Model save
Browse files- all_results.json +9 -0
- generation_config.json +14 -0
- train_results.json +9 -0
- trainer_state.json +940 -0
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.03677848608291457,
|
| 5 |
+
"train_runtime": 1919.659,
|
| 6 |
+
"train_samples": 4332,
|
| 7 |
+
"train_samples_per_second": 2.257,
|
| 8 |
+
"train_steps_per_second": 0.282
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
151645,
|
| 6 |
+
151643
|
| 7 |
+
],
|
| 8 |
+
"pad_token_id": 151643,
|
| 9 |
+
"repetition_penalty": 1.05,
|
| 10 |
+
"temperature": 0.7,
|
| 11 |
+
"top_k": 20,
|
| 12 |
+
"top_p": 0.8,
|
| 13 |
+
"transformers_version": "4.53.3"
|
| 14 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.03677848608291457,
|
| 5 |
+
"train_runtime": 1919.659,
|
| 6 |
+
"train_samples": 4332,
|
| 7 |
+
"train_samples_per_second": 2.257,
|
| 8 |
+
"train_steps_per_second": 0.282
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,940 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 542,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0018450184501845018,
|
| 14 |
+
"grad_norm": 174.53293244150143,
|
| 15 |
+
"learning_rate": 0.0,
|
| 16 |
+
"logits/chosen": -0.28515625,
|
| 17 |
+
"logits/rejected": -0.365234375,
|
| 18 |
+
"logps/chosen": -556.0,
|
| 19 |
+
"logps/rejected": -448.0,
|
| 20 |
+
"loss": 0.6914,
|
| 21 |
+
"nll_loss": 1.453125,
|
| 22 |
+
"rewards/accuracies": 0.0,
|
| 23 |
+
"rewards/chosen": 0.0,
|
| 24 |
+
"rewards/margins": 0.0,
|
| 25 |
+
"rewards/rejected": 0.0,
|
| 26 |
+
"step": 1
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.01845018450184502,
|
| 30 |
+
"grad_norm": 346.4662455777756,
|
| 31 |
+
"learning_rate": 8.181818181818182e-08,
|
| 32 |
+
"logits/chosen": -0.9115939736366272,
|
| 33 |
+
"logits/rejected": -0.8186306357383728,
|
| 34 |
+
"logps/chosen": -508.8888854980469,
|
| 35 |
+
"logps/rejected": -512.2222290039062,
|
| 36 |
+
"loss": 0.6986,
|
| 37 |
+
"nll_loss": 1.6076388359069824,
|
| 38 |
+
"rewards/accuracies": 0.2638888955116272,
|
| 39 |
+
"rewards/chosen": 0.0194566510617733,
|
| 40 |
+
"rewards/margins": 0.011149088852107525,
|
| 41 |
+
"rewards/rejected": 0.0083482526242733,
|
| 42 |
+
"step": 10
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 0.03690036900369004,
|
| 46 |
+
"grad_norm": 136.41193032953723,
|
| 47 |
+
"learning_rate": 1.7272727272727272e-07,
|
| 48 |
+
"logits/chosen": -0.9195312261581421,
|
| 49 |
+
"logits/rejected": -0.21855469048023224,
|
| 50 |
+
"logps/chosen": -479.0,
|
| 51 |
+
"logps/rejected": -481.6000061035156,
|
| 52 |
+
"loss": 0.5387,
|
| 53 |
+
"nll_loss": 1.8468749523162842,
|
| 54 |
+
"rewards/accuracies": 0.737500011920929,
|
| 55 |
+
"rewards/chosen": 0.36181640625,
|
| 56 |
+
"rewards/margins": 0.3769287168979645,
|
| 57 |
+
"rewards/rejected": -0.015087890438735485,
|
| 58 |
+
"step": 20
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.055350553505535055,
|
| 62 |
+
"grad_norm": 130.6864319952881,
|
| 63 |
+
"learning_rate": 2.636363636363636e-07,
|
| 64 |
+
"logits/chosen": -0.724609375,
|
| 65 |
+
"logits/rejected": -0.31591796875,
|
| 66 |
+
"logps/chosen": -523.0,
|
| 67 |
+
"logps/rejected": -518.0,
|
| 68 |
+
"loss": 0.259,
|
| 69 |
+
"nll_loss": 1.5070312023162842,
|
| 70 |
+
"rewards/accuracies": 0.949999988079071,
|
| 71 |
+
"rewards/chosen": 1.501562476158142,
|
| 72 |
+
"rewards/margins": 1.534765601158142,
|
| 73 |
+
"rewards/rejected": -0.03243408352136612,
|
| 74 |
+
"step": 30
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.07380073800738007,
|
| 78 |
+
"grad_norm": 9.172076329512516,
|
| 79 |
+
"learning_rate": 3.545454545454545e-07,
|
| 80 |
+
"logits/chosen": -0.8726562261581421,
|
| 81 |
+
"logits/rejected": -0.361572265625,
|
| 82 |
+
"logps/chosen": -484.0,
|
| 83 |
+
"logps/rejected": -476.79998779296875,
|
| 84 |
+
"loss": 0.0942,
|
| 85 |
+
"nll_loss": 1.7843749523162842,
|
| 86 |
+
"rewards/accuracies": 1.0,
|
| 87 |
+
"rewards/chosen": 2.7953124046325684,
|
| 88 |
+
"rewards/margins": 3.5882811546325684,
|
| 89 |
+
"rewards/rejected": -0.788769543170929,
|
| 90 |
+
"step": 40
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.09225092250922509,
|
| 94 |
+
"grad_norm": 133.45013172512645,
|
| 95 |
+
"learning_rate": 4.4545454545454544e-07,
|
| 96 |
+
"logits/chosen": -0.667187511920929,
|
| 97 |
+
"logits/rejected": -0.2557128965854645,
|
| 98 |
+
"logps/chosen": -466.1000061035156,
|
| 99 |
+
"logps/rejected": -514.0,
|
| 100 |
+
"loss": 0.0757,
|
| 101 |
+
"nll_loss": 1.747656226158142,
|
| 102 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 103 |
+
"rewards/chosen": 3.504687547683716,
|
| 104 |
+
"rewards/margins": 5.074999809265137,
|
| 105 |
+
"rewards/rejected": -1.5654296875,
|
| 106 |
+
"step": 50
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 0.11070110701107011,
|
| 110 |
+
"grad_norm": 27.938955497888248,
|
| 111 |
+
"learning_rate": 4.958932238193018e-07,
|
| 112 |
+
"logits/chosen": -0.443359375,
|
| 113 |
+
"logits/rejected": -0.11607666313648224,
|
| 114 |
+
"logps/chosen": -424.1000061035156,
|
| 115 |
+
"logps/rejected": -622.0,
|
| 116 |
+
"loss": 0.0719,
|
| 117 |
+
"nll_loss": 1.6359374523162842,
|
| 118 |
+
"rewards/accuracies": 0.987500011920929,
|
| 119 |
+
"rewards/chosen": 4.342187404632568,
|
| 120 |
+
"rewards/margins": 7.287499904632568,
|
| 121 |
+
"rewards/rejected": -2.9453125,
|
| 122 |
+
"step": 60
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.12915129151291513,
|
| 126 |
+
"grad_norm": 2.667667834720549,
|
| 127 |
+
"learning_rate": 4.856262833675565e-07,
|
| 128 |
+
"logits/chosen": -0.34898680448532104,
|
| 129 |
+
"logits/rejected": -0.0115966796875,
|
| 130 |
+
"logps/chosen": -544.7999877929688,
|
| 131 |
+
"logps/rejected": -636.4000244140625,
|
| 132 |
+
"loss": 0.0561,
|
| 133 |
+
"nll_loss": 1.5499999523162842,
|
| 134 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 135 |
+
"rewards/chosen": 4.755078315734863,
|
| 136 |
+
"rewards/margins": 9.284375190734863,
|
| 137 |
+
"rewards/rejected": -4.529687404632568,
|
| 138 |
+
"step": 70
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 0.14760147601476015,
|
| 142 |
+
"grad_norm": 49.19225093797519,
|
| 143 |
+
"learning_rate": 4.753593429158111e-07,
|
| 144 |
+
"logits/chosen": -0.33574217557907104,
|
| 145 |
+
"logits/rejected": -0.006542968563735485,
|
| 146 |
+
"logps/chosen": -437.6000061035156,
|
| 147 |
+
"logps/rejected": -614.2000122070312,
|
| 148 |
+
"loss": 0.0096,
|
| 149 |
+
"nll_loss": 1.533593773841858,
|
| 150 |
+
"rewards/accuracies": 1.0,
|
| 151 |
+
"rewards/chosen": 4.714062690734863,
|
| 152 |
+
"rewards/margins": 9.884374618530273,
|
| 153 |
+
"rewards/rejected": -5.159375190734863,
|
| 154 |
+
"step": 80
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 0.16605166051660517,
|
| 158 |
+
"grad_norm": 7.869193854585077,
|
| 159 |
+
"learning_rate": 4.6509240246406566e-07,
|
| 160 |
+
"logits/chosen": -0.44111329317092896,
|
| 161 |
+
"logits/rejected": -0.08427734673023224,
|
| 162 |
+
"logps/chosen": -421.3999938964844,
|
| 163 |
+
"logps/rejected": -608.4000244140625,
|
| 164 |
+
"loss": 0.0053,
|
| 165 |
+
"nll_loss": 1.591406226158142,
|
| 166 |
+
"rewards/accuracies": 1.0,
|
| 167 |
+
"rewards/chosen": 4.628125190734863,
|
| 168 |
+
"rewards/margins": 11.384374618530273,
|
| 169 |
+
"rewards/rejected": -6.768750190734863,
|
| 170 |
+
"step": 90
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.18450184501845018,
|
| 174 |
+
"grad_norm": 0.07766290165816067,
|
| 175 |
+
"learning_rate": 4.548254620123203e-07,
|
| 176 |
+
"logits/chosen": -0.3090454041957855,
|
| 177 |
+
"logits/rejected": -0.009716796688735485,
|
| 178 |
+
"logps/chosen": -431.3999938964844,
|
| 179 |
+
"logps/rejected": -543.4000244140625,
|
| 180 |
+
"loss": 0.004,
|
| 181 |
+
"nll_loss": 1.316796898841858,
|
| 182 |
+
"rewards/accuracies": 1.0,
|
| 183 |
+
"rewards/chosen": 6.25,
|
| 184 |
+
"rewards/margins": 12.481249809265137,
|
| 185 |
+
"rewards/rejected": -6.221875190734863,
|
| 186 |
+
"step": 100
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.2029520295202952,
|
| 190 |
+
"grad_norm": 2.219444743672593,
|
| 191 |
+
"learning_rate": 4.445585215605749e-07,
|
| 192 |
+
"logits/chosen": -0.2965331971645355,
|
| 193 |
+
"logits/rejected": -0.10786132514476776,
|
| 194 |
+
"logps/chosen": -410.0,
|
| 195 |
+
"logps/rejected": -554.2000122070312,
|
| 196 |
+
"loss": 0.0035,
|
| 197 |
+
"nll_loss": 1.5421874523162842,
|
| 198 |
+
"rewards/accuracies": 1.0,
|
| 199 |
+
"rewards/chosen": 4.670312404632568,
|
| 200 |
+
"rewards/margins": 12.443750381469727,
|
| 201 |
+
"rewards/rejected": -7.778124809265137,
|
| 202 |
+
"step": 110
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.22140221402214022,
|
| 206 |
+
"grad_norm": 193.37089815524854,
|
| 207 |
+
"learning_rate": 4.3429158110882957e-07,
|
| 208 |
+
"logits/chosen": -0.48212891817092896,
|
| 209 |
+
"logits/rejected": 0.2857910096645355,
|
| 210 |
+
"logps/chosen": -447.6000061035156,
|
| 211 |
+
"logps/rejected": -562.7999877929688,
|
| 212 |
+
"loss": 0.0252,
|
| 213 |
+
"nll_loss": 1.564062476158142,
|
| 214 |
+
"rewards/accuracies": 0.987500011920929,
|
| 215 |
+
"rewards/chosen": 5.265625,
|
| 216 |
+
"rewards/margins": 14.199999809265137,
|
| 217 |
+
"rewards/rejected": -8.940625190734863,
|
| 218 |
+
"step": 120
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.23985239852398524,
|
| 222 |
+
"grad_norm": 3.835693063482046,
|
| 223 |
+
"learning_rate": 4.2402464065708417e-07,
|
| 224 |
+
"logits/chosen": -0.23652343451976776,
|
| 225 |
+
"logits/rejected": 0.23701171576976776,
|
| 226 |
+
"logps/chosen": -455.3999938964844,
|
| 227 |
+
"logps/rejected": -596.7999877929688,
|
| 228 |
+
"loss": 0.0055,
|
| 229 |
+
"nll_loss": 1.6085937023162842,
|
| 230 |
+
"rewards/accuracies": 1.0,
|
| 231 |
+
"rewards/chosen": 5.831250190734863,
|
| 232 |
+
"rewards/margins": 15.287500381469727,
|
| 233 |
+
"rewards/rejected": -9.482812881469727,
|
| 234 |
+
"step": 130
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.25830258302583026,
|
| 238 |
+
"grad_norm": 0.24648110266264378,
|
| 239 |
+
"learning_rate": 4.137577002053388e-07,
|
| 240 |
+
"logits/chosen": -0.25221556425094604,
|
| 241 |
+
"logits/rejected": 0.06816406548023224,
|
| 242 |
+
"logps/chosen": -432.79998779296875,
|
| 243 |
+
"logps/rejected": -649.7999877929688,
|
| 244 |
+
"loss": 0.0079,
|
| 245 |
+
"nll_loss": 1.424218773841858,
|
| 246 |
+
"rewards/accuracies": 1.0,
|
| 247 |
+
"rewards/chosen": 7.181250095367432,
|
| 248 |
+
"rewards/margins": 15.356249809265137,
|
| 249 |
+
"rewards/rejected": -8.184374809265137,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.2767527675276753,
|
| 254 |
+
"grad_norm": 0.03622304658168317,
|
| 255 |
+
"learning_rate": 4.0349075975359343e-07,
|
| 256 |
+
"logits/chosen": -0.32623291015625,
|
| 257 |
+
"logits/rejected": 0.09409179538488388,
|
| 258 |
+
"logps/chosen": -373.79998779296875,
|
| 259 |
+
"logps/rejected": -513.2000122070312,
|
| 260 |
+
"loss": 0.0027,
|
| 261 |
+
"nll_loss": 1.293359398841858,
|
| 262 |
+
"rewards/accuracies": 1.0,
|
| 263 |
+
"rewards/chosen": 6.625,
|
| 264 |
+
"rewards/margins": 13.981249809265137,
|
| 265 |
+
"rewards/rejected": -7.340624809265137,
|
| 266 |
+
"step": 150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.2952029520295203,
|
| 270 |
+
"grad_norm": 0.020076000811240717,
|
| 271 |
+
"learning_rate": 3.932238193018481e-07,
|
| 272 |
+
"logits/chosen": -0.31867676973342896,
|
| 273 |
+
"logits/rejected": 0.16098633408546448,
|
| 274 |
+
"logps/chosen": -477.0,
|
| 275 |
+
"logps/rejected": -640.4000244140625,
|
| 276 |
+
"loss": 0.0022,
|
| 277 |
+
"nll_loss": 1.346093773841858,
|
| 278 |
+
"rewards/accuracies": 1.0,
|
| 279 |
+
"rewards/chosen": 7.65625,
|
| 280 |
+
"rewards/margins": 16.581249237060547,
|
| 281 |
+
"rewards/rejected": -8.909375190734863,
|
| 282 |
+
"step": 160
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.31365313653136534,
|
| 286 |
+
"grad_norm": 0.041859392866214304,
|
| 287 |
+
"learning_rate": 3.8295687885010263e-07,
|
| 288 |
+
"logits/chosen": -0.4400878846645355,
|
| 289 |
+
"logits/rejected": 0.0843505859375,
|
| 290 |
+
"logps/chosen": -390.6000061035156,
|
| 291 |
+
"logps/rejected": -612.7999877929688,
|
| 292 |
+
"loss": 0.0027,
|
| 293 |
+
"nll_loss": 1.485937476158142,
|
| 294 |
+
"rewards/accuracies": 1.0,
|
| 295 |
+
"rewards/chosen": 7.912499904632568,
|
| 296 |
+
"rewards/margins": 16.293750762939453,
|
| 297 |
+
"rewards/rejected": -8.393750190734863,
|
| 298 |
+
"step": 170
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 0.33210332103321033,
|
| 302 |
+
"grad_norm": 0.05035591243969461,
|
| 303 |
+
"learning_rate": 3.726899383983573e-07,
|
| 304 |
+
"logits/chosen": -0.51611328125,
|
| 305 |
+
"logits/rejected": 0.005908203311264515,
|
| 306 |
+
"logps/chosen": -419.79998779296875,
|
| 307 |
+
"logps/rejected": -572.2000122070312,
|
| 308 |
+
"loss": 0.0016,
|
| 309 |
+
"nll_loss": 1.400781273841858,
|
| 310 |
+
"rewards/accuracies": 1.0,
|
| 311 |
+
"rewards/chosen": 8.496874809265137,
|
| 312 |
+
"rewards/margins": 16.387500762939453,
|
| 313 |
+
"rewards/rejected": -7.909375190734863,
|
| 314 |
+
"step": 180
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.3505535055350554,
|
| 318 |
+
"grad_norm": 0.10815032342544012,
|
| 319 |
+
"learning_rate": 3.624229979466119e-07,
|
| 320 |
+
"logits/chosen": -0.37841796875,
|
| 321 |
+
"logits/rejected": 0.03369140625,
|
| 322 |
+
"logps/chosen": -444.3999938964844,
|
| 323 |
+
"logps/rejected": -575.0,
|
| 324 |
+
"loss": 0.0024,
|
| 325 |
+
"nll_loss": 1.273828148841858,
|
| 326 |
+
"rewards/accuracies": 1.0,
|
| 327 |
+
"rewards/chosen": 9.0625,
|
| 328 |
+
"rewards/margins": 14.831250190734863,
|
| 329 |
+
"rewards/rejected": -5.7578125,
|
| 330 |
+
"step": 190
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 0.36900369003690037,
|
| 334 |
+
"grad_norm": 0.04004367261026864,
|
| 335 |
+
"learning_rate": 3.521560574948665e-07,
|
| 336 |
+
"logits/chosen": -0.40791016817092896,
|
| 337 |
+
"logits/rejected": -0.10537109524011612,
|
| 338 |
+
"logps/chosen": -368.20001220703125,
|
| 339 |
+
"logps/rejected": -501.79998779296875,
|
| 340 |
+
"loss": 0.0018,
|
| 341 |
+
"nll_loss": 1.365625023841858,
|
| 342 |
+
"rewards/accuracies": 1.0,
|
| 343 |
+
"rewards/chosen": 8.859375,
|
| 344 |
+
"rewards/margins": 15.050000190734863,
|
| 345 |
+
"rewards/rejected": -6.193749904632568,
|
| 346 |
+
"step": 200
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.3874538745387454,
|
| 350 |
+
"grad_norm": 0.008541615515180453,
|
| 351 |
+
"learning_rate": 3.418891170431211e-07,
|
| 352 |
+
"logits/chosen": -0.30952149629592896,
|
| 353 |
+
"logits/rejected": 0.10322265326976776,
|
| 354 |
+
"logps/chosen": -470.0,
|
| 355 |
+
"logps/rejected": -550.7999877929688,
|
| 356 |
+
"loss": 0.0029,
|
| 357 |
+
"nll_loss": 1.355078101158142,
|
| 358 |
+
"rewards/accuracies": 1.0,
|
| 359 |
+
"rewards/chosen": 10.175000190734863,
|
| 360 |
+
"rewards/margins": 16.53125,
|
| 361 |
+
"rewards/rejected": -6.373437404632568,
|
| 362 |
+
"step": 210
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.4059040590405904,
|
| 366 |
+
"grad_norm": 0.02023002394682239,
|
| 367 |
+
"learning_rate": 3.3162217659137575e-07,
|
| 368 |
+
"logits/chosen": -0.2582031190395355,
|
| 369 |
+
"logits/rejected": 0.18632812798023224,
|
| 370 |
+
"logps/chosen": -386.3999938964844,
|
| 371 |
+
"logps/rejected": -556.5999755859375,
|
| 372 |
+
"loss": 0.0031,
|
| 373 |
+
"nll_loss": 1.3464844226837158,
|
| 374 |
+
"rewards/accuracies": 1.0,
|
| 375 |
+
"rewards/chosen": 8.5625,
|
| 376 |
+
"rewards/margins": 15.675000190734863,
|
| 377 |
+
"rewards/rejected": -7.118750095367432,
|
| 378 |
+
"step": 220
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 0.42435424354243545,
|
| 382 |
+
"grad_norm": 37.84292558111141,
|
| 383 |
+
"learning_rate": 3.213552361396304e-07,
|
| 384 |
+
"logits/chosen": -0.41533201932907104,
|
| 385 |
+
"logits/rejected": -0.09333495795726776,
|
| 386 |
+
"logps/chosen": -491.0,
|
| 387 |
+
"logps/rejected": -661.5999755859375,
|
| 388 |
+
"loss": 0.0038,
|
| 389 |
+
"nll_loss": 1.3156249523162842,
|
| 390 |
+
"rewards/accuracies": 1.0,
|
| 391 |
+
"rewards/chosen": 10.203125,
|
| 392 |
+
"rewards/margins": 16.450000762939453,
|
| 393 |
+
"rewards/rejected": -6.25,
|
| 394 |
+
"step": 230
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.44280442804428044,
|
| 398 |
+
"grad_norm": 0.23253825124415758,
|
| 399 |
+
"learning_rate": 3.1108829568788496e-07,
|
| 400 |
+
"logits/chosen": -0.15390625596046448,
|
| 401 |
+
"logits/rejected": 0.15219727158546448,
|
| 402 |
+
"logps/chosen": -409.29998779296875,
|
| 403 |
+
"logps/rejected": -550.0,
|
| 404 |
+
"loss": 0.0245,
|
| 405 |
+
"nll_loss": 1.2214844226837158,
|
| 406 |
+
"rewards/accuracies": 0.987500011920929,
|
| 407 |
+
"rewards/chosen": 10.21875,
|
| 408 |
+
"rewards/margins": 18.28125,
|
| 409 |
+
"rewards/rejected": -8.065625190734863,
|
| 410 |
+
"step": 240
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.4612546125461255,
|
| 414 |
+
"grad_norm": 0.6855044368096641,
|
| 415 |
+
"learning_rate": 3.008213552361396e-07,
|
| 416 |
+
"logits/chosen": -0.2812744081020355,
|
| 417 |
+
"logits/rejected": 0.24277344346046448,
|
| 418 |
+
"logps/chosen": -502.20001220703125,
|
| 419 |
+
"logps/rejected": -546.2000122070312,
|
| 420 |
+
"loss": 0.0018,
|
| 421 |
+
"nll_loss": 1.3703124523162842,
|
| 422 |
+
"rewards/accuracies": 1.0,
|
| 423 |
+
"rewards/chosen": 10.5,
|
| 424 |
+
"rewards/margins": 17.024999618530273,
|
| 425 |
+
"rewards/rejected": -6.518750190734863,
|
| 426 |
+
"step": 250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 0.4797047970479705,
|
| 430 |
+
"grad_norm": 0.7753136116337657,
|
| 431 |
+
"learning_rate": 2.9055441478439427e-07,
|
| 432 |
+
"logits/chosen": -0.26799315214157104,
|
| 433 |
+
"logits/rejected": 0.06092529371380806,
|
| 434 |
+
"logps/chosen": -447.79998779296875,
|
| 435 |
+
"logps/rejected": -633.7999877929688,
|
| 436 |
+
"loss": 0.0025,
|
| 437 |
+
"nll_loss": 1.353124976158142,
|
| 438 |
+
"rewards/accuracies": 1.0,
|
| 439 |
+
"rewards/chosen": 10.787500381469727,
|
| 440 |
+
"rewards/margins": 16.918750762939453,
|
| 441 |
+
"rewards/rejected": -6.120312690734863,
|
| 442 |
+
"step": 260
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 0.4981549815498155,
|
| 446 |
+
"grad_norm": 0.10339419092952733,
|
| 447 |
+
"learning_rate": 2.8028747433264887e-07,
|
| 448 |
+
"logits/chosen": -0.22612304985523224,
|
| 449 |
+
"logits/rejected": 0.2598632872104645,
|
| 450 |
+
"logps/chosen": -411.3999938964844,
|
| 451 |
+
"logps/rejected": -534.5999755859375,
|
| 452 |
+
"loss": 0.0025,
|
| 453 |
+
"nll_loss": 1.2843749523162842,
|
| 454 |
+
"rewards/accuracies": 1.0,
|
| 455 |
+
"rewards/chosen": 9.537500381469727,
|
| 456 |
+
"rewards/margins": 15.912500381469727,
|
| 457 |
+
"rewards/rejected": -6.354687690734863,
|
| 458 |
+
"step": 270
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.5166051660516605,
|
| 462 |
+
"grad_norm": 0.1215998717046616,
|
| 463 |
+
"learning_rate": 2.7002053388090347e-07,
|
| 464 |
+
"logits/chosen": -0.34821778535842896,
|
| 465 |
+
"logits/rejected": 0.02799682691693306,
|
| 466 |
+
"logps/chosen": -348.3999938964844,
|
| 467 |
+
"logps/rejected": -595.2000122070312,
|
| 468 |
+
"loss": 0.0015,
|
| 469 |
+
"nll_loss": 1.345312476158142,
|
| 470 |
+
"rewards/accuracies": 1.0,
|
| 471 |
+
"rewards/chosen": 10.699999809265137,
|
| 472 |
+
"rewards/margins": 17.975000381469727,
|
| 473 |
+
"rewards/rejected": -7.259375095367432,
|
| 474 |
+
"step": 280
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.5350553505535055,
|
| 478 |
+
"grad_norm": 0.02878650522676844,
|
| 479 |
+
"learning_rate": 2.597535934291581e-07,
|
| 480 |
+
"logits/chosen": -0.378662109375,
|
| 481 |
+
"logits/rejected": 0.10524902492761612,
|
| 482 |
+
"logps/chosen": -372.1000061035156,
|
| 483 |
+
"logps/rejected": -542.0,
|
| 484 |
+
"loss": 0.0014,
|
| 485 |
+
"nll_loss": 1.271875023841858,
|
| 486 |
+
"rewards/accuracies": 1.0,
|
| 487 |
+
"rewards/chosen": 10.925000190734863,
|
| 488 |
+
"rewards/margins": 18.943750381469727,
|
| 489 |
+
"rewards/rejected": -8.015625,
|
| 490 |
+
"step": 290
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 0.5535055350553506,
|
| 494 |
+
"grad_norm": 0.03536097923100537,
|
| 495 |
+
"learning_rate": 2.4948665297741273e-07,
|
| 496 |
+
"logits/chosen": -0.3155273497104645,
|
| 497 |
+
"logits/rejected": -0.10869140923023224,
|
| 498 |
+
"logps/chosen": -419.1000061035156,
|
| 499 |
+
"logps/rejected": -685.2000122070312,
|
| 500 |
+
"loss": 0.0025,
|
| 501 |
+
"nll_loss": 1.337890625,
|
| 502 |
+
"rewards/accuracies": 1.0,
|
| 503 |
+
"rewards/chosen": 10.925000190734863,
|
| 504 |
+
"rewards/margins": 17.868749618530273,
|
| 505 |
+
"rewards/rejected": -6.954687595367432,
|
| 506 |
+
"step": 300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.5719557195571956,
|
| 510 |
+
"grad_norm": 0.03473546153791582,
|
| 511 |
+
"learning_rate": 2.3921971252566733e-07,
|
| 512 |
+
"logits/chosen": -0.31220704317092896,
|
| 513 |
+
"logits/rejected": 0.03559570387005806,
|
| 514 |
+
"logps/chosen": -417.1000061035156,
|
| 515 |
+
"logps/rejected": -614.0,
|
| 516 |
+
"loss": 0.0014,
|
| 517 |
+
"nll_loss": 1.290624976158142,
|
| 518 |
+
"rewards/accuracies": 1.0,
|
| 519 |
+
"rewards/chosen": 10.956250190734863,
|
| 520 |
+
"rewards/margins": 17.387500762939453,
|
| 521 |
+
"rewards/rejected": -6.418749809265137,
|
| 522 |
+
"step": 310
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.5904059040590406,
|
| 526 |
+
"grad_norm": 36.361485146275975,
|
| 527 |
+
"learning_rate": 2.2895277207392196e-07,
|
| 528 |
+
"logits/chosen": -0.28837889432907104,
|
| 529 |
+
"logits/rejected": 0.21499022841453552,
|
| 530 |
+
"logps/chosen": -387.79998779296875,
|
| 531 |
+
"logps/rejected": -581.2000122070312,
|
| 532 |
+
"loss": 0.0045,
|
| 533 |
+
"nll_loss": 1.299218773841858,
|
| 534 |
+
"rewards/accuracies": 1.0,
|
| 535 |
+
"rewards/chosen": 11.431249618530273,
|
| 536 |
+
"rewards/margins": 17.537500381469727,
|
| 537 |
+
"rewards/rejected": -6.108593940734863,
|
| 538 |
+
"step": 320
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 0.6088560885608856,
|
| 542 |
+
"grad_norm": 0.017578308664404376,
|
| 543 |
+
"learning_rate": 2.1868583162217656e-07,
|
| 544 |
+
"logits/chosen": -0.16363525390625,
|
| 545 |
+
"logits/rejected": 0.0478515625,
|
| 546 |
+
"logps/chosen": -438.3999938964844,
|
| 547 |
+
"logps/rejected": -585.5999755859375,
|
| 548 |
+
"loss": 0.0013,
|
| 549 |
+
"nll_loss": 1.249609351158142,
|
| 550 |
+
"rewards/accuracies": 1.0,
|
| 551 |
+
"rewards/chosen": 11.34375,
|
| 552 |
+
"rewards/margins": 17.706249237060547,
|
| 553 |
+
"rewards/rejected": -6.364062309265137,
|
| 554 |
+
"step": 330
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.6273062730627307,
|
| 558 |
+
"grad_norm": 0.13503151970629174,
|
| 559 |
+
"learning_rate": 2.0841889117043122e-07,
|
| 560 |
+
"logits/chosen": -0.276611328125,
|
| 561 |
+
"logits/rejected": 0.21477051079273224,
|
| 562 |
+
"logps/chosen": -387.79998779296875,
|
| 563 |
+
"logps/rejected": -552.0,
|
| 564 |
+
"loss": 0.0014,
|
| 565 |
+
"nll_loss": 1.246484398841858,
|
| 566 |
+
"rewards/accuracies": 1.0,
|
| 567 |
+
"rewards/chosen": 12.106249809265137,
|
| 568 |
+
"rewards/margins": 18.924999237060547,
|
| 569 |
+
"rewards/rejected": -6.785937309265137,
|
| 570 |
+
"step": 340
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.6457564575645757,
|
| 574 |
+
"grad_norm": 0.2744112671801272,
|
| 575 |
+
"learning_rate": 1.9815195071868582e-07,
|
| 576 |
+
"logits/chosen": -0.26923829317092896,
|
| 577 |
+
"logits/rejected": 0.15705566108226776,
|
| 578 |
+
"logps/chosen": -448.6000061035156,
|
| 579 |
+
"logps/rejected": -711.0,
|
| 580 |
+
"loss": 0.0028,
|
| 581 |
+
"nll_loss": 1.2371094226837158,
|
| 582 |
+
"rewards/accuracies": 1.0,
|
| 583 |
+
"rewards/chosen": 12.068750381469727,
|
| 584 |
+
"rewards/margins": 18.393749237060547,
|
| 585 |
+
"rewards/rejected": -6.315625190734863,
|
| 586 |
+
"step": 350
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 0.6642066420664207,
|
| 590 |
+
"grad_norm": 1.288974393127278,
|
| 591 |
+
"learning_rate": 1.8788501026694045e-07,
|
| 592 |
+
"logits/chosen": -0.2408447265625,
|
| 593 |
+
"logits/rejected": 0.12333984673023224,
|
| 594 |
+
"logps/chosen": -428.20001220703125,
|
| 595 |
+
"logps/rejected": -714.4000244140625,
|
| 596 |
+
"loss": 0.0014,
|
| 597 |
+
"nll_loss": 1.2585937976837158,
|
| 598 |
+
"rewards/accuracies": 1.0,
|
| 599 |
+
"rewards/chosen": 12.168749809265137,
|
| 600 |
+
"rewards/margins": 18.75,
|
| 601 |
+
"rewards/rejected": -6.59375,
|
| 602 |
+
"step": 360
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.6826568265682657,
|
| 606 |
+
"grad_norm": 0.010817526846263152,
|
| 607 |
+
"learning_rate": 1.7761806981519505e-07,
|
| 608 |
+
"logits/chosen": -0.4193359315395355,
|
| 609 |
+
"logits/rejected": 0.12089844048023224,
|
| 610 |
+
"logps/chosen": -387.6000061035156,
|
| 611 |
+
"logps/rejected": -546.5999755859375,
|
| 612 |
+
"loss": 0.0013,
|
| 613 |
+
"nll_loss": 1.220312476158142,
|
| 614 |
+
"rewards/accuracies": 1.0,
|
| 615 |
+
"rewards/chosen": 12.699999809265137,
|
| 616 |
+
"rewards/margins": 20.431249618530273,
|
| 617 |
+
"rewards/rejected": -7.745312690734863,
|
| 618 |
+
"step": 370
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.7011070110701108,
|
| 622 |
+
"grad_norm": 0.9313842568161499,
|
| 623 |
+
"learning_rate": 1.673511293634497e-07,
|
| 624 |
+
"logits/chosen": -0.2831054627895355,
|
| 625 |
+
"logits/rejected": 0.2867675721645355,
|
| 626 |
+
"logps/chosen": -284.8999938964844,
|
| 627 |
+
"logps/rejected": -574.7999877929688,
|
| 628 |
+
"loss": 0.0015,
|
| 629 |
+
"nll_loss": 1.189062476158142,
|
| 630 |
+
"rewards/accuracies": 1.0,
|
| 631 |
+
"rewards/chosen": 11.875,
|
| 632 |
+
"rewards/margins": 19.59375,
|
| 633 |
+
"rewards/rejected": -7.71875,
|
| 634 |
+
"step": 380
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.7195571955719557,
|
| 638 |
+
"grad_norm": 0.015457285064348653,
|
| 639 |
+
"learning_rate": 1.570841889117043e-07,
|
| 640 |
+
"logits/chosen": -0.38752442598342896,
|
| 641 |
+
"logits/rejected": 0.01973876915872097,
|
| 642 |
+
"logps/chosen": -436.3999938964844,
|
| 643 |
+
"logps/rejected": -622.4000244140625,
|
| 644 |
+
"loss": 0.0106,
|
| 645 |
+
"nll_loss": 1.1261718273162842,
|
| 646 |
+
"rewards/accuracies": 0.987500011920929,
|
| 647 |
+
"rewards/chosen": 11.868749618530273,
|
| 648 |
+
"rewards/margins": 17.862499237060547,
|
| 649 |
+
"rewards/rejected": -5.993750095367432,
|
| 650 |
+
"step": 390
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.7380073800738007,
|
| 654 |
+
"grad_norm": 0.14413562107865965,
|
| 655 |
+
"learning_rate": 1.4681724845995894e-07,
|
| 656 |
+
"logits/chosen": -0.24350586533546448,
|
| 657 |
+
"logits/rejected": 0.20991210639476776,
|
| 658 |
+
"logps/chosen": -371.1000061035156,
|
| 659 |
+
"logps/rejected": -562.5999755859375,
|
| 660 |
+
"loss": 0.0014,
|
| 661 |
+
"nll_loss": 1.131250023841858,
|
| 662 |
+
"rewards/accuracies": 1.0,
|
| 663 |
+
"rewards/chosen": 11.506250381469727,
|
| 664 |
+
"rewards/margins": 18.356250762939453,
|
| 665 |
+
"rewards/rejected": -6.846875190734863,
|
| 666 |
+
"step": 400
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 0.7564575645756457,
|
| 670 |
+
"grad_norm": 0.015979900222072605,
|
| 671 |
+
"learning_rate": 1.3655030800821354e-07,
|
| 672 |
+
"logits/chosen": -0.2548828125,
|
| 673 |
+
"logits/rejected": 0.2605957090854645,
|
| 674 |
+
"logps/chosen": -368.70001220703125,
|
| 675 |
+
"logps/rejected": -534.5999755859375,
|
| 676 |
+
"loss": 0.0012,
|
| 677 |
+
"nll_loss": 1.148046851158142,
|
| 678 |
+
"rewards/accuracies": 1.0,
|
| 679 |
+
"rewards/chosen": 13.168749809265137,
|
| 680 |
+
"rewards/margins": 20.318750381469727,
|
| 681 |
+
"rewards/rejected": -7.157812595367432,
|
| 682 |
+
"step": 410
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.7749077490774908,
|
| 686 |
+
"grad_norm": 0.009958077789392873,
|
| 687 |
+
"learning_rate": 1.2628336755646817e-07,
|
| 688 |
+
"logits/chosen": -0.34492188692092896,
|
| 689 |
+
"logits/rejected": 0.02458496019244194,
|
| 690 |
+
"logps/chosen": -404.3999938964844,
|
| 691 |
+
"logps/rejected": -572.5999755859375,
|
| 692 |
+
"loss": 0.0334,
|
| 693 |
+
"nll_loss": 1.1804687976837158,
|
| 694 |
+
"rewards/accuracies": 0.987500011920929,
|
| 695 |
+
"rewards/chosen": 11.949999809265137,
|
| 696 |
+
"rewards/margins": 17.006250381469727,
|
| 697 |
+
"rewards/rejected": -5.053124904632568,
|
| 698 |
+
"step": 420
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.7933579335793358,
|
| 702 |
+
"grad_norm": 0.018519608671113973,
|
| 703 |
+
"learning_rate": 1.1601642710472278e-07,
|
| 704 |
+
"logits/chosen": -0.3464599549770355,
|
| 705 |
+
"logits/rejected": 0.13808593153953552,
|
| 706 |
+
"logps/chosen": -344.5,
|
| 707 |
+
"logps/rejected": -527.0,
|
| 708 |
+
"loss": 0.0015,
|
| 709 |
+
"nll_loss": 1.1984374523162842,
|
| 710 |
+
"rewards/accuracies": 1.0,
|
| 711 |
+
"rewards/chosen": 13.162500381469727,
|
| 712 |
+
"rewards/margins": 19.318750381469727,
|
| 713 |
+
"rewards/rejected": -6.146874904632568,
|
| 714 |
+
"step": 430
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.8118081180811808,
|
| 718 |
+
"grad_norm": 0.014726861292652616,
|
| 719 |
+
"learning_rate": 1.0574948665297741e-07,
|
| 720 |
+
"logits/chosen": -0.3399902284145355,
|
| 721 |
+
"logits/rejected": 0.012939453125,
|
| 722 |
+
"logps/chosen": -365.29998779296875,
|
| 723 |
+
"logps/rejected": -544.5999755859375,
|
| 724 |
+
"loss": 0.0012,
|
| 725 |
+
"nll_loss": 1.1847655773162842,
|
| 726 |
+
"rewards/accuracies": 1.0,
|
| 727 |
+
"rewards/chosen": 12.768750190734863,
|
| 728 |
+
"rewards/margins": 20.362499237060547,
|
| 729 |
+
"rewards/rejected": -7.607812404632568,
|
| 730 |
+
"step": 440
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.8302583025830258,
|
| 734 |
+
"grad_norm": 0.224159824109424,
|
| 735 |
+
"learning_rate": 9.548254620123203e-08,
|
| 736 |
+
"logits/chosen": -0.2451171875,
|
| 737 |
+
"logits/rejected": 0.18867187201976776,
|
| 738 |
+
"logps/chosen": -461.20001220703125,
|
| 739 |
+
"logps/rejected": -573.2000122070312,
|
| 740 |
+
"loss": 0.0012,
|
| 741 |
+
"nll_loss": 1.203515648841858,
|
| 742 |
+
"rewards/accuracies": 1.0,
|
| 743 |
+
"rewards/chosen": 12.550000190734863,
|
| 744 |
+
"rewards/margins": 18.918750762939453,
|
| 745 |
+
"rewards/rejected": -6.364843845367432,
|
| 746 |
+
"step": 450
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.8487084870848709,
|
| 750 |
+
"grad_norm": 0.3216885287305295,
|
| 751 |
+
"learning_rate": 8.521560574948664e-08,
|
| 752 |
+
"logits/chosen": -0.4150390625,
|
| 753 |
+
"logits/rejected": 0.19716796278953552,
|
| 754 |
+
"logps/chosen": -384.70001220703125,
|
| 755 |
+
"logps/rejected": -514.5999755859375,
|
| 756 |
+
"loss": 0.0014,
|
| 757 |
+
"nll_loss": 1.2429687976837158,
|
| 758 |
+
"rewards/accuracies": 1.0,
|
| 759 |
+
"rewards/chosen": 12.600000381469727,
|
| 760 |
+
"rewards/margins": 19.018749237060547,
|
| 761 |
+
"rewards/rejected": -6.389843940734863,
|
| 762 |
+
"step": 460
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 0.8671586715867159,
|
| 766 |
+
"grad_norm": 0.036868046311022036,
|
| 767 |
+
"learning_rate": 7.494866529774127e-08,
|
| 768 |
+
"logits/chosen": -0.44775390625,
|
| 769 |
+
"logits/rejected": 0.04902343824505806,
|
| 770 |
+
"logps/chosen": -364.3999938964844,
|
| 771 |
+
"logps/rejected": -635.5999755859375,
|
| 772 |
+
"loss": 0.0013,
|
| 773 |
+
"nll_loss": 1.2976562976837158,
|
| 774 |
+
"rewards/accuracies": 1.0,
|
| 775 |
+
"rewards/chosen": 11.993749618530273,
|
| 776 |
+
"rewards/margins": 18.274999618530273,
|
| 777 |
+
"rewards/rejected": -6.285937309265137,
|
| 778 |
+
"step": 470
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 0.8856088560885609,
|
| 782 |
+
"grad_norm": 0.5078671632789603,
|
| 783 |
+
"learning_rate": 6.468172484599589e-08,
|
| 784 |
+
"logits/chosen": -0.24858398735523224,
|
| 785 |
+
"logits/rejected": -0.013867187313735485,
|
| 786 |
+
"logps/chosen": -319.5,
|
| 787 |
+
"logps/rejected": -668.7999877929688,
|
| 788 |
+
"loss": 0.0013,
|
| 789 |
+
"nll_loss": 1.154296875,
|
| 790 |
+
"rewards/accuracies": 1.0,
|
| 791 |
+
"rewards/chosen": 12.774999618530273,
|
| 792 |
+
"rewards/margins": 19.331249237060547,
|
| 793 |
+
"rewards/rejected": -6.546875,
|
| 794 |
+
"step": 480
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.9040590405904059,
|
| 798 |
+
"grad_norm": 0.02980417558547798,
|
| 799 |
+
"learning_rate": 5.441478439425051e-08,
|
| 800 |
+
"logits/chosen": -0.26640623807907104,
|
| 801 |
+
"logits/rejected": 0.18837890028953552,
|
| 802 |
+
"logps/chosen": -387.79998779296875,
|
| 803 |
+
"logps/rejected": -545.5999755859375,
|
| 804 |
+
"loss": 0.002,
|
| 805 |
+
"nll_loss": 1.148046851158142,
|
| 806 |
+
"rewards/accuracies": 1.0,
|
| 807 |
+
"rewards/chosen": 12.412500381469727,
|
| 808 |
+
"rewards/margins": 19.231250762939453,
|
| 809 |
+
"rewards/rejected": -6.824999809265137,
|
| 810 |
+
"step": 490
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.922509225092251,
|
| 814 |
+
"grad_norm": 0.187006563020737,
|
| 815 |
+
"learning_rate": 4.414784394250513e-08,
|
| 816 |
+
"logits/chosen": -0.24520263075828552,
|
| 817 |
+
"logits/rejected": 0.14360351860523224,
|
| 818 |
+
"logps/chosen": -381.20001220703125,
|
| 819 |
+
"logps/rejected": -571.5999755859375,
|
| 820 |
+
"loss": 0.0012,
|
| 821 |
+
"nll_loss": 1.130468726158142,
|
| 822 |
+
"rewards/accuracies": 1.0,
|
| 823 |
+
"rewards/chosen": 12.1875,
|
| 824 |
+
"rewards/margins": 19.200000762939453,
|
| 825 |
+
"rewards/rejected": -6.991406440734863,
|
| 826 |
+
"step": 500
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.940959409594096,
|
| 830 |
+
"grad_norm": 0.1976693035630184,
|
| 831 |
+
"learning_rate": 3.3880903490759755e-08,
|
| 832 |
+
"logits/chosen": -0.35673826932907104,
|
| 833 |
+
"logits/rejected": -0.04693603515625,
|
| 834 |
+
"logps/chosen": -359.70001220703125,
|
| 835 |
+
"logps/rejected": -559.4000244140625,
|
| 836 |
+
"loss": 0.0017,
|
| 837 |
+
"nll_loss": 1.214453101158142,
|
| 838 |
+
"rewards/accuracies": 1.0,
|
| 839 |
+
"rewards/chosen": 12.074999809265137,
|
| 840 |
+
"rewards/margins": 18.293750762939453,
|
| 841 |
+
"rewards/rejected": -6.229687690734863,
|
| 842 |
+
"step": 510
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.959409594095941,
|
| 846 |
+
"grad_norm": 0.19235015104809136,
|
| 847 |
+
"learning_rate": 2.361396303901437e-08,
|
| 848 |
+
"logits/chosen": -0.4139648377895355,
|
| 849 |
+
"logits/rejected": 0.10092773288488388,
|
| 850 |
+
"logps/chosen": -356.79998779296875,
|
| 851 |
+
"logps/rejected": -668.4000244140625,
|
| 852 |
+
"loss": 0.0012,
|
| 853 |
+
"nll_loss": 1.1335937976837158,
|
| 854 |
+
"rewards/accuracies": 1.0,
|
| 855 |
+
"rewards/chosen": 12.293749809265137,
|
| 856 |
+
"rewards/margins": 19.274999618530273,
|
| 857 |
+
"rewards/rejected": -7.0,
|
| 858 |
+
"step": 520
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.977859778597786,
|
| 862 |
+
"grad_norm": 0.011034995549329435,
|
| 863 |
+
"learning_rate": 1.3347022587268993e-08,
|
| 864 |
+
"logits/chosen": -0.3798828125,
|
| 865 |
+
"logits/rejected": 0.1337890625,
|
| 866 |
+
"logps/chosen": -325.29998779296875,
|
| 867 |
+
"logps/rejected": -492.3999938964844,
|
| 868 |
+
"loss": 0.0012,
|
| 869 |
+
"nll_loss": 1.123437523841858,
|
| 870 |
+
"rewards/accuracies": 1.0,
|
| 871 |
+
"rewards/chosen": 12.818750381469727,
|
| 872 |
+
"rewards/margins": 20.15625,
|
| 873 |
+
"rewards/rejected": -7.365624904632568,
|
| 874 |
+
"step": 530
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.996309963099631,
|
| 878 |
+
"grad_norm": 0.0682979395748976,
|
| 879 |
+
"learning_rate": 3.080082135523614e-09,
|
| 880 |
+
"logits/chosen": -0.31596678495407104,
|
| 881 |
+
"logits/rejected": 0.13836669921875,
|
| 882 |
+
"logps/chosen": -375.0,
|
| 883 |
+
"logps/rejected": -526.0,
|
| 884 |
+
"loss": 0.0012,
|
| 885 |
+
"nll_loss": 1.1609375476837158,
|
| 886 |
+
"rewards/accuracies": 1.0,
|
| 887 |
+
"rewards/chosen": 12.78125,
|
| 888 |
+
"rewards/margins": 19.537500381469727,
|
| 889 |
+
"rewards/rejected": -6.743750095367432,
|
| 890 |
+
"step": 540
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.0,
|
| 894 |
+
"eval_logits/chosen": -0.31317609548568726,
|
| 895 |
+
"eval_logits/rejected": 0.21189998090267181,
|
| 896 |
+
"eval_logps/chosen": -294.4615478515625,
|
| 897 |
+
"eval_logps/rejected": -450.1538391113281,
|
| 898 |
+
"eval_loss": 0.004863434005528688,
|
| 899 |
+
"eval_nll_loss": 1.1430288553237915,
|
| 900 |
+
"eval_rewards/accuracies": 1.0,
|
| 901 |
+
"eval_rewards/chosen": 12.75,
|
| 902 |
+
"eval_rewards/margins": 18.009614944458008,
|
| 903 |
+
"eval_rewards/rejected": -5.266225814819336,
|
| 904 |
+
"eval_runtime": 15.407,
|
| 905 |
+
"eval_samples_per_second": 6.491,
|
| 906 |
+
"eval_steps_per_second": 0.844,
|
| 907 |
+
"step": 542
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.0,
|
| 911 |
+
"step": 542,
|
| 912 |
+
"total_flos": 0.0,
|
| 913 |
+
"train_loss": 0.03677848608291457,
|
| 914 |
+
"train_runtime": 1919.659,
|
| 915 |
+
"train_samples_per_second": 2.257,
|
| 916 |
+
"train_steps_per_second": 0.282
|
| 917 |
+
}
|
| 918 |
+
],
|
| 919 |
+
"logging_steps": 10,
|
| 920 |
+
"max_steps": 542,
|
| 921 |
+
"num_input_tokens_seen": 0,
|
| 922 |
+
"num_train_epochs": 1,
|
| 923 |
+
"save_steps": 500,
|
| 924 |
+
"stateful_callbacks": {
|
| 925 |
+
"TrainerControl": {
|
| 926 |
+
"args": {
|
| 927 |
+
"should_epoch_stop": false,
|
| 928 |
+
"should_evaluate": false,
|
| 929 |
+
"should_log": false,
|
| 930 |
+
"should_save": true,
|
| 931 |
+
"should_training_stop": true
|
| 932 |
+
},
|
| 933 |
+
"attributes": {}
|
| 934 |
+
}
|
| 935 |
+
},
|
| 936 |
+
"total_flos": 0.0,
|
| 937 |
+
"train_batch_size": 2,
|
| 938 |
+
"trial_name": null,
|
| 939 |
+
"trial_params": null
|
| 940 |
+
}
|