Training in progress, epoch 5
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:ed8dbf802d5a2c81cf908bbae2367db6319b18403ab8b65d201359e0240a88eb
|
| 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 sha256:35030500b6390522588c3834f0022ba1f0ce92386bd51724a9cf763a4b3b72fe
|
| 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:c999710310a4a0dfb87c713d1d881e2d953ef96e475596f1159b9c3cf20e533a
|
| 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 sha256:9dee06d18ba54b89f17c9a1aa41f4227ce91ff4820faefd664c85b2ba9a292e3
|
| 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,
|
|
@@ -60,11 +60,26 @@
|
|
| 60 |
"eval_samples_per_second": 237.3,
|
| 61 |
"eval_steps_per_second": 3.76,
|
| 62 |
"step": 1720
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
}
|
| 64 |
],
|
| 65 |
"max_steps": 8600,
|
| 66 |
"num_train_epochs": 20,
|
| 67 |
-
"total_flos":
|
| 68 |
"trial_name": null,
|
| 69 |
"trial_params": null
|
| 70 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.508567750453949,
|
| 3 |
"best_model_checkpoint": "bert_final_model/checkpoint-430",
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"global_step": 2150,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
|
|
| 60 |
"eval_samples_per_second": 237.3,
|
| 61 |
"eval_steps_per_second": 3.76,
|
| 62 |
"step": 1720
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 4.65,
|
| 66 |
+
"learning_rate": 7.674418604651164e-06,
|
| 67 |
+
"loss": 0.3066,
|
| 68 |
+
"step": 2000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 5.0,
|
| 72 |
+
"eval_accuracy": 0.7900396151669496,
|
| 73 |
+
"eval_loss": 0.5998191237449646,
|
| 74 |
+
"eval_runtime": 14.9204,
|
| 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": 7767267398282700.0,
|
| 83 |
"trial_name": null,
|
| 84 |
"trial_params": null
|
| 85 |
}
|
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 sha256:35030500b6390522588c3834f0022ba1f0ce92386bd51724a9cf763a4b3b72fe
|
| 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 sha256:1698a6098a4a4e792101aae817dd686e3601a376ab447fd84113ece218c023fa
|
| 3 |
+
size 6436
|