Training in progress, epoch 6
Browse files- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/pytorch_model.bin +1 -1
- last-checkpoint/rng_state.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +18 -3
- pytorch_model.bin +1 -1
- runs/Apr30_07-42-25_943994db1dae/events.out.tfevents.1682840548.943994db1dae.1004.2 +2 -2
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 875994245
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:522bdd86ebd56e3333cdf8f810403663c349a8c2d525890cad99b5627f76250e
|
| 3 |
size 875994245
|
last-checkpoint/pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 438010997
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0d6840f85ddd265707a68fb9bf7067bbd0034b02771a3dbd962f3eb684d9f780
|
| 3 |
size 438010997
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14575
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0cbe35c04904b2049403d241b359ccc73738ea1279ee49a485c25599b2a82200
|
| 3 |
size 14575
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 627
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c8c9aa20f02423f5d8434388216a7a927bd1ffc39067a5a30f8d2a3c4561bae5
|
| 3 |
size 627
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.508567750453949,
|
| 3 |
"best_model_checkpoint": "bert_final_model/checkpoint-430",
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
@@ -75,11 +75,26 @@
|
|
| 75 |
"eval_samples_per_second": 236.857,
|
| 76 |
"eval_steps_per_second": 3.753,
|
| 77 |
"step": 2150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 78 |
}
|
| 79 |
],
|
| 80 |
"max_steps": 8600,
|
| 81 |
"num_train_epochs": 20,
|
| 82 |
-
"total_flos":
|
| 83 |
"trial_name": null,
|
| 84 |
"trial_params": null
|
| 85 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.508567750453949,
|
| 3 |
"best_model_checkpoint": "bert_final_model/checkpoint-430",
|
| 4 |
+
"epoch": 6.0,
|
| 5 |
+
"global_step": 2580,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
|
|
| 75 |
"eval_samples_per_second": 236.857,
|
| 76 |
"eval_steps_per_second": 3.753,
|
| 77 |
"step": 2150
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 5.81,
|
| 81 |
+
"learning_rate": 7.0930232558139545e-06,
|
| 82 |
+
"loss": 0.2395,
|
| 83 |
+
"step": 2500
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 6.0,
|
| 87 |
+
"eval_accuracy": 0.7843803056027164,
|
| 88 |
+
"eval_loss": 0.6628095507621765,
|
| 89 |
+
"eval_runtime": 14.9429,
|
| 90 |
+
"eval_samples_per_second": 236.5,
|
| 91 |
+
"eval_steps_per_second": 3.748,
|
| 92 |
+
"step": 2580
|
| 93 |
}
|
| 94 |
],
|
| 95 |
"max_steps": 8600,
|
| 96 |
"num_train_epochs": 20,
|
| 97 |
+
"total_flos": 9320720877939240.0,
|
| 98 |
"trial_name": null,
|
| 99 |
"trial_params": null
|
| 100 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 438010997
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0d6840f85ddd265707a68fb9bf7067bbd0034b02771a3dbd962f3eb684d9f780
|
| 3 |
size 438010997
|
runs/Apr30_07-42-25_943994db1dae/events.out.tfevents.1682840548.943994db1dae.1004.2
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32fcab748c43a2cddb66d4e59fbe2543ae8be76dd950f04dd89cf7fed0588147
|
| 3 |
+
size 6916
|