": 0.8641,
|
4536 |
+
"step": 641
|
4537 |
+
},
|
4538 |
+
{
|
4539 |
+
"epoch": 0.3223161241737093,
|
4540 |
+
"grad_norm": 0.4448348879814148,
|
4541 |
+
"learning_rate": 6.239336969522932e-05,
|
4542 |
+
"loss": 0.9077,
|
4543 |
+
"step": 642
|
4544 |
+
},
|
4545 |
+
{
|
4546 |
+
"epoch": 0.32281817421136305,
|
4547 |
+
"grad_norm": 0.4691510796546936,
|
4548 |
+
"learning_rate": 6.229051901274137e-05,
|
4549 |
+
"loss": 0.8585,
|
4550 |
+
"step": 643
|
4551 |
+
},
|
4552 |
+
{
|
4553 |
+
"epoch": 0.3233202242490168,
|
4554 |
+
"grad_norm": 0.4641557037830353,
|
4555 |
+
"learning_rate": 6.218761295109208e-05,
|
4556 |
+
"loss": 0.8527,
|
4557 |
+
"step": 644
|
4558 |
+
},
|
4559 |
+
{
|
4560 |
+
"epoch": 0.3238222742866706,
|
4561 |
+
"grad_norm": 0.5288779735565186,
|
4562 |
+
"learning_rate": 6.208465197396013e-05,
|
4563 |
+
"loss": 0.8489,
|
4564 |
+
"step": 645
|
4565 |
+
},
|
4566 |
+
{
|
4567 |
+
"epoch": 0.32432432432432434,
|
4568 |
+
"grad_norm": 0.45869073271751404,
|
4569 |
+
"learning_rate": 6.19816365452716e-05,
|
4570 |
+
"loss": 0.8505,
|
4571 |
+
"step": 646
|
4572 |
+
},
|
4573 |
+
{
|
4574 |
+
"epoch": 0.3248263743619781,
|
4575 |
+
"grad_norm": 0.49422523379325867,
|
4576 |
+
"learning_rate": 6.187856712919795e-05,
|
4577 |
+
"loss": 0.8555,
|
4578 |
+
"step": 647
|
4579 |
+
},
|
4580 |
+
{
|
4581 |
+
"epoch": 0.32532842439963183,
|
4582 |
+
"grad_norm": 0.5668922066688538,
|
4583 |
+
"learning_rate": 6.177544419015388e-05,
|
4584 |
+
"loss": 0.7629,
|
4585 |
+
"step": 648
|
4586 |
+
},
|
4587 |
+
{
|
4588 |
+
"epoch": 0.3258304744372856,
|
4589 |
+
"grad_norm": 0.5716300010681152,
|
4590 |
+
"learning_rate": 6.167226819279528e-05,
|
4591 |
+
"loss": 0.8643,
|
4592 |
+
"step": 649
|
4593 |
+
},
|
4594 |
+
{
|
4595 |
+
"epoch": 0.3263325244749393,
|
4596 |
+
"grad_norm": 0.6652288436889648,
|
4597 |
+
"learning_rate": 6.156903960201709e-05,
|
4598 |
+
"loss": 0.7433,
|
4599 |
+
"step": 650
|
4600 |
+
},
|
4601 |
+
{
|
4602 |
+
"epoch": 0.32683457451259307,
|
4603 |
+
"grad_norm": 0.6001056432723999,
|
4604 |
+
"learning_rate": 6.146575888295123e-05,
|
4605 |
+
"loss": 1.2497,
|
4606 |
+
"step": 651
|
4607 |
+
},
|
4608 |
+
{
|
4609 |
+
"epoch": 0.32733662455024687,
|
4610 |
+
"grad_norm": 0.3522529900074005,
|
4611 |
+
"learning_rate": 6.136242650096451e-05,
|
4612 |
+
"loss": 1.177,
|
4613 |
+
"step": 652
|
4614 |
+
},
|
4615 |
+
{
|
4616 |
+
"epoch": 0.3278386745879006,
|
4617 |
+
"grad_norm": 0.3846982717514038,
|
4618 |
+
"learning_rate": 6.125904292165652e-05,
|
4619 |
+
"loss": 1.1357,
|
4620 |
+
"step": 653
|
4621 |
+
},
|
4622 |
+
{
|
4623 |
+
"epoch": 0.32834072462555436,
|
4624 |
+
"grad_norm": 0.389482706785202,
|
4625 |
+
"learning_rate": 6.115560861085756e-05,
|
4626 |
+
"loss": 1.0675,
|
4627 |
+
"step": 654
|
4628 |
+
},
|
4629 |
+
{
|
4630 |
+
"epoch": 0.3288427746632081,
|
4631 |
+
"grad_norm": 0.41399508714675903,
|
4632 |
+
"learning_rate": 6.105212403462651e-05,
|
4633 |
+
"loss": 1.1065,
|
4634 |
+
"step": 655
|
4635 |
+
},
|
4636 |
+
{
|
4637 |
+
"epoch": 0.32934482470086185,
|
4638 |
+
"grad_norm": 0.5792128443717957,
|
4639 |
+
"learning_rate": 6.0948589659248654e-05,
|
4640 |
+
"loss": 1.1188,
|
4641 |
+
"step": 656
|
4642 |
+
},
|
4643 |
+
{
|
4644 |
+
"epoch": 0.3298468747385156,
|
4645 |
+
"grad_norm": 0.3753111958503723,
|
4646 |
+
"learning_rate": 6.084500595123383e-05,
|
4647 |
+
"loss": 1.1127,
|
4648 |
+
"step": 657
|
4649 |
+
},
|
4650 |
+
{
|
4651 |
+
"epoch": 0.33034892477616934,
|
4652 |
+
"grad_norm": 0.3663425147533417,
|
4653 |
+
"learning_rate": 6.0741373377314005e-05,
|
4654 |
+
"loss": 1.019,
|
4655 |
+
"step": 658
|
4656 |
+
},
|
4657 |
+
{
|
4658 |
+
"epoch": 0.3308509748138231,
|
4659 |
+
"grad_norm": 0.39105096459388733,
|
4660 |
+
"learning_rate": 6.0637692404441416e-05,
|
4661 |
+
"loss": 1.0186,
|
4662 |
+
"step": 659
|
4663 |
+
},
|
4664 |
+
{
|
4665 |
+
"epoch": 0.3313530248514769,
|
4666 |
+
"grad_norm": 0.38673144578933716,
|
4667 |
+
"learning_rate": 6.0533963499786314e-05,
|
4668 |
+
"loss": 1.0256,
|
4669 |
+
"step": 660
|
4670 |
+
},
|
4671 |
+
{
|
4672 |
+
"epoch": 0.33185507488913063,
|
4673 |
+
"grad_norm": 0.3633407950401306,
|
4674 |
+
"learning_rate": 6.0430187130735016e-05,
|
4675 |
+
"loss": 1.0332,
|
4676 |
+
"step": 661
|
4677 |
+
},
|
4678 |
+
{
|
4679 |
+
"epoch": 0.3323571249267844,
|
4680 |
+
"grad_norm": 0.35200172662734985,
|
4681 |
+
"learning_rate": 6.032636376488763e-05,
|
4682 |
+
"loss": 0.9356,
|
4683 |
+
"step": 662
|
4684 |
+
},
|
4685 |
+
{
|
4686 |
+
"epoch": 0.3328591749644381,
|
4687 |
+
"grad_norm": 0.3665078282356262,
|
4688 |
+
"learning_rate": 6.0222493870056044e-05,
|
4689 |
+
"loss": 1.0154,
|
4690 |
+
"step": 663
|
4691 |
+
},
|
4692 |
+
{
|
4693 |
+
"epoch": 0.33336122500209187,
|
4694 |
+
"grad_norm": 0.3591248095035553,
|
4695 |
+
"learning_rate": 6.0118577914261784e-05,
|
4696 |
+
"loss": 0.9798,
|
4697 |
+
"step": 664
|
4698 |
+
},
|
4699 |
+
{
|
4700 |
+
"epoch": 0.3338632750397456,
|
4701 |
+
"grad_norm": 0.361217200756073,
|
4702 |
+
"learning_rate": 6.001461636573397e-05,
|
4703 |
+
"loss": 0.9813,
|
4704 |
+
"step": 665
|
4705 |
+
},
|
4706 |
+
{
|
4707 |
+
"epoch": 0.33436532507739936,
|
4708 |
+
"grad_norm": 0.37569659948349,
|
4709 |
+
"learning_rate": 5.99106096929071e-05,
|
4710 |
+
"loss": 1.011,
|
4711 |
+
"step": 666
|
4712 |
+
},
|
4713 |
+
{
|
4714 |
+
"epoch": 0.33486737511505316,
|
4715 |
+
"grad_norm": 0.3692183494567871,
|
4716 |
+
"learning_rate": 5.980655836441902e-05,
|
4717 |
+
"loss": 1.0294,
|
4718 |
+
"step": 667
|
4719 |
+
},
|
4720 |
+
{
|
4721 |
+
"epoch": 0.3353694251527069,
|
4722 |
+
"grad_norm": 0.374726802110672,
|
4723 |
+
"learning_rate": 5.970246284910876e-05,
|
4724 |
+
"loss": 0.9654,
|
4725 |
+
"step": 668
|
4726 |
+
},
|
4727 |
+
{
|
4728 |
+
"epoch": 0.33587147519036065,
|
4729 |
+
"grad_norm": 0.3687571585178375,
|
4730 |
+
"learning_rate": 5.959832361601453e-05,
|
4731 |
+
"loss": 1.0423,
|
4732 |
+
"step": 669
|
4733 |
+
},
|
4734 |
+
{
|
4735 |
+
"epoch": 0.3363735252280144,
|
4736 |
+
"grad_norm": 0.36362433433532715,
|
4737 |
+
"learning_rate": 5.949414113437142e-05,
|
4738 |
+
"loss": 0.8874,
|
4739 |
+
"step": 670
|
4740 |
+
},
|
4741 |
+
{
|
4742 |
+
"epoch": 0.33687557526566814,
|
4743 |
+
"grad_norm": 0.34844672679901123,
|
4744 |
+
"learning_rate": 5.938991587360946e-05,
|
4745 |
+
"loss": 0.8979,
|
4746 |
+
"step": 671
|
4747 |
+
},
|
4748 |
+
{
|
4749 |
+
"epoch": 0.3373776253033219,
|
4750 |
+
"grad_norm": 0.3646034598350525,
|
4751 |
+
"learning_rate": 5.9285648303351404e-05,
|
4752 |
+
"loss": 0.9435,
|
4753 |
+
"step": 672
|
4754 |
+
},
|
4755 |
+
{
|
4756 |
+
"epoch": 0.3378796753409756,
|
4757 |
+
"grad_norm": 0.37094947695732117,
|
4758 |
+
"learning_rate": 5.9181338893410663e-05,
|
4759 |
+
"loss": 0.9679,
|
4760 |
+
"step": 673
|
4761 |
+
},
|
4762 |
+
{
|
4763 |
+
"epoch": 0.3383817253786294,
|
4764 |
+
"grad_norm": 0.385873943567276,
|
4765 |
+
"learning_rate": 5.907698811378919e-05,
|
4766 |
+
"loss": 0.9898,
|
4767 |
+
"step": 674
|
4768 |
+
},
|
4769 |
+
{
|
4770 |
+
"epoch": 0.3388837754162832,
|
4771 |
+
"grad_norm": 0.38623571395874023,
|
4772 |
+
"learning_rate": 5.897259643467527e-05,
|
4773 |
+
"loss": 0.987,
|
4774 |
+
"step": 675
|
4775 |
+
},
|
4776 |
+
{
|
4777 |
+
"epoch": 0.3393858254539369,
|
4778 |
+
"grad_norm": 0.3703857362270355,
|
4779 |
+
"learning_rate": 5.8868164326441546e-05,
|
4780 |
+
"loss": 0.919,
|
4781 |
+
"step": 676
|
4782 |
+
},
|
4783 |
+
{
|
4784 |
+
"epoch": 0.33988787549159066,
|
4785 |
+
"grad_norm": 0.3874402344226837,
|
4786 |
+
"learning_rate": 5.876369225964283e-05,
|
4787 |
+
"loss": 0.959,
|
4788 |
+
"step": 677
|
4789 |
+
},
|
4790 |
+
{
|
4791 |
+
"epoch": 0.3403899255292444,
|
4792 |
+
"grad_norm": 0.37169700860977173,
|
4793 |
+
"learning_rate": 5.8659180705013936e-05,
|
4794 |
+
"loss": 0.9883,
|
4795 |
+
"step": 678
|
4796 |
+
},
|
4797 |
+
{
|
4798 |
+
"epoch": 0.34089197556689815,
|
4799 |
+
"grad_norm": 0.4187929332256317,
|
4800 |
+
"learning_rate": 5.8554630133467624e-05,
|
4801 |
+
"loss": 0.9527,
|
4802 |
+
"step": 679
|
4803 |
+
},
|
4804 |
+
{
|
4805 |
+
"epoch": 0.3413940256045519,
|
4806 |
+
"grad_norm": 0.39550694823265076,
|
4807 |
+
"learning_rate": 5.8450041016092464e-05,
|
4808 |
+
"loss": 0.9152,
|
4809 |
+
"step": 680
|
4810 |
+
},
|
4811 |
+
{
|
4812 |
+
"epoch": 0.3418960756422057,
|
4813 |
+
"grad_norm": 0.40294429659843445,
|
4814 |
+
"learning_rate": 5.83454138241507e-05,
|
4815 |
+
"loss": 0.95,
|
4816 |
+
"step": 681
|
4817 |
+
},
|
4818 |
+
{
|
4819 |
+
"epoch": 0.34239812567985944,
|
4820 |
+
"grad_norm": 0.38999685645103455,
|
4821 |
+
"learning_rate": 5.8240749029076134e-05,
|
4822 |
+
"loss": 0.9475,
|
4823 |
+
"step": 682
|
4824 |
+
},
|
4825 |
+
{
|
4826 |
+
"epoch": 0.3429001757175132,
|
4827 |
+
"grad_norm": 0.40788596868515015,
|
4828 |
+
"learning_rate": 5.8136047102472e-05,
|
4829 |
+
"loss": 1.01,
|
4830 |
+
"step": 683
|
4831 |
+
},
|
4832 |
+
{
|
4833 |
+
"epoch": 0.34340222575516693,
|
4834 |
+
"grad_norm": 0.4204280972480774,
|
4835 |
+
"learning_rate": 5.803130851610886e-05,
|
4836 |
+
"loss": 0.934,
|
4837 |
+
"step": 684
|
4838 |
+
},
|
4839 |
+
{
|
4840 |
+
"epoch": 0.3439042757928207,
|
4841 |
+
"grad_norm": 0.4102809429168701,
|
4842 |
+
"learning_rate": 5.792653374192245e-05,
|
4843 |
+
"loss": 0.9398,
|
4844 |
+
"step": 685
|
4845 |
+
},
|
4846 |
+
{
|
4847 |
+
"epoch": 0.3444063258304744,
|
4848 |
+
"grad_norm": 0.4025559723377228,
|
4849 |
+
"learning_rate": 5.782172325201155e-05,
|
4850 |
+
"loss": 0.9245,
|
4851 |
+
"step": 686
|
4852 |
+
},
|
4853 |
+
{
|
4854 |
+
"epoch": 0.34490837586812817,
|
4855 |
+
"grad_norm": 0.4101907014846802,
|
4856 |
+
"learning_rate": 5.771687751863587e-05,
|
4857 |
+
"loss": 0.9279,
|
4858 |
+
"step": 687
|
4859 |
+
},
|
4860 |
+
{
|
4861 |
+
"epoch": 0.34541042590578197,
|
4862 |
+
"grad_norm": 0.43221110105514526,
|
4863 |
+
"learning_rate": 5.761199701421391e-05,
|
4864 |
+
"loss": 0.8831,
|
4865 |
+
"step": 688
|
4866 |
+
},
|
4867 |
+
{
|
4868 |
+
"epoch": 0.3459124759434357,
|
4869 |
+
"grad_norm": 0.42259782552719116,
|
4870 |
+
"learning_rate": 5.750708221132092e-05,
|
4871 |
+
"loss": 0.8903,
|
4872 |
+
"step": 689
|
4873 |
+
},
|
4874 |
+
{
|
4875 |
+
"epoch": 0.34641452598108946,
|
4876 |
+
"grad_norm": 0.4195202887058258,
|
4877 |
+
"learning_rate": 5.7402133582686576e-05,
|
4878 |
+
"loss": 0.8291,
|
4879 |
+
"step": 690
|
4880 |
+
},
|
4881 |
+
{
|
4882 |
+
"epoch": 0.3469165760187432,
|
4883 |
+
"grad_norm": 0.4531534016132355,
|
4884 |
+
"learning_rate": 5.7297151601193056e-05,
|
4885 |
+
"loss": 0.8893,
|
4886 |
+
"step": 691
|
4887 |
+
},
|
4888 |
+
{
|
4889 |
+
"epoch": 0.34741862605639695,
|
4890 |
+
"grad_norm": 0.46428826451301575,
|
4891 |
+
"learning_rate": 5.719213673987277e-05,
|
4892 |
+
"loss": 0.9049,
|
4893 |
+
"step": 692
|
4894 |
+
},
|
4895 |
+
{
|
4896 |
+
"epoch": 0.3479206760940507,
|
4897 |
+
"grad_norm": 0.4338727295398712,
|
4898 |
+
"learning_rate": 5.708708947190634e-05,
|
4899 |
+
"loss": 0.8142,
|
4900 |
+
"step": 693
|
4901 |
+
},
|
4902 |
+
{
|
4903 |
+
"epoch": 0.34842272613170444,
|
4904 |
+
"grad_norm": 0.44543692469596863,
|
4905 |
+
"learning_rate": 5.698201027062034e-05,
|
4906 |
+
"loss": 0.8463,
|
4907 |
+
"step": 694
|
4908 |
+
},
|
4909 |
+
{
|
4910 |
+
"epoch": 0.3489247761693582,
|
4911 |
+
"grad_norm": 0.4769425094127655,
|
4912 |
+
"learning_rate": 5.6876899609485256e-05,
|
4913 |
+
"loss": 0.8931,
|
4914 |
+
"step": 695
|
4915 |
+
},
|
4916 |
+
{
|
4917 |
+
"epoch": 0.349426826207012,
|
4918 |
+
"grad_norm": 0.49232223629951477,
|
4919 |
+
"learning_rate": 5.6771757962113323e-05,
|
4920 |
+
"loss": 0.8189,
|
4921 |
+
"step": 696
|
4922 |
+
},
|
4923 |
+
{
|
4924 |
+
"epoch": 0.34992887624466573,
|
4925 |
+
"grad_norm": 0.49148690700531006,
|
4926 |
+
"learning_rate": 5.666658580225643e-05,
|
4927 |
+
"loss": 0.8153,
|
4928 |
+
"step": 697
|
4929 |
+
},
|
4930 |
+
{
|
4931 |
+
"epoch": 0.3504309262823195,
|
4932 |
+
"grad_norm": 0.5055503845214844,
|
4933 |
+
"learning_rate": 5.656138360380391e-05,
|
4934 |
+
"loss": 0.8018,
|
4935 |
+
"step": 698
|
4936 |
+
},
|
4937 |
+
{
|
4938 |
+
"epoch": 0.3509329763199732,
|
4939 |
+
"grad_norm": 0.5481170415878296,
|
4940 |
+
"learning_rate": 5.645615184078044e-05,
|
4941 |
+
"loss": 0.8587,
|
4942 |
+
"step": 699
|
4943 |
+
},
|
4944 |
+
{
|
4945 |
+
"epoch": 0.35143502635762697,
|
4946 |
+
"grad_norm": 0.6615381240844727,
|
4947 |
+
"learning_rate": 5.6350890987343944e-05,
|
4948 |
+
"loss": 0.777,
|
4949 |
+
"step": 700
|
4950 |
+
},
|
4951 |
+
{
|
4952 |
+
"epoch": 0.3519370763952807,
|
4953 |
+
"grad_norm": 0.434299111366272,
|
4954 |
+
"learning_rate": 5.6245601517783406e-05,
|
4955 |
+
"loss": 1.2088,
|
4956 |
+
"step": 701
|
4957 |
+
},
|
4958 |
+
{
|
4959 |
+
"epoch": 0.35243912643293446,
|
4960 |
+
"grad_norm": 0.39533907175064087,
|
4961 |
+
"learning_rate": 5.614028390651675e-05,
|
4962 |
+
"loss": 1.1814,
|
4963 |
+
"step": 702
|
4964 |
+
},
|
4965 |
+
{
|
4966 |
+
"epoch": 0.35294117647058826,
|
4967 |
+
"grad_norm": 0.3828687369823456,
|
4968 |
+
"learning_rate": 5.6034938628088705e-05,
|
4969 |
+
"loss": 1.1873,
|
4970 |
+
"step": 703
|
4971 |
+
},
|
4972 |
+
{
|
4973 |
+
"epoch": 0.353443226508242,
|
4974 |
+
"grad_norm": 0.3660382628440857,
|
4975 |
+
"learning_rate": 5.5929566157168665e-05,
|
4976 |
+
"loss": 1.0862,
|
4977 |
+
"step": 704
|
4978 |
+
},
|
4979 |
+
{
|
4980 |
+
"epoch": 0.35394527654589575,
|
4981 |
+
"grad_norm": 0.39876964688301086,
|
4982 |
+
"learning_rate": 5.582416696854853e-05,
|
4983 |
+
"loss": 1.0083,
|
4984 |
+
"step": 705
|
4985 |
+
},
|
4986 |
+
{
|
4987 |
+
"epoch": 0.3544473265835495,
|
4988 |
+
"grad_norm": 0.409247487783432,
|
4989 |
+
"learning_rate": 5.571874153714063e-05,
|
4990 |
+
"loss": 1.0714,
|
4991 |
+
"step": 706
|
4992 |
+
},
|
4993 |
+
{
|
4994 |
+
"epoch": 0.35494937662120324,
|
4995 |
+
"grad_norm": 0.3872778117656708,
|
4996 |
+
"learning_rate": 5.561329033797547e-05,
|
4997 |
+
"loss": 1.085,
|
4998 |
+
"step": 707
|
4999 |
+
},
|
5000 |
+
{
|
5001 |
+
"epoch": 0.355451426658857,
|
5002 |
+
"grad_norm": 0.38185930252075195,
|
5003 |
+
"learning_rate": 5.550781384619973e-05,
|
5004 |
+
"loss": 1.0762,
|
5005 |
+
"step": 708
|
5006 |
+
},
|
5007 |
+
{
|
5008 |
+
"epoch": 0.35595347669651073,
|
5009 |
+
"grad_norm": 0.3866881728172302,
|
5010 |
+
"learning_rate": 5.540231253707403e-05,
|
5011 |
+
"loss": 1.0326,
|
5012 |
+
"step": 709
|
5013 |
+
},
|
5014 |
+
{
|
5015 |
+
"epoch": 0.35645552673416453,
|
5016 |
+
"grad_norm": 0.37910160422325134,
|
5017 |
+
"learning_rate": 5.5296786885970805e-05,
|
5018 |
+
"loss": 1.0769,
|
5019 |
+
"step": 710
|
5020 |
+
},
|
5021 |
+
{
|
5022 |
+
"epoch": 0.3569575767718183,
|
5023 |
+
"grad_norm": 0.3608991205692291,
|
5024 |
+
"learning_rate": 5.519123736837217e-05,
|
5025 |
+
"loss": 1.0523,
|
5026 |
+
"step": 711
|
5027 |
+
},
|
5028 |
+
{
|
5029 |
+
"epoch": 0.357459626809472,
|
5030 |
+
"grad_norm": 0.36697694659233093,
|
5031 |
+
"learning_rate": 5.50856644598678e-05,
|
5032 |
+
"loss": 0.9778,
|
5033 |
+
"step": 712
|
5034 |
+
},
|
5035 |
+
{
|
5036 |
+
"epoch": 0.35796167684712576,
|
5037 |
+
"grad_norm": 0.4545275568962097,
|
5038 |
+
"learning_rate": 5.498006863615275e-05,
|
5039 |
+
"loss": 1.0207,
|
5040 |
+
"step": 713
|
5041 |
+
},
|
5042 |
+
{
|
5043 |
+
"epoch": 0.3584637268847795,
|
5044 |
+
"grad_norm": 0.3483712375164032,
|
5045 |
+
"learning_rate": 5.487445037302531e-05,
|
5046 |
+
"loss": 1.0002,
|
5047 |
+
"step": 714
|
5048 |
+
},
|
5049 |
+
{
|
5050 |
+
"epoch": 0.35896577692243326,
|
5051 |
+
"grad_norm": 0.3665158152580261,
|
5052 |
+
"learning_rate": 5.476881014638491e-05,
|
5053 |
+
"loss": 1.0274,
|
5054 |
+
"step": 715
|
5055 |
+
},
|
5056 |
+
{
|
5057 |
+
"epoch": 0.359467826960087,
|
5058 |
+
"grad_norm": 0.35564157366752625,
|
5059 |
+
"learning_rate": 5.466314843222993e-05,
|
5060 |
+
"loss": 0.9884,
|
5061 |
+
"step": 716
|
5062 |
+
},
|
5063 |
+
{
|
5064 |
+
"epoch": 0.3599698769977408,
|
5065 |
+
"grad_norm": 0.3559761345386505,
|
5066 |
+
"learning_rate": 5.4557465706655564e-05,
|
5067 |
+
"loss": 1.0143,
|
5068 |
+
"step": 717
|
5069 |
+
},
|
5070 |
+
{
|
5071 |
+
"epoch": 0.36047192703539455,
|
5072 |
+
"grad_norm": 0.38508090376853943,
|
5073 |
+
"learning_rate": 5.4451762445851705e-05,
|
5074 |
+
"loss": 1.0679,
|
5075 |
+
"step": 718
|
5076 |
+
},
|
5077 |
+
{
|
5078 |
+
"epoch": 0.3609739770730483,
|
5079 |
+
"grad_norm": 0.3513292670249939,
|
5080 |
+
"learning_rate": 5.4346039126100733e-05,
|
5081 |
+
"loss": 0.948,
|
5082 |
+
"step": 719
|
5083 |
+
},
|
5084 |
+
{
|
5085 |
+
"epoch": 0.36147602711070204,
|
5086 |
+
"grad_norm": 0.36502474546432495,
|
5087 |
+
"learning_rate": 5.4240296223775465e-05,
|
5088 |
+
"loss": 1.0246,
|
5089 |
+
"step": 720
|
5090 |
+
},
|
5091 |
+
{
|
5092 |
+
"epoch": 0.3619780771483558,
|
5093 |
+
"grad_norm": 0.3846004605293274,
|
5094 |
+
"learning_rate": 5.41345342153369e-05,
|
5095 |
+
"loss": 1.0332,
|
5096 |
+
"step": 721
|
5097 |
+
},
|
5098 |
+
{
|
5099 |
+
"epoch": 0.3624801271860095,
|
5100 |
+
"grad_norm": 0.35061997175216675,
|
5101 |
+
"learning_rate": 5.4028753577332146e-05,
|
5102 |
+
"loss": 0.9286,
|
5103 |
+
"step": 722
|
5104 |
+
},
|
5105 |
+
{
|
5106 |
+
"epoch": 0.36298217722366327,
|
5107 |
+
"grad_norm": 0.37235984206199646,
|
5108 |
+
"learning_rate": 5.392295478639225e-05,
|
5109 |
+
"loss": 1.0385,
|
5110 |
+
"step": 723
|
5111 |
+
},
|
5112 |
+
{
|
5113 |
+
"epoch": 0.36348422726131707,
|
5114 |
+
"grad_norm": 0.3770149350166321,
|
5115 |
+
"learning_rate": 5.3817138319230076e-05,
|
5116 |
+
"loss": 0.9865,
|
5117 |
+
"step": 724
|
5118 |
+
},
|
5119 |
+
{
|
5120 |
+
"epoch": 0.3639862772989708,
|
5121 |
+
"grad_norm": 0.3904590606689453,
|
5122 |
+
"learning_rate": 5.3711304652638126e-05,
|
5123 |
+
"loss": 0.934,
|
5124 |
+
"step": 725
|
5125 |
+
},
|
5126 |
+
{
|
5127 |
+
"epoch": 0.36448832733662456,
|
5128 |
+
"grad_norm": 0.3823120892047882,
|
5129 |
+
"learning_rate": 5.360545426348638e-05,
|
5130 |
+
"loss": 0.9394,
|
5131 |
+
"step": 726
|
5132 |
+
},
|
5133 |
+
{
|
5134 |
+
"epoch": 0.3649903773742783,
|
5135 |
+
"grad_norm": 0.36231666803359985,
|
5136 |
+
"learning_rate": 5.349958762872016e-05,
|
5137 |
+
"loss": 0.9282,
|
5138 |
+
"step": 727
|
5139 |
+
},
|
5140 |
+
{
|
5141 |
+
"epoch": 0.36549242741193205,
|
5142 |
+
"grad_norm": 0.3757944405078888,
|
5143 |
+
"learning_rate": 5.3393705225358046e-05,
|
5144 |
+
"loss": 0.8884,
|
5145 |
+
"step": 728
|
5146 |
+
},
|
5147 |
+
{
|
5148 |
+
"epoch": 0.3659944774495858,
|
5149 |
+
"grad_norm": 0.4007607102394104,
|
5150 |
+
"learning_rate": 5.32878075304896e-05,
|
5151 |
+
"loss": 0.9739,
|
5152 |
+
"step": 729
|
5153 |
+
},
|
5154 |
+
{
|
5155 |
+
"epoch": 0.36649652748723954,
|
5156 |
+
"grad_norm": 0.40476924180984497,
|
5157 |
+
"learning_rate": 5.318189502127332e-05,
|
5158 |
+
"loss": 0.9458,
|
5159 |
+
"step": 730
|
5160 |
+
},
|
5161 |
+
{
|
5162 |
+
"epoch": 0.3669985775248933,
|
5163 |
+
"grad_norm": 0.39884302020072937,
|
5164 |
+
"learning_rate": 5.307596817493445e-05,
|
5165 |
+
"loss": 0.8989,
|
5166 |
+
"step": 731
|
5167 |
+
},
|
5168 |
+
{
|
5169 |
+
"epoch": 0.3675006275625471,
|
5170 |
+
"grad_norm": 0.42604318261146545,
|
5171 |
+
"learning_rate": 5.297002746876284e-05,
|
5172 |
+
"loss": 0.9337,
|
5173 |
+
"step": 732
|
5174 |
+
},
|
5175 |
+
{
|
5176 |
+
"epoch": 0.36800267760020083,
|
5177 |
+
"grad_norm": 0.41235285997390747,
|
5178 |
+
"learning_rate": 5.286407338011079e-05,
|
5179 |
+
"loss": 0.9191,
|
5180 |
+
"step": 733
|
5181 |
+
},
|
5182 |
+
{
|
5183 |
+
"epoch": 0.3685047276378546,
|
5184 |
+
"grad_norm": 0.40768033266067505,
|
5185 |
+
"learning_rate": 5.275810638639088e-05,
|
5186 |
+
"loss": 0.957,
|
5187 |
+
"step": 734
|
5188 |
+
},
|
5189 |
+
{
|
5190 |
+
"epoch": 0.3690067776755083,
|
5191 |
+
"grad_norm": 0.42073965072631836,
|
5192 |
+
"learning_rate": 5.265212696507387e-05,
|
5193 |
+
"loss": 0.9503,
|
5194 |
+
"step": 735
|
5195 |
+
},
|
5196 |
+
{
|
5197 |
+
"epoch": 0.36950882771316207,
|
5198 |
+
"grad_norm": 0.40175575017929077,
|
5199 |
+
"learning_rate": 5.254613559368649e-05,
|
5200 |
+
"loss": 0.9277,
|
5201 |
+
"step": 736
|
5202 |
+
},
|
5203 |
+
{
|
5204 |
+
"epoch": 0.3700108777508158,
|
5205 |
+
"grad_norm": 0.39959418773651123,
|
5206 |
+
"learning_rate": 5.2440132749809313e-05,
|
5207 |
+
"loss": 0.9021,
|
5208 |
+
"step": 737
|
5209 |
+
},
|
5210 |
+
{
|
5211 |
+
"epoch": 0.37051292778846956,
|
5212 |
+
"grad_norm": 0.45893776416778564,
|
5213 |
+
"learning_rate": 5.2334118911074635e-05,
|
5214 |
+
"loss": 0.9413,
|
5215 |
+
"step": 738
|
5216 |
+
},
|
5217 |
+
{
|
5218 |
+
"epoch": 0.37101497782612336,
|
5219 |
+
"grad_norm": 0.4203508794307709,
|
5220 |
+
"learning_rate": 5.2228094555164265e-05,
|
5221 |
+
"loss": 0.9131,
|
5222 |
+
"step": 739
|
5223 |
+
},
|
5224 |
+
{
|
5225 |
+
"epoch": 0.3715170278637771,
|
5226 |
+
"grad_norm": 0.4097796082496643,
|
5227 |
+
"learning_rate": 5.212206015980742e-05,
|
5228 |
+
"loss": 0.881,
|
5229 |
+
"step": 740
|
5230 |
+
},
|
5231 |
+
{
|
5232 |
+
"epoch": 0.37201907790143085,
|
5233 |
+
"grad_norm": 0.44615375995635986,
|
5234 |
+
"learning_rate": 5.201601620277854e-05,
|
5235 |
+
"loss": 0.8147,
|
5236 |
+
"step": 741
|
5237 |
+
},
|
5238 |
+
{
|
5239 |
+
"epoch": 0.3725211279390846,
|
5240 |
+
"grad_norm": 0.4491327702999115,
|
5241 |
+
"learning_rate": 5.190996316189515e-05,
|
5242 |
+
"loss": 0.8368,
|
5243 |
+
"step": 742
|
5244 |
+
},
|
5245 |
+
{
|
5246 |
+
"epoch": 0.37302317797673834,
|
5247 |
+
"grad_norm": 0.4489690065383911,
|
5248 |
+
"learning_rate": 5.180390151501569e-05,
|
5249 |
+
"loss": 0.9062,
|
5250 |
+
"step": 743
|
5251 |
+
},
|
5252 |
+
{
|
5253 |
+
"epoch": 0.3735252280143921,
|
5254 |
+
"grad_norm": 0.4554278552532196,
|
5255 |
+
"learning_rate": 5.1697831740037436e-05,
|
5256 |
+
"loss": 0.841,
|
5257 |
+
"step": 744
|
5258 |
+
},
|
5259 |
+
{
|
5260 |
+
"epoch": 0.37402727805204583,
|
5261 |
+
"grad_norm": 0.4591432213783264,
|
5262 |
+
"learning_rate": 5.159175431489424e-05,
|
5263 |
+
"loss": 0.8241,
|
5264 |
+
"step": 745
|
5265 |
+
},
|
5266 |
+
{
|
5267 |
+
"epoch": 0.37452932808969963,
|
5268 |
+
"grad_norm": 0.4552235007286072,
|
5269 |
+
"learning_rate": 5.1485669717554396e-05,
|
5270 |
+
"loss": 0.7784,
|
5271 |
+
"step": 746
|
5272 |
+
},
|
5273 |
+
{
|
5274 |
+
"epoch": 0.3750313781273534,
|
5275 |
+
"grad_norm": 0.4900113046169281,
|
5276 |
+
"learning_rate": 5.137957842601856e-05,
|
5277 |
+
"loss": 0.7905,
|
5278 |
+
"step": 747
|
5279 |
+
},
|
5280 |
+
{
|
5281 |
+
"epoch": 0.3755334281650071,
|
5282 |
+
"grad_norm": 0.5452777743339539,
|
5283 |
+
"learning_rate": 5.1273480918317554e-05,
|
5284 |
+
"loss": 0.8248,
|
5285 |
+
"step": 748
|
5286 |
+
},
|
5287 |
+
{
|
5288 |
+
"epoch": 0.37603547820266087,
|
5289 |
+
"grad_norm": 0.5230666399002075,
|
5290 |
+
"learning_rate": 5.116737767251021e-05,
|
5291 |
+
"loss": 0.781,
|
5292 |
+
"step": 749
|
5293 |
+
},
|
5294 |
+
{
|
5295 |
+
"epoch": 0.3765375282403146,
|
5296 |
+
"grad_norm": 0.632352352142334,
|
5297 |
+
"learning_rate": 5.1061269166681183e-05,
|
5298 |
+
"loss": 0.7272,
|
5299 |
+
"step": 750
|
5300 |
+
},
|
5301 |
+
{
|
5302 |
+
"epoch": 0.3765375282403146,
|
5303 |
+
"eval_loss": 0.9616905450820923,
|
5304 |
+
"eval_runtime": 709.9548,
|
5305 |
+
"eval_samples_per_second": 21.265,
|
5306 |
+
"eval_steps_per_second": 2.659,
|
5307 |
+
"step": 750
|
5308 |
}
|
5309 |
],
|
5310 |
"logging_steps": 1,
|
|
|
5333 |
"attributes": {}
|
5334 |
}
|
5335 |
},
|
5336 |
+
"total_flos": 5.207200309633352e+18,
|
5337 |
"train_batch_size": 12,
|
5338 |
"trial_name": null,
|
5339 |
"trial_params": null
|