Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +34 -0
trainer_state.json
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 3.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 60,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 2.5,
|
13 |
+
"learning_rate": 3.3333333333333333e-06,
|
14 |
+
"loss": 0.7025,
|
15 |
+
"step": 50
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 3.0,
|
19 |
+
"step": 60,
|
20 |
+
"total_flos": 96150646293120.0,
|
21 |
+
"train_loss": 0.7014555931091309,
|
22 |
+
"train_runtime": 9.5902,
|
23 |
+
"train_samples_per_second": 198.641,
|
24 |
+
"train_steps_per_second": 6.256
|
25 |
+
}
|
26 |
+
],
|
27 |
+
"logging_steps": 50,
|
28 |
+
"max_steps": 60,
|
29 |
+
"num_train_epochs": 3,
|
30 |
+
"save_steps": 500,
|
31 |
+
"total_flos": 96150646293120.0,
|
32 |
+
"trial_name": null,
|
33 |
+
"trial_params": null
|
34 |
+
}
|