End of training
Browse files- all_results.json +14 -0
- eval_results.json +9 -0
- train_results.json +8 -0
- trainer_state.json +61 -0
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.96,
|
| 3 |
+
"eval_accuracy": 0.7140672782874617,
|
| 4 |
+
"eval_loss": 0.7000637054443359,
|
| 5 |
+
"eval_runtime": 10.4472,
|
| 6 |
+
"eval_samples": 3270,
|
| 7 |
+
"eval_samples_per_second": 313.003,
|
| 8 |
+
"eval_steps_per_second": 4.977,
|
| 9 |
+
"train_loss": 0.4717037253183861,
|
| 10 |
+
"train_runtime": 435.6602,
|
| 11 |
+
"train_samples": 9427,
|
| 12 |
+
"train_samples_per_second": 86.554,
|
| 13 |
+
"train_steps_per_second": 0.67
|
| 14 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.96,
|
| 3 |
+
"eval_accuracy": 0.7140672782874617,
|
| 4 |
+
"eval_loss": 0.7000637054443359,
|
| 5 |
+
"eval_runtime": 10.4472,
|
| 6 |
+
"eval_samples": 3270,
|
| 7 |
+
"eval_samples_per_second": 313.003,
|
| 8 |
+
"eval_steps_per_second": 4.977
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.96,
|
| 3 |
+
"train_loss": 0.4717037253183861,
|
| 4 |
+
"train_runtime": 435.6602,
|
| 5 |
+
"train_samples": 9427,
|
| 6 |
+
"train_samples_per_second": 86.554,
|
| 7 |
+
"train_steps_per_second": 0.67
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.959322033898305,
|
| 5 |
+
"global_step": 292,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.99,
|
| 12 |
+
"eval_accuracy": 0.6107033639143731,
|
| 13 |
+
"eval_loss": 0.6423391103744507,
|
| 14 |
+
"eval_runtime": 10.5728,
|
| 15 |
+
"eval_samples_per_second": 309.284,
|
| 16 |
+
"eval_steps_per_second": 4.918,
|
| 17 |
+
"step": 73
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 1.99,
|
| 21 |
+
"eval_accuracy": 0.7097859327217125,
|
| 22 |
+
"eval_loss": 0.5650969743728638,
|
| 23 |
+
"eval_runtime": 10.5393,
|
| 24 |
+
"eval_samples_per_second": 310.267,
|
| 25 |
+
"eval_steps_per_second": 4.934,
|
| 26 |
+
"step": 147
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 3.0,
|
| 30 |
+
"eval_accuracy": 0.7149847094801223,
|
| 31 |
+
"eval_loss": 0.6035652160644531,
|
| 32 |
+
"eval_runtime": 10.5633,
|
| 33 |
+
"eval_samples_per_second": 309.563,
|
| 34 |
+
"eval_steps_per_second": 4.923,
|
| 35 |
+
"step": 221
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"epoch": 3.96,
|
| 39 |
+
"eval_accuracy": 0.7140672782874617,
|
| 40 |
+
"eval_loss": 0.7000637054443359,
|
| 41 |
+
"eval_runtime": 10.5393,
|
| 42 |
+
"eval_samples_per_second": 310.266,
|
| 43 |
+
"eval_steps_per_second": 4.934,
|
| 44 |
+
"step": 292
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 3.96,
|
| 48 |
+
"step": 292,
|
| 49 |
+
"total_flos": 9823777473976320.0,
|
| 50 |
+
"train_loss": 0.4717037253183861,
|
| 51 |
+
"train_runtime": 435.6602,
|
| 52 |
+
"train_samples_per_second": 86.554,
|
| 53 |
+
"train_steps_per_second": 0.67
|
| 54 |
+
}
|
| 55 |
+
],
|
| 56 |
+
"max_steps": 292,
|
| 57 |
+
"num_train_epochs": 4,
|
| 58 |
+
"total_flos": 9823777473976320.0,
|
| 59 |
+
"trial_name": null,
|
| 60 |
+
"trial_params": null
|
| 61 |
+
}
|