Training in progress, epoch 6, checkpoint
Browse files- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +620 -4
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 78207176
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3dd432db5da8f4c023892dca2807f126bf83de59027dfc1d4a10f342ffdda95d
|
| 3 |
size 78207176
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 40178148
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:05e2f776dc4d831cb32aeef6e02b3bc0156fb70dd6d29462ba03dc27efc51f90
|
| 3 |
size 40178148
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed226574950c635e8c4596429289e17dc17336a36854ca7cfdc0ab58e7860de4
|
| 3 |
size 15984
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d483aff5929017d517922beeffe48cf9cfd0b9ab20c1e5d41ec02204efc567bb
|
| 3 |
size 15984
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b2dc138a9008b46ab343634e564d1c559d53f57ffb3333d53c3eef66571d03f
|
| 3 |
size 15984
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58a645b2ce0616ed60d06cabe300d984ae0c18d86997cf72a44efb98fa92ced9
|
| 3 |
size 15984
|
last-checkpoint/rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6fbcd3f3595bbc5ea5bd431178cd3f1b0831803dc4919831bba50c4c8a45094e
|
| 3 |
size 15984
|
last-checkpoint/rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0d5c16b2554df18b9d35a36260bbe168d7cd2b6d9a912d01d6a206fab0446748
|
| 3 |
size 15984
|
last-checkpoint/rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b73f06fa2fa898def7d477689270dd204662ecac6f9a4adeed837d035cd659c8
|
| 3 |
size 15984
|
last-checkpoint/rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6e011d70987bb6c392832ce34365144ee4fe9c012d18cf12cef6672234b1b9fd
|
| 3 |
size 15984
|
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:b4ac31dcfba988823188c44099086577dd09e162577217892c3f4fd5ad8489c2
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 6.
|
| 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,
|
|
@@ -4291,6 +4291,622 @@
|
|
| 4291 |
"learning_rate": 1.0533112215819298e-05,
|
| 4292 |
"loss": 0.0,
|
| 4293 |
"step": 612
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4294 |
}
|
| 4295 |
],
|
| 4296 |
"logging_steps": 1,
|
|
@@ -4305,12 +4921,12 @@
|
|
| 4305 |
"should_evaluate": false,
|
| 4306 |
"should_log": false,
|
| 4307 |
"should_save": true,
|
| 4308 |
-
"should_training_stop":
|
| 4309 |
},
|
| 4310 |
"attributes": {}
|
| 4311 |
}
|
| 4312 |
},
|
| 4313 |
-
"total_flos":
|
| 4314 |
"train_batch_size": 8,
|
| 4315 |
"trial_name": null,
|
| 4316 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 6.862745098039216,
|
| 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,
|
|
|
|
| 4291 |
"learning_rate": 1.0533112215819298e-05,
|
| 4292 |
"loss": 0.0,
|
| 4293 |
"step": 612
|
| 4294 |
+
},
|
| 4295 |
+
{
|
| 4296 |
+
"epoch": 6.009803921568627,
|
| 4297 |
+
"grad_norm": 0.00028710978222079575,
|
| 4298 |
+
"learning_rate": 1.029921484817783e-05,
|
| 4299 |
+
"loss": 0.0,
|
| 4300 |
+
"step": 613
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 6.019607843137255,
|
| 4304 |
+
"grad_norm": 0.0004612678021658212,
|
| 4305 |
+
"learning_rate": 1.0067804088378455e-05,
|
| 4306 |
+
"loss": 0.0,
|
| 4307 |
+
"step": 614
|
| 4308 |
+
},
|
| 4309 |
+
{
|
| 4310 |
+
"epoch": 6.029411764705882,
|
| 4311 |
+
"grad_norm": 0.00030615812283940613,
|
| 4312 |
+
"learning_rate": 9.8388862806641e-06,
|
| 4313 |
+
"loss": 0.0,
|
| 4314 |
+
"step": 615
|
| 4315 |
+
},
|
| 4316 |
+
{
|
| 4317 |
+
"epoch": 6.03921568627451,
|
| 4318 |
+
"grad_norm": 0.0002808566496241838,
|
| 4319 |
+
"learning_rate": 9.612467700932045e-06,
|
| 4320 |
+
"loss": 0.0,
|
| 4321 |
+
"step": 616
|
| 4322 |
+
},
|
| 4323 |
+
{
|
| 4324 |
+
"epoch": 6.049019607843137,
|
| 4325 |
+
"grad_norm": 0.00028527501854114234,
|
| 4326 |
+
"learning_rate": 9.388554556562049e-06,
|
| 4327 |
+
"loss": 0.0,
|
| 4328 |
+
"step": 617
|
| 4329 |
+
},
|
| 4330 |
+
{
|
| 4331 |
+
"epoch": 6.0588235294117645,
|
| 4332 |
+
"grad_norm": 0.00022610844462178648,
|
| 4333 |
+
"learning_rate": 9.167152986246078e-06,
|
| 4334 |
+
"loss": 0.0,
|
| 4335 |
+
"step": 618
|
| 4336 |
+
},
|
| 4337 |
+
{
|
| 4338 |
+
"epoch": 6.068627450980392,
|
| 4339 |
+
"grad_norm": 0.0007680747075937688,
|
| 4340 |
+
"learning_rate": 8.948269059820025e-06,
|
| 4341 |
+
"loss": 0.0,
|
| 4342 |
+
"step": 619
|
| 4343 |
+
},
|
| 4344 |
+
{
|
| 4345 |
+
"epoch": 6.078431372549019,
|
| 4346 |
+
"grad_norm": 0.000270091142738238,
|
| 4347 |
+
"learning_rate": 8.731908778097302e-06,
|
| 4348 |
+
"loss": 0.0,
|
| 4349 |
+
"step": 620
|
| 4350 |
+
},
|
| 4351 |
+
{
|
| 4352 |
+
"epoch": 6.088235294117647,
|
| 4353 |
+
"grad_norm": 0.00030641083139926195,
|
| 4354 |
+
"learning_rate": 8.518078072704338e-06,
|
| 4355 |
+
"loss": 0.0,
|
| 4356 |
+
"step": 621
|
| 4357 |
+
},
|
| 4358 |
+
{
|
| 4359 |
+
"epoch": 6.098039215686274,
|
| 4360 |
+
"grad_norm": 0.000176057088538073,
|
| 4361 |
+
"learning_rate": 8.306782805917904e-06,
|
| 4362 |
+
"loss": 0.0,
|
| 4363 |
+
"step": 622
|
| 4364 |
+
},
|
| 4365 |
+
{
|
| 4366 |
+
"epoch": 6.107843137254902,
|
| 4367 |
+
"grad_norm": 0.0003552499692887068,
|
| 4368 |
+
"learning_rate": 8.098028770504494e-06,
|
| 4369 |
+
"loss": 0.0,
|
| 4370 |
+
"step": 623
|
| 4371 |
+
},
|
| 4372 |
+
{
|
| 4373 |
+
"epoch": 6.117647058823529,
|
| 4374 |
+
"grad_norm": 0.00022495270241051912,
|
| 4375 |
+
"learning_rate": 7.891821689561459e-06,
|
| 4376 |
+
"loss": 0.0,
|
| 4377 |
+
"step": 624
|
| 4378 |
+
},
|
| 4379 |
+
{
|
| 4380 |
+
"epoch": 6.127450980392156,
|
| 4381 |
+
"grad_norm": 0.00021133887639734894,
|
| 4382 |
+
"learning_rate": 7.68816721636004e-06,
|
| 4383 |
+
"loss": 0.0,
|
| 4384 |
+
"step": 625
|
| 4385 |
+
},
|
| 4386 |
+
{
|
| 4387 |
+
"epoch": 6.137254901960785,
|
| 4388 |
+
"grad_norm": 0.00021480493887793273,
|
| 4389 |
+
"learning_rate": 7.487070934190532e-06,
|
| 4390 |
+
"loss": 0.0,
|
| 4391 |
+
"step": 626
|
| 4392 |
+
},
|
| 4393 |
+
{
|
| 4394 |
+
"epoch": 6.147058823529412,
|
| 4395 |
+
"grad_norm": 0.00021525472402572632,
|
| 4396 |
+
"learning_rate": 7.288538356209092e-06,
|
| 4397 |
+
"loss": 0.0,
|
| 4398 |
+
"step": 627
|
| 4399 |
+
},
|
| 4400 |
+
{
|
| 4401 |
+
"epoch": 6.1568627450980395,
|
| 4402 |
+
"grad_norm": 0.00030373220215551555,
|
| 4403 |
+
"learning_rate": 7.092574925286614e-06,
|
| 4404 |
+
"loss": 0.0,
|
| 4405 |
+
"step": 628
|
| 4406 |
+
},
|
| 4407 |
+
{
|
| 4408 |
+
"epoch": 6.166666666666667,
|
| 4409 |
+
"grad_norm": 0.00036649382673203945,
|
| 4410 |
+
"learning_rate": 6.899186013859561e-06,
|
| 4411 |
+
"loss": 0.0,
|
| 4412 |
+
"step": 629
|
| 4413 |
+
},
|
| 4414 |
+
{
|
| 4415 |
+
"epoch": 6.176470588235294,
|
| 4416 |
+
"grad_norm": 0.0003223404346499592,
|
| 4417 |
+
"learning_rate": 6.708376923782635e-06,
|
| 4418 |
+
"loss": 0.0,
|
| 4419 |
+
"step": 630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 6.186274509803922,
|
| 4423 |
+
"grad_norm": 0.0003951244871132076,
|
| 4424 |
+
"learning_rate": 6.520152886183406e-06,
|
| 4425 |
+
"loss": 0.0,
|
| 4426 |
+
"step": 631
|
| 4427 |
+
},
|
| 4428 |
+
{
|
| 4429 |
+
"epoch": 6.196078431372549,
|
| 4430 |
+
"grad_norm": 0.0002750719431787729,
|
| 4431 |
+
"learning_rate": 6.3345190613189635e-06,
|
| 4432 |
+
"loss": 0.0,
|
| 4433 |
+
"step": 632
|
| 4434 |
+
},
|
| 4435 |
+
{
|
| 4436 |
+
"epoch": 6.205882352941177,
|
| 4437 |
+
"grad_norm": 0.00025263812858611345,
|
| 4438 |
+
"learning_rate": 6.151480538434382e-06,
|
| 4439 |
+
"loss": 0.0,
|
| 4440 |
+
"step": 633
|
| 4441 |
+
},
|
| 4442 |
+
{
|
| 4443 |
+
"epoch": 6.215686274509804,
|
| 4444 |
+
"grad_norm": 0.00031022890470921993,
|
| 4445 |
+
"learning_rate": 5.971042335623229e-06,
|
| 4446 |
+
"loss": 0.0,
|
| 4447 |
+
"step": 634
|
| 4448 |
+
},
|
| 4449 |
+
{
|
| 4450 |
+
"epoch": 6.2254901960784315,
|
| 4451 |
+
"grad_norm": 0.0007206103182397783,
|
| 4452 |
+
"learning_rate": 5.793209399689978e-06,
|
| 4453 |
+
"loss": 0.0,
|
| 4454 |
+
"step": 635
|
| 4455 |
+
},
|
| 4456 |
+
{
|
| 4457 |
+
"epoch": 6.235294117647059,
|
| 4458 |
+
"grad_norm": 0.00041089122532866895,
|
| 4459 |
+
"learning_rate": 5.617986606014419e-06,
|
| 4460 |
+
"loss": 0.0,
|
| 4461 |
+
"step": 636
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 6.245098039215686,
|
| 4465 |
+
"grad_norm": 0.0001943662209669128,
|
| 4466 |
+
"learning_rate": 5.445378758417925e-06,
|
| 4467 |
+
"loss": 0.0,
|
| 4468 |
+
"step": 637
|
| 4469 |
+
},
|
| 4470 |
+
{
|
| 4471 |
+
"epoch": 6.254901960784314,
|
| 4472 |
+
"grad_norm": 0.00022811333474237472,
|
| 4473 |
+
"learning_rate": 5.275390589031859e-06,
|
| 4474 |
+
"loss": 0.0,
|
| 4475 |
+
"step": 638
|
| 4476 |
+
},
|
| 4477 |
+
{
|
| 4478 |
+
"epoch": 6.264705882352941,
|
| 4479 |
+
"grad_norm": 0.00030687302933074534,
|
| 4480 |
+
"learning_rate": 5.108026758167719e-06,
|
| 4481 |
+
"loss": 0.0,
|
| 4482 |
+
"step": 639
|
| 4483 |
+
},
|
| 4484 |
+
{
|
| 4485 |
+
"epoch": 6.2745098039215685,
|
| 4486 |
+
"grad_norm": 0.00025758627452887595,
|
| 4487 |
+
"learning_rate": 4.943291854189493e-06,
|
| 4488 |
+
"loss": 0.0,
|
| 4489 |
+
"step": 640
|
| 4490 |
+
},
|
| 4491 |
+
{
|
| 4492 |
+
"epoch": 6.284313725490196,
|
| 4493 |
+
"grad_norm": 0.00033247433020733297,
|
| 4494 |
+
"learning_rate": 4.781190393387796e-06,
|
| 4495 |
+
"loss": 0.0,
|
| 4496 |
+
"step": 641
|
| 4497 |
+
},
|
| 4498 |
+
{
|
| 4499 |
+
"epoch": 6.294117647058823,
|
| 4500 |
+
"grad_norm": 0.0002540280984248966,
|
| 4501 |
+
"learning_rate": 4.6217268198560404e-06,
|
| 4502 |
+
"loss": 0.0,
|
| 4503 |
+
"step": 642
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 6.303921568627451,
|
| 4507 |
+
"grad_norm": 0.0003001320001203567,
|
| 4508 |
+
"learning_rate": 4.464905505368658e-06,
|
| 4509 |
+
"loss": 0.0,
|
| 4510 |
+
"step": 643
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"epoch": 6.313725490196078,
|
| 4514 |
+
"grad_norm": 0.0005657892324961722,
|
| 4515 |
+
"learning_rate": 4.3107307492612086e-06,
|
| 4516 |
+
"loss": 0.0,
|
| 4517 |
+
"step": 644
|
| 4518 |
+
},
|
| 4519 |
+
{
|
| 4520 |
+
"epoch": 6.323529411764706,
|
| 4521 |
+
"grad_norm": 0.00026279667508788407,
|
| 4522 |
+
"learning_rate": 4.1592067783125015e-06,
|
| 4523 |
+
"loss": 0.0,
|
| 4524 |
+
"step": 645
|
| 4525 |
+
},
|
| 4526 |
+
{
|
| 4527 |
+
"epoch": 6.333333333333333,
|
| 4528 |
+
"grad_norm": 0.00045822173706255853,
|
| 4529 |
+
"learning_rate": 4.010337746628751e-06,
|
| 4530 |
+
"loss": 0.0,
|
| 4531 |
+
"step": 646
|
| 4532 |
+
},
|
| 4533 |
+
{
|
| 4534 |
+
"epoch": 6.3431372549019605,
|
| 4535 |
+
"grad_norm": 0.0007906205719336867,
|
| 4536 |
+
"learning_rate": 3.864127735529656e-06,
|
| 4537 |
+
"loss": 0.0,
|
| 4538 |
+
"step": 647
|
| 4539 |
+
},
|
| 4540 |
+
{
|
| 4541 |
+
"epoch": 6.352941176470588,
|
| 4542 |
+
"grad_norm": 0.00023586759925819933,
|
| 4543 |
+
"learning_rate": 3.7205807534365315e-06,
|
| 4544 |
+
"loss": 0.0,
|
| 4545 |
+
"step": 648
|
| 4546 |
+
},
|
| 4547 |
+
{
|
| 4548 |
+
"epoch": 6.362745098039215,
|
| 4549 |
+
"grad_norm": 0.0002759952622000128,
|
| 4550 |
+
"learning_rate": 3.5797007357623945e-06,
|
| 4551 |
+
"loss": 0.0,
|
| 4552 |
+
"step": 649
|
| 4553 |
+
},
|
| 4554 |
+
{
|
| 4555 |
+
"epoch": 6.372549019607844,
|
| 4556 |
+
"grad_norm": 0.00020269464584998786,
|
| 4557 |
+
"learning_rate": 3.441491544804112e-06,
|
| 4558 |
+
"loss": 0.0,
|
| 4559 |
+
"step": 650
|
| 4560 |
+
},
|
| 4561 |
+
{
|
| 4562 |
+
"epoch": 6.382352941176471,
|
| 4563 |
+
"grad_norm": 0.00032984919380396605,
|
| 4564 |
+
"learning_rate": 3.3059569696364502e-06,
|
| 4565 |
+
"loss": 0.0,
|
| 4566 |
+
"step": 651
|
| 4567 |
+
},
|
| 4568 |
+
{
|
| 4569 |
+
"epoch": 6.392156862745098,
|
| 4570 |
+
"grad_norm": 0.0002803058014251292,
|
| 4571 |
+
"learning_rate": 3.1731007260082616e-06,
|
| 4572 |
+
"loss": 0.0,
|
| 4573 |
+
"step": 652
|
| 4574 |
+
},
|
| 4575 |
+
{
|
| 4576 |
+
"epoch": 6.401960784313726,
|
| 4577 |
+
"grad_norm": 0.0007452854770235717,
|
| 4578 |
+
"learning_rate": 3.0429264562405776e-06,
|
| 4579 |
+
"loss": 0.0,
|
| 4580 |
+
"step": 653
|
| 4581 |
+
},
|
| 4582 |
+
{
|
| 4583 |
+
"epoch": 6.411764705882353,
|
| 4584 |
+
"grad_norm": 0.00031766341999173164,
|
| 4585 |
+
"learning_rate": 2.9154377291267674e-06,
|
| 4586 |
+
"loss": 0.0,
|
| 4587 |
+
"step": 654
|
| 4588 |
+
},
|
| 4589 |
+
{
|
| 4590 |
+
"epoch": 6.421568627450981,
|
| 4591 |
+
"grad_norm": 0.00020729843527078629,
|
| 4592 |
+
"learning_rate": 2.790638039834668e-06,
|
| 4593 |
+
"loss": 0.0,
|
| 4594 |
+
"step": 655
|
| 4595 |
+
},
|
| 4596 |
+
{
|
| 4597 |
+
"epoch": 6.431372549019608,
|
| 4598 |
+
"grad_norm": 0.0007090336875990033,
|
| 4599 |
+
"learning_rate": 2.6685308098108106e-06,
|
| 4600 |
+
"loss": 0.0,
|
| 4601 |
+
"step": 656
|
| 4602 |
+
},
|
| 4603 |
+
{
|
| 4604 |
+
"epoch": 6.4411764705882355,
|
| 4605 |
+
"grad_norm": 0.00028745399322360754,
|
| 4606 |
+
"learning_rate": 2.5491193866866025e-06,
|
| 4607 |
+
"loss": 0.0,
|
| 4608 |
+
"step": 657
|
| 4609 |
+
},
|
| 4610 |
+
{
|
| 4611 |
+
"epoch": 6.450980392156863,
|
| 4612 |
+
"grad_norm": 0.00036435428773984313,
|
| 4613 |
+
"learning_rate": 2.432407044186509e-06,
|
| 4614 |
+
"loss": 0.0,
|
| 4615 |
+
"step": 658
|
| 4616 |
+
},
|
| 4617 |
+
{
|
| 4618 |
+
"epoch": 6.46078431372549,
|
| 4619 |
+
"grad_norm": 0.00039886683225631714,
|
| 4620 |
+
"learning_rate": 2.3183969820383735e-06,
|
| 4621 |
+
"loss": 0.0,
|
| 4622 |
+
"step": 659
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 6.470588235294118,
|
| 4626 |
+
"grad_norm": 0.0003374480293132365,
|
| 4627 |
+
"learning_rate": 2.2070923258856255e-06,
|
| 4628 |
+
"loss": 0.0,
|
| 4629 |
+
"step": 660
|
| 4630 |
+
},
|
| 4631 |
+
{
|
| 4632 |
+
"epoch": 6.480392156862745,
|
| 4633 |
+
"grad_norm": 0.00023650114599149674,
|
| 4634 |
+
"learning_rate": 2.098496127201648e-06,
|
| 4635 |
+
"loss": 0.0,
|
| 4636 |
+
"step": 661
|
| 4637 |
+
},
|
| 4638 |
+
{
|
| 4639 |
+
"epoch": 6.490196078431373,
|
| 4640 |
+
"grad_norm": 0.0002508611942175776,
|
| 4641 |
+
"learning_rate": 1.992611363206103e-06,
|
| 4642 |
+
"loss": 0.0,
|
| 4643 |
+
"step": 662
|
| 4644 |
+
},
|
| 4645 |
+
{
|
| 4646 |
+
"epoch": 6.5,
|
| 4647 |
+
"grad_norm": 0.0002132109657395631,
|
| 4648 |
+
"learning_rate": 1.889440936783242e-06,
|
| 4649 |
+
"loss": 0.0,
|
| 4650 |
+
"step": 663
|
| 4651 |
+
},
|
| 4652 |
+
{
|
| 4653 |
+
"epoch": 6.509803921568627,
|
| 4654 |
+
"grad_norm": 0.0003078288573306054,
|
| 4655 |
+
"learning_rate": 1.7889876764024505e-06,
|
| 4656 |
+
"loss": 0.0,
|
| 4657 |
+
"step": 664
|
| 4658 |
+
},
|
| 4659 |
+
{
|
| 4660 |
+
"epoch": 6.519607843137255,
|
| 4661 |
+
"grad_norm": 0.0003129029937554151,
|
| 4662 |
+
"learning_rate": 1.691254336040595e-06,
|
| 4663 |
+
"loss": 0.0,
|
| 4664 |
+
"step": 665
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 6.529411764705882,
|
| 4668 |
+
"grad_norm": 0.0001735873956931755,
|
| 4669 |
+
"learning_rate": 1.59624359510657e-06,
|
| 4670 |
+
"loss": 0.0,
|
| 4671 |
+
"step": 666
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 6.53921568627451,
|
| 4675 |
+
"grad_norm": 0.00021285780530888587,
|
| 4676 |
+
"learning_rate": 1.5039580583678393e-06,
|
| 4677 |
+
"loss": 0.0,
|
| 4678 |
+
"step": 667
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 6.549019607843137,
|
| 4682 |
+
"grad_norm": 0.00029559744871221483,
|
| 4683 |
+
"learning_rate": 1.414400255879008e-06,
|
| 4684 |
+
"loss": 0.0,
|
| 4685 |
+
"step": 668
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 6.5588235294117645,
|
| 4689 |
+
"grad_norm": 0.00028173308237455785,
|
| 4690 |
+
"learning_rate": 1.327572642912468e-06,
|
| 4691 |
+
"loss": 0.0,
|
| 4692 |
+
"step": 669
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 6.568627450980392,
|
| 4696 |
+
"grad_norm": 0.00024257514451164752,
|
| 4697 |
+
"learning_rate": 1.2434775998910964e-06,
|
| 4698 |
+
"loss": 0.0,
|
| 4699 |
+
"step": 670
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 6.578431372549019,
|
| 4703 |
+
"grad_norm": 0.0003676264896057546,
|
| 4704 |
+
"learning_rate": 1.1621174323229612e-06,
|
| 4705 |
+
"loss": 0.0,
|
| 4706 |
+
"step": 671
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 6.588235294117647,
|
| 4710 |
+
"grad_norm": 0.00032732580439187586,
|
| 4711 |
+
"learning_rate": 1.0834943707381784e-06,
|
| 4712 |
+
"loss": 0.0,
|
| 4713 |
+
"step": 672
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 6.598039215686274,
|
| 4717 |
+
"grad_norm": 0.00023518149100709707,
|
| 4718 |
+
"learning_rate": 1.0076105706276888e-06,
|
| 4719 |
+
"loss": 0.0,
|
| 4720 |
+
"step": 673
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 6.607843137254902,
|
| 4724 |
+
"grad_norm": 0.00025342023582197726,
|
| 4725 |
+
"learning_rate": 9.344681123841967e-07,
|
| 4726 |
+
"loss": 0.0,
|
| 4727 |
+
"step": 674
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 6.617647058823529,
|
| 4731 |
+
"grad_norm": 0.0003005929756909609,
|
| 4732 |
+
"learning_rate": 8.640690012451515e-07,
|
| 4733 |
+
"loss": 0.0,
|
| 4734 |
+
"step": 675
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 6.627450980392156,
|
| 4738 |
+
"grad_norm": 0.00025066285161301494,
|
| 4739 |
+
"learning_rate": 7.964151672377458e-07,
|
| 4740 |
+
"loss": 0.0,
|
| 4741 |
+
"step": 676
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 6.637254901960784,
|
| 4745 |
+
"grad_norm": 0.00024955839035101235,
|
| 4746 |
+
"learning_rate": 7.315084651260009e-07,
|
| 4747 |
+
"loss": 0.0,
|
| 4748 |
+
"step": 677
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 6.647058823529412,
|
| 4752 |
+
"grad_norm": 0.000446314865257591,
|
| 4753 |
+
"learning_rate": 6.69350674359959e-07,
|
| 4754 |
+
"loss": 0.0,
|
| 4755 |
+
"step": 678
|
| 4756 |
+
},
|
| 4757 |
+
{
|
| 4758 |
+
"epoch": 6.6568627450980395,
|
| 4759 |
+
"grad_norm": 0.0002682288468349725,
|
| 4760 |
+
"learning_rate": 6.099434990268609e-07,
|
| 4761 |
+
"loss": 0.0,
|
| 4762 |
+
"step": 679
|
| 4763 |
+
},
|
| 4764 |
+
{
|
| 4765 |
+
"epoch": 6.666666666666667,
|
| 4766 |
+
"grad_norm": 0.006764828227460384,
|
| 4767 |
+
"learning_rate": 5.532885678043977e-07,
|
| 4768 |
+
"loss": 0.0,
|
| 4769 |
+
"step": 680
|
| 4770 |
+
},
|
| 4771 |
+
{
|
| 4772 |
+
"epoch": 6.676470588235294,
|
| 4773 |
+
"grad_norm": 0.0003307235019747168,
|
| 4774 |
+
"learning_rate": 4.9938743391615e-07,
|
| 4775 |
+
"loss": 0.0,
|
| 4776 |
+
"step": 681
|
| 4777 |
+
},
|
| 4778 |
+
{
|
| 4779 |
+
"epoch": 6.686274509803922,
|
| 4780 |
+
"grad_norm": 0.00019659138342831284,
|
| 4781 |
+
"learning_rate": 4.482415750889204e-07,
|
| 4782 |
+
"loss": 0.0,
|
| 4783 |
+
"step": 682
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 6.696078431372549,
|
| 4787 |
+
"grad_norm": 0.000775125517975539,
|
| 4788 |
+
"learning_rate": 3.998523935122772e-07,
|
| 4789 |
+
"loss": 0.0,
|
| 4790 |
+
"step": 683
|
| 4791 |
+
},
|
| 4792 |
+
{
|
| 4793 |
+
"epoch": 6.705882352941177,
|
| 4794 |
+
"grad_norm": 0.00031790934735909104,
|
| 4795 |
+
"learning_rate": 3.5422121580005864e-07,
|
| 4796 |
+
"loss": 0.0,
|
| 4797 |
+
"step": 684
|
| 4798 |
+
},
|
| 4799 |
+
{
|
| 4800 |
+
"epoch": 6.715686274509804,
|
| 4801 |
+
"grad_norm": 0.00035337384906597435,
|
| 4802 |
+
"learning_rate": 3.1134929295407564e-07,
|
| 4803 |
+
"loss": 0.0,
|
| 4804 |
+
"step": 685
|
| 4805 |
+
},
|
| 4806 |
+
{
|
| 4807 |
+
"epoch": 6.7254901960784315,
|
| 4808 |
+
"grad_norm": 0.00026295497082173824,
|
| 4809 |
+
"learning_rate": 2.7123780032973235e-07,
|
| 4810 |
+
"loss": 0.0,
|
| 4811 |
+
"step": 686
|
| 4812 |
+
},
|
| 4813 |
+
{
|
| 4814 |
+
"epoch": 6.735294117647059,
|
| 4815 |
+
"grad_norm": 0.0004518682253547013,
|
| 4816 |
+
"learning_rate": 2.3388783760386601e-07,
|
| 4817 |
+
"loss": 0.0,
|
| 4818 |
+
"step": 687
|
| 4819 |
+
},
|
| 4820 |
+
{
|
| 4821 |
+
"epoch": 6.745098039215686,
|
| 4822 |
+
"grad_norm": 0.00031162338564172387,
|
| 4823 |
+
"learning_rate": 1.9930042874457254e-07,
|
| 4824 |
+
"loss": 0.0,
|
| 4825 |
+
"step": 688
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 6.754901960784314,
|
| 4829 |
+
"grad_norm": 0.000260617962339893,
|
| 4830 |
+
"learning_rate": 1.6747652198313957e-07,
|
| 4831 |
+
"loss": 0.0,
|
| 4832 |
+
"step": 689
|
| 4833 |
+
},
|
| 4834 |
+
{
|
| 4835 |
+
"epoch": 6.764705882352941,
|
| 4836 |
+
"grad_norm": 0.00021623096836265177,
|
| 4837 |
+
"learning_rate": 1.3841698978804285e-07,
|
| 4838 |
+
"loss": 0.0,
|
| 4839 |
+
"step": 690
|
| 4840 |
+
},
|
| 4841 |
+
{
|
| 4842 |
+
"epoch": 6.7745098039215685,
|
| 4843 |
+
"grad_norm": 0.00037732068449258804,
|
| 4844 |
+
"learning_rate": 1.1212262884103974e-07,
|
| 4845 |
+
"loss": 0.0,
|
| 4846 |
+
"step": 691
|
| 4847 |
+
},
|
| 4848 |
+
{
|
| 4849 |
+
"epoch": 6.784313725490196,
|
| 4850 |
+
"grad_norm": 0.000314529548631981,
|
| 4851 |
+
"learning_rate": 8.85941600153033e-08,
|
| 4852 |
+
"loss": 0.0,
|
| 4853 |
+
"step": 692
|
| 4854 |
+
},
|
| 4855 |
+
{
|
| 4856 |
+
"epoch": 6.794117647058823,
|
| 4857 |
+
"grad_norm": 0.0002145230391761288,
|
| 4858 |
+
"learning_rate": 6.783222835572055e-08,
|
| 4859 |
+
"loss": 0.0,
|
| 4860 |
+
"step": 693
|
| 4861 |
+
},
|
| 4862 |
+
{
|
| 4863 |
+
"epoch": 6.803921568627451,
|
| 4864 |
+
"grad_norm": 0.00027028913609683514,
|
| 4865 |
+
"learning_rate": 4.98374030611084e-08,
|
| 4866 |
+
"loss": 0.0,
|
| 4867 |
+
"step": 694
|
| 4868 |
+
},
|
| 4869 |
+
{
|
| 4870 |
+
"epoch": 6.813725490196078,
|
| 4871 |
+
"grad_norm": 0.0003521139733493328,
|
| 4872 |
+
"learning_rate": 3.461017746871675e-08,
|
| 4873 |
+
"loss": 0.0,
|
| 4874 |
+
"step": 695
|
| 4875 |
+
},
|
| 4876 |
+
{
|
| 4877 |
+
"epoch": 6.823529411764706,
|
| 4878 |
+
"grad_norm": 0.00044955802150070667,
|
| 4879 |
+
"learning_rate": 2.215096904060454e-08,
|
| 4880 |
+
"loss": 0.0,
|
| 4881 |
+
"step": 696
|
| 4882 |
+
},
|
| 4883 |
+
{
|
| 4884 |
+
"epoch": 6.833333333333333,
|
| 4885 |
+
"grad_norm": 0.00033390807220712304,
|
| 4886 |
+
"learning_rate": 1.246011935228064e-08,
|
| 4887 |
+
"loss": 0.0,
|
| 4888 |
+
"step": 697
|
| 4889 |
+
},
|
| 4890 |
+
{
|
| 4891 |
+
"epoch": 6.8431372549019605,
|
| 4892 |
+
"grad_norm": 0.00022902774799149483,
|
| 4893 |
+
"learning_rate": 5.537894083273543e-09,
|
| 4894 |
+
"loss": 0.0,
|
| 4895 |
+
"step": 698
|
| 4896 |
+
},
|
| 4897 |
+
{
|
| 4898 |
+
"epoch": 6.852941176470588,
|
| 4899 |
+
"grad_norm": 0.00023710480309091508,
|
| 4900 |
+
"learning_rate": 1.384483009898796e-09,
|
| 4901 |
+
"loss": 0.0,
|
| 4902 |
+
"step": 699
|
| 4903 |
+
},
|
| 4904 |
+
{
|
| 4905 |
+
"epoch": 6.862745098039216,
|
| 4906 |
+
"grad_norm": 0.0005341597134247422,
|
| 4907 |
+
"learning_rate": 0.0,
|
| 4908 |
+
"loss": 0.0,
|
| 4909 |
+
"step": 700
|
| 4910 |
}
|
| 4911 |
],
|
| 4912 |
"logging_steps": 1,
|
|
|
|
| 4921 |
"should_evaluate": false,
|
| 4922 |
"should_log": false,
|
| 4923 |
"should_save": true,
|
| 4924 |
+
"should_training_stop": true
|
| 4925 |
},
|
| 4926 |
"attributes": {}
|
| 4927 |
}
|
| 4928 |
},
|
| 4929 |
+
"total_flos": 7.629560224940032e+16,
|
| 4930 |
"train_batch_size": 8,
|
| 4931 |
"trial_name": null,
|
| 4932 |
"trial_params": null
|