End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 0.6194185018539429,
|
| 4 |
+
"eval_runtime": 4.6183,
|
| 5 |
+
"eval_samples": 32,
|
| 6 |
+
"eval_samples_per_second": 6.929,
|
| 7 |
+
"eval_steps_per_second": 0.217,
|
| 8 |
+
"eval_wer": 0.392811296534018,
|
| 9 |
+
"train_loss": 0.4858972082535426,
|
| 10 |
+
"train_runtime": 84.4207,
|
| 11 |
+
"train_samples": 128,
|
| 12 |
+
"train_samples_per_second": 4.549,
|
| 13 |
+
"train_steps_per_second": 0.142
|
| 14 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 0.6194185018539429,
|
| 4 |
+
"eval_runtime": 4.6183,
|
| 5 |
+
"eval_samples": 32,
|
| 6 |
+
"eval_samples_per_second": 6.929,
|
| 7 |
+
"eval_steps_per_second": 0.217,
|
| 8 |
+
"eval_wer": 0.392811296534018
|
| 9 |
+
}
|
runs/Jun01_08-35-19_bc6e99d30a9c/events.out.tfevents.1685609988.bc6e99d30a9c.7397.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:134623d32fddcb67425f36c8bfb0ffff1ca9845edc66d94e676842d1ab74cbb6
|
| 3 |
+
size 352
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"train_loss": 0.4858972082535426,
|
| 4 |
+
"train_runtime": 84.4207,
|
| 5 |
+
"train_samples": 128,
|
| 6 |
+
"train_samples_per_second": 4.549,
|
| 7 |
+
"train_steps_per_second": 0.142
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 12,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 1.0,
|
| 12 |
+
"eval_loss": 0.6447093486785889,
|
| 13 |
+
"eval_runtime": 4.5652,
|
| 14 |
+
"eval_samples_per_second": 7.01,
|
| 15 |
+
"eval_steps_per_second": 0.219,
|
| 16 |
+
"eval_wer": 0.4030808729139923,
|
| 17 |
+
"step": 4
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 2.0,
|
| 21 |
+
"eval_loss": 0.6389366984367371,
|
| 22 |
+
"eval_runtime": 4.147,
|
| 23 |
+
"eval_samples_per_second": 7.716,
|
| 24 |
+
"eval_steps_per_second": 0.241,
|
| 25 |
+
"eval_wer": 0.3992297817715019,
|
| 26 |
+
"step": 8
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 2.5,
|
| 30 |
+
"learning_rate": 9.000000000000001e-07,
|
| 31 |
+
"loss": 0.4891,
|
| 32 |
+
"step": 10
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 3.0,
|
| 36 |
+
"eval_loss": 0.6194185018539429,
|
| 37 |
+
"eval_runtime": 4.1704,
|
| 38 |
+
"eval_samples_per_second": 7.673,
|
| 39 |
+
"eval_steps_per_second": 0.24,
|
| 40 |
+
"eval_wer": 0.392811296534018,
|
| 41 |
+
"step": 12
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 3.0,
|
| 45 |
+
"step": 12,
|
| 46 |
+
"total_flos": 1.1081679372288e+17,
|
| 47 |
+
"train_loss": 0.4858972082535426,
|
| 48 |
+
"train_runtime": 84.4207,
|
| 49 |
+
"train_samples_per_second": 4.549,
|
| 50 |
+
"train_steps_per_second": 0.142
|
| 51 |
+
}
|
| 52 |
+
],
|
| 53 |
+
"max_steps": 12,
|
| 54 |
+
"num_train_epochs": 3,
|
| 55 |
+
"total_flos": 1.1081679372288e+17,
|
| 56 |
+
"trial_name": null,
|
| 57 |
+
"trial_params": null
|
| 58 |
+
}
|