Model save
Browse files- README.md +21 -1
- all_results.json +5 -5
- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- train_results.json +5 -5
- trainer_state.json +1359 -69
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -16,6 +16,16 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
# zephyr-7b-dpo-full
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the None dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
@@ -37,7 +47,7 @@ The following hyperparameters were used during training:
|
|
| 37 |
- learning_rate: 5e-07
|
| 38 |
- train_batch_size: 8
|
| 39 |
- eval_batch_size: 8
|
| 40 |
-
- seed:
|
| 41 |
- distributed_type: multi-GPU
|
| 42 |
- num_devices: 8
|
| 43 |
- gradient_accumulation_steps: 2
|
|
@@ -50,6 +60,16 @@ The following hyperparameters were used during training:
|
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
|
| 54 |
|
| 55 |
### Framework versions
|
|
|
|
| 16 |
# zephyr-7b-dpo-full
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the None dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.0465
|
| 21 |
+
- Rewards/chosen: -2.6400
|
| 22 |
+
- Rewards/rejected: -3.4900
|
| 23 |
+
- Rewards/accuracies: 0.7227
|
| 24 |
+
- Rewards/margins: 0.8499
|
| 25 |
+
- Logps/rejected: -606.3505
|
| 26 |
+
- Logps/chosen: -521.0439
|
| 27 |
+
- Logits/rejected: -1.9091
|
| 28 |
+
- Logits/chosen: -1.9501
|
| 29 |
|
| 30 |
## Model description
|
| 31 |
|
|
|
|
| 47 |
- learning_rate: 5e-07
|
| 48 |
- train_batch_size: 8
|
| 49 |
- eval_batch_size: 8
|
| 50 |
+
- seed: 5
|
| 51 |
- distributed_type: multi-GPU
|
| 52 |
- num_devices: 8
|
| 53 |
- gradient_accumulation_steps: 2
|
|
|
|
| 60 |
|
| 61 |
### Training results
|
| 62 |
|
| 63 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 64 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 65 |
+
| 0.1485 | 0.11 | 100 | 0.1803 | -0.5621 | -0.7737 | 0.6406 | 0.2117 | -334.7263 | -313.2471 | -2.4998 | -2.5133 |
|
| 66 |
+
| 0.0592 | 0.23 | 200 | 0.0662 | -1.7402 | -2.3280 | 0.6797 | 0.5878 | -490.1518 | -431.0574 | -2.2396 | -2.2729 |
|
| 67 |
+
| 0.0394 | 0.34 | 300 | 0.0494 | -2.3707 | -2.9767 | 0.6953 | 0.6061 | -555.0248 | -494.1047 | -2.1101 | -2.1389 |
|
| 68 |
+
| 0.0401 | 0.45 | 400 | 0.0523 | -2.4275 | -3.1076 | 0.7031 | 0.6801 | -568.1116 | -499.7916 | -2.0429 | -2.0799 |
|
| 69 |
+
| 0.0335 | 0.57 | 500 | 0.0461 | -2.4063 | -3.2276 | 0.7148 | 0.8213 | -580.1129 | -497.6727 | -2.0057 | -2.0456 |
|
| 70 |
+
| 0.0273 | 0.68 | 600 | 0.0409 | -2.8465 | -3.7152 | 0.7070 | 0.8687 | -628.8741 | -541.6862 | -1.9162 | -1.9558 |
|
| 71 |
+
| 0.0377 | 0.79 | 700 | 0.0496 | -2.5317 | -3.3682 | 0.7227 | 0.8365 | -594.1712 | -510.2102 | -1.9274 | -1.9673 |
|
| 72 |
+
| 0.0352 | 0.91 | 800 | 0.0465 | -2.6400 | -3.4900 | 0.7227 | 0.8499 | -606.3505 | -521.0439 | -1.9091 | -1.9501 |
|
| 73 |
|
| 74 |
|
| 75 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.06580274857555349,
|
| 4 |
+
"train_runtime": 7965.4661,
|
| 5 |
+
"train_samples": 113028,
|
| 6 |
+
"train_samples_per_second": 14.19,
|
| 7 |
+
"train_steps_per_second": 0.111
|
| 8 |
}
|
model-00001-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4943162336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93a30ad43f88322bd82e3026803a89f38f46679584ddb5a8576ed49805581903
|
| 3 |
size 4943162336
|
model-00002-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999819336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dc2c468bd767d9f03fd1f458561b58b4a1064074ab3091b661b4a07ba1978269
|
| 3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4540516344
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:11f222d47eda9d06a27b7b27cbe6f579baf4af911dd2d99a0968a94755f3f515
|
| 3 |
size 4540516344
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.06580274857555349,
|
| 4 |
+
"train_runtime": 7965.4661,
|
| 5 |
+
"train_samples": 113028,
|
| 6 |
+
"train_samples_per_second": 14.19,
|
| 7 |
+
"train_steps_per_second": 0.111
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,21 +1,21 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"learning_rate":
|
| 14 |
-
"logits/chosen": -2.
|
| 15 |
-
"logits/rejected": -2.
|
| 16 |
-
"logps/chosen": -
|
| 17 |
-
"logps/rejected": -
|
| 18 |
-
"loss": 0.
|
| 19 |
"rewards/accuracies": 0.0,
|
| 20 |
"rewards/chosen": 0.0,
|
| 21 |
"rewards/margins": 0.0,
|
|
@@ -23,87 +23,1377 @@
|
|
| 23 |
"step": 1
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"learning_rate":
|
| 28 |
-
"logits/chosen": -2.
|
| 29 |
-
"logits/rejected": -2.
|
| 30 |
-
"logps/chosen": -
|
| 31 |
-
"logps/rejected": -
|
| 32 |
-
"loss": 0.
|
| 33 |
-
"rewards/accuracies": 0.
|
| 34 |
-
"rewards/chosen": 0.
|
| 35 |
-
"rewards/margins": 0.
|
| 36 |
-
"rewards/rejected":
|
| 37 |
"step": 10
|
| 38 |
},
|
| 39 |
{
|
| 40 |
-
"epoch": 0.
|
| 41 |
-
"learning_rate":
|
| 42 |
-
"logits/chosen": -2.
|
| 43 |
-
"logits/rejected": -2.
|
| 44 |
-
"logps/chosen": -
|
| 45 |
-
"logps/rejected": -
|
| 46 |
-
"loss": 0.
|
| 47 |
-
"rewards/accuracies": 0.
|
| 48 |
-
"rewards/chosen": 0.
|
| 49 |
-
"rewards/margins": 0.
|
| 50 |
-
"rewards/rejected": -0.
|
| 51 |
"step": 20
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate":
|
| 56 |
-
"logits/chosen": -2.
|
| 57 |
-
"logits/rejected": -2.
|
| 58 |
-
"logps/chosen": -
|
| 59 |
-
"logps/rejected": -
|
| 60 |
-
"loss": 0.
|
| 61 |
-
"rewards/accuracies": 0.
|
| 62 |
-
"rewards/chosen": 0.
|
| 63 |
-
"rewards/margins": 0.
|
| 64 |
-
"rewards/rejected": -0.
|
| 65 |
"step": 30
|
| 66 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
{
|
| 68 |
"epoch": 0.75,
|
| 69 |
-
"learning_rate":
|
| 70 |
-
"logits/chosen": -
|
| 71 |
-
"logits/rejected": -
|
| 72 |
-
"logps/chosen": -
|
| 73 |
-
"logps/rejected": -
|
| 74 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 75 |
"rewards/accuracies": 0.699999988079071,
|
| 76 |
-
"rewards/chosen":
|
| 77 |
-
"rewards/margins":
|
| 78 |
-
"rewards/rejected": -
|
| 79 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.94,
|
| 83 |
-
"learning_rate": 5.
|
| 84 |
-
"logits/chosen": -
|
| 85 |
-
"logits/rejected": -
|
| 86 |
-
"logps/chosen": -
|
| 87 |
-
"logps/rejected": -
|
| 88 |
-
"loss": 0.
|
| 89 |
-
"rewards/accuracies": 0.
|
| 90 |
-
"rewards/chosen":
|
| 91 |
-
"rewards/margins":
|
| 92 |
-
"rewards/rejected": -
|
| 93 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 1.0,
|
| 97 |
-
"step":
|
| 98 |
"total_flos": 0.0,
|
| 99 |
-
"train_loss": 0.
|
| 100 |
-
"train_runtime":
|
| 101 |
-
"train_samples_per_second":
|
| 102 |
-
"train_steps_per_second": 0.
|
| 103 |
}
|
| 104 |
],
|
| 105 |
"logging_steps": 10,
|
| 106 |
-
"max_steps":
|
| 107 |
"num_train_epochs": 1,
|
| 108 |
"save_steps": 100,
|
| 109 |
"total_flos": 0.0,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9994340690435767,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 883,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 5.617977528089887e-09,
|
| 14 |
+
"logits/chosen": -2.8261122703552246,
|
| 15 |
+
"logits/rejected": -2.782524824142456,
|
| 16 |
+
"logps/chosen": -386.01312255859375,
|
| 17 |
+
"logps/rejected": -174.26467895507812,
|
| 18 |
+
"loss": 0.2845,
|
| 19 |
"rewards/accuracies": 0.0,
|
| 20 |
"rewards/chosen": 0.0,
|
| 21 |
"rewards/margins": 0.0,
|
|
|
|
| 23 |
"step": 1
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"learning_rate": 5.617977528089887e-08,
|
| 28 |
+
"logits/chosen": -2.770503044128418,
|
| 29 |
+
"logits/rejected": -2.7067270278930664,
|
| 30 |
+
"logps/chosen": -337.03875732421875,
|
| 31 |
+
"logps/rejected": -169.81399536132812,
|
| 32 |
+
"loss": 0.2812,
|
| 33 |
+
"rewards/accuracies": 0.3541666567325592,
|
| 34 |
+
"rewards/chosen": -0.0004154888156335801,
|
| 35 |
+
"rewards/margins": -0.0005714126164093614,
|
| 36 |
+
"rewards/rejected": 0.00015592378622386605,
|
| 37 |
"step": 10
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 0.02,
|
| 41 |
+
"learning_rate": 1.1235955056179774e-07,
|
| 42 |
+
"logits/chosen": -2.804004192352295,
|
| 43 |
+
"logits/rejected": -2.800266742706299,
|
| 44 |
+
"logps/chosen": -306.04486083984375,
|
| 45 |
+
"logps/rejected": -189.47840881347656,
|
| 46 |
+
"loss": 0.2839,
|
| 47 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 48 |
+
"rewards/chosen": 0.001905800774693489,
|
| 49 |
+
"rewards/margins": 0.003199492348358035,
|
| 50 |
+
"rewards/rejected": -0.0012936916900798678,
|
| 51 |
"step": 20
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.03,
|
| 55 |
+
"learning_rate": 1.6853932584269663e-07,
|
| 56 |
+
"logits/chosen": -2.7842857837677,
|
| 57 |
+
"logits/rejected": -2.7480905055999756,
|
| 58 |
+
"logps/chosen": -340.1336669921875,
|
| 59 |
+
"logps/rejected": -176.36483764648438,
|
| 60 |
+
"loss": 0.2736,
|
| 61 |
+
"rewards/accuracies": 0.65625,
|
| 62 |
+
"rewards/chosen": 0.013033352792263031,
|
| 63 |
+
"rewards/margins": 0.02086206153035164,
|
| 64 |
+
"rewards/rejected": -0.007828707806766033,
|
| 65 |
"step": 30
|
| 66 |
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05,
|
| 69 |
+
"learning_rate": 2.2471910112359549e-07,
|
| 70 |
+
"logits/chosen": -2.7320685386657715,
|
| 71 |
+
"logits/rejected": -2.694530963897705,
|
| 72 |
+
"logps/chosen": -318.46612548828125,
|
| 73 |
+
"logps/rejected": -170.09690856933594,
|
| 74 |
+
"loss": 0.2779,
|
| 75 |
+
"rewards/accuracies": 0.643750011920929,
|
| 76 |
+
"rewards/chosen": 0.011926891282200813,
|
| 77 |
+
"rewards/margins": 0.05010765790939331,
|
| 78 |
+
"rewards/rejected": -0.038180768489837646,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.06,
|
| 83 |
+
"learning_rate": 2.8089887640449437e-07,
|
| 84 |
+
"logits/chosen": -2.638516902923584,
|
| 85 |
+
"logits/rejected": -2.6276156902313232,
|
| 86 |
+
"logps/chosen": -326.9003601074219,
|
| 87 |
+
"logps/rejected": -187.60606384277344,
|
| 88 |
+
"loss": 0.2637,
|
| 89 |
+
"rewards/accuracies": 0.699999988079071,
|
| 90 |
+
"rewards/chosen": 0.04087567329406738,
|
| 91 |
+
"rewards/margins": 0.14330647885799408,
|
| 92 |
+
"rewards/rejected": -0.1024308055639267,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07,
|
| 97 |
+
"learning_rate": 3.3707865168539325e-07,
|
| 98 |
+
"logits/chosen": -2.584765672683716,
|
| 99 |
+
"logits/rejected": -2.537776231765747,
|
| 100 |
+
"logps/chosen": -350.67724609375,
|
| 101 |
+
"logps/rejected": -225.81283569335938,
|
| 102 |
+
"loss": 0.2493,
|
| 103 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 104 |
+
"rewards/chosen": -0.013082382269203663,
|
| 105 |
+
"rewards/margins": 0.19787843525409698,
|
| 106 |
+
"rewards/rejected": -0.21096083521842957,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.08,
|
| 111 |
+
"learning_rate": 3.9325842696629214e-07,
|
| 112 |
+
"logits/chosen": -2.5709831714630127,
|
| 113 |
+
"logits/rejected": -2.5257468223571777,
|
| 114 |
+
"logps/chosen": -342.55291748046875,
|
| 115 |
+
"logps/rejected": -237.45059204101562,
|
| 116 |
+
"loss": 0.2242,
|
| 117 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 118 |
+
"rewards/chosen": -0.19872836768627167,
|
| 119 |
+
"rewards/margins": 0.2264036238193512,
|
| 120 |
+
"rewards/rejected": -0.42513203620910645,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.09,
|
| 125 |
+
"learning_rate": 4.4943820224719097e-07,
|
| 126 |
+
"logits/chosen": -2.46580171585083,
|
| 127 |
+
"logits/rejected": -2.4357619285583496,
|
| 128 |
+
"logps/chosen": -338.7400817871094,
|
| 129 |
+
"logps/rejected": -255.5446319580078,
|
| 130 |
+
"loss": 0.187,
|
| 131 |
+
"rewards/accuracies": 0.643750011920929,
|
| 132 |
+
"rewards/chosen": -0.2514013946056366,
|
| 133 |
+
"rewards/margins": 0.39069053530693054,
|
| 134 |
+
"rewards/rejected": -0.6420919895172119,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1,
|
| 139 |
+
"learning_rate": 4.999980431020109e-07,
|
| 140 |
+
"logits/chosen": -2.492363452911377,
|
| 141 |
+
"logits/rejected": -2.4689993858337402,
|
| 142 |
+
"logps/chosen": -373.74151611328125,
|
| 143 |
+
"logps/rejected": -278.97564697265625,
|
| 144 |
+
"loss": 0.1471,
|
| 145 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 146 |
+
"rewards/chosen": -0.2644743025302887,
|
| 147 |
+
"rewards/margins": 0.5433769822120667,
|
| 148 |
+
"rewards/rejected": -0.8078513145446777,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.11,
|
| 153 |
+
"learning_rate": 4.997632524101301e-07,
|
| 154 |
+
"logits/chosen": -2.5200695991516113,
|
| 155 |
+
"logits/rejected": -2.493776798248291,
|
| 156 |
+
"logps/chosen": -365.40191650390625,
|
| 157 |
+
"logps/rejected": -278.2363586425781,
|
| 158 |
+
"loss": 0.1485,
|
| 159 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 160 |
+
"rewards/chosen": -0.31133976578712463,
|
| 161 |
+
"rewards/margins": 0.6167228817939758,
|
| 162 |
+
"rewards/rejected": -0.9280625581741333,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.11,
|
| 167 |
+
"eval_logits/chosen": -2.513258934020996,
|
| 168 |
+
"eval_logits/rejected": -2.4997854232788086,
|
| 169 |
+
"eval_logps/chosen": -313.2470703125,
|
| 170 |
+
"eval_logps/rejected": -334.7263488769531,
|
| 171 |
+
"eval_loss": 0.18025632202625275,
|
| 172 |
+
"eval_rewards/accuracies": 0.640625,
|
| 173 |
+
"eval_rewards/chosen": -0.5620743632316589,
|
| 174 |
+
"eval_rewards/margins": 0.2116563767194748,
|
| 175 |
+
"eval_rewards/rejected": -0.7737306952476501,
|
| 176 |
+
"eval_runtime": 53.274,
|
| 177 |
+
"eval_samples_per_second": 37.542,
|
| 178 |
+
"eval_steps_per_second": 0.601,
|
| 179 |
+
"step": 100
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.12,
|
| 183 |
+
"learning_rate": 4.991375032514749e-07,
|
| 184 |
+
"logits/chosen": -2.4822776317596436,
|
| 185 |
+
"logits/rejected": -2.445827007293701,
|
| 186 |
+
"logps/chosen": -388.3408508300781,
|
| 187 |
+
"logps/rejected": -346.55419921875,
|
| 188 |
+
"loss": 0.1265,
|
| 189 |
+
"rewards/accuracies": 0.65625,
|
| 190 |
+
"rewards/chosen": -0.5192545652389526,
|
| 191 |
+
"rewards/margins": 0.6625990867614746,
|
| 192 |
+
"rewards/rejected": -1.1818536520004272,
|
| 193 |
+
"step": 110
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.14,
|
| 197 |
+
"learning_rate": 4.98121775121344e-07,
|
| 198 |
+
"logits/chosen": -2.4391703605651855,
|
| 199 |
+
"logits/rejected": -2.4168076515197754,
|
| 200 |
+
"logps/chosen": -354.26580810546875,
|
| 201 |
+
"logps/rejected": -301.9068298339844,
|
| 202 |
+
"loss": 0.1549,
|
| 203 |
+
"rewards/accuracies": 0.699999988079071,
|
| 204 |
+
"rewards/chosen": -0.4692833423614502,
|
| 205 |
+
"rewards/margins": 0.6351093053817749,
|
| 206 |
+
"rewards/rejected": -1.104392647743225,
|
| 207 |
+
"step": 120
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.15,
|
| 211 |
+
"learning_rate": 4.96717657955441e-07,
|
| 212 |
+
"logits/chosen": -2.456784963607788,
|
| 213 |
+
"logits/rejected": -2.402589797973633,
|
| 214 |
+
"logps/chosen": -380.1900939941406,
|
| 215 |
+
"logps/rejected": -345.0179443359375,
|
| 216 |
+
"loss": 0.1076,
|
| 217 |
+
"rewards/accuracies": 0.6875,
|
| 218 |
+
"rewards/chosen": -0.7707170248031616,
|
| 219 |
+
"rewards/margins": 0.8249391317367554,
|
| 220 |
+
"rewards/rejected": -1.595656156539917,
|
| 221 |
+
"step": 130
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.16,
|
| 225 |
+
"learning_rate": 4.949273496411216e-07,
|
| 226 |
+
"logits/chosen": -2.3544960021972656,
|
| 227 |
+
"logits/rejected": -2.34092378616333,
|
| 228 |
+
"logps/chosen": -446.9285583496094,
|
| 229 |
+
"logps/rejected": -376.6002197265625,
|
| 230 |
+
"loss": 0.0663,
|
| 231 |
+
"rewards/accuracies": 0.78125,
|
| 232 |
+
"rewards/chosen": -1.0127520561218262,
|
| 233 |
+
"rewards/margins": 1.086138367652893,
|
| 234 |
+
"rewards/rejected": -2.098890542984009,
|
| 235 |
+
"step": 140
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.17,
|
| 239 |
+
"learning_rate": 4.927536525770046e-07,
|
| 240 |
+
"logits/chosen": -2.376620054244995,
|
| 241 |
+
"logits/rejected": -2.3480007648468018,
|
| 242 |
+
"logps/chosen": -434.1917419433594,
|
| 243 |
+
"logps/rejected": -401.83587646484375,
|
| 244 |
+
"loss": 0.067,
|
| 245 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 246 |
+
"rewards/chosen": -1.034598708152771,
|
| 247 |
+
"rewards/margins": 0.8842358589172363,
|
| 248 |
+
"rewards/rejected": -1.9188346862792969,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.18,
|
| 253 |
+
"learning_rate": 4.901999692863326e-07,
|
| 254 |
+
"logits/chosen": -2.378688335418701,
|
| 255 |
+
"logits/rejected": -2.3023369312286377,
|
| 256 |
+
"logps/chosen": -454.9664611816406,
|
| 257 |
+
"logps/rejected": -411.72027587890625,
|
| 258 |
+
"loss": 0.0667,
|
| 259 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 260 |
+
"rewards/chosen": -1.2886327505111694,
|
| 261 |
+
"rewards/margins": 0.9617815017700195,
|
| 262 |
+
"rewards/rejected": -2.2504146099090576,
|
| 263 |
+
"step": 160
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.19,
|
| 267 |
+
"learning_rate": 4.872702970909464e-07,
|
| 268 |
+
"logits/chosen": -2.327148914337158,
|
| 269 |
+
"logits/rejected": -2.255551338195801,
|
| 270 |
+
"logps/chosen": -474.45538330078125,
|
| 271 |
+
"logps/rejected": -428.8330078125,
|
| 272 |
+
"loss": 0.0538,
|
| 273 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 274 |
+
"rewards/chosen": -1.456198811531067,
|
| 275 |
+
"rewards/margins": 0.9602320790290833,
|
| 276 |
+
"rewards/rejected": -2.416430950164795,
|
| 277 |
+
"step": 170
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 0.2,
|
| 281 |
+
"learning_rate": 4.839692218542131e-07,
|
| 282 |
+
"logits/chosen": -2.3194737434387207,
|
| 283 |
+
"logits/rejected": -2.289604902267456,
|
| 284 |
+
"logps/chosen": -398.30853271484375,
|
| 285 |
+
"logps/rejected": -385.2601013183594,
|
| 286 |
+
"loss": 0.0682,
|
| 287 |
+
"rewards/accuracies": 0.78125,
|
| 288 |
+
"rewards/chosen": -0.9385713338851929,
|
| 289 |
+
"rewards/margins": 1.0706655979156494,
|
| 290 |
+
"rewards/rejected": -2.009237051010132,
|
| 291 |
+
"step": 180
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.22,
|
| 295 |
+
"learning_rate": 4.803019108026997e-07,
|
| 296 |
+
"logits/chosen": -2.3338797092437744,
|
| 297 |
+
"logits/rejected": -2.2716145515441895,
|
| 298 |
+
"logps/chosen": -423.6573181152344,
|
| 299 |
+
"logps/rejected": -402.0962219238281,
|
| 300 |
+
"loss": 0.0702,
|
| 301 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 302 |
+
"rewards/chosen": -1.050208568572998,
|
| 303 |
+
"rewards/margins": 0.8981093168258667,
|
| 304 |
+
"rewards/rejected": -1.9483178853988647,
|
| 305 |
+
"step": 190
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.23,
|
| 309 |
+
"learning_rate": 4.7627410443782887e-07,
|
| 310 |
+
"logits/chosen": -2.3091745376586914,
|
| 311 |
+
"logits/rejected": -2.2324705123901367,
|
| 312 |
+
"logps/chosen": -481.0274353027344,
|
| 313 |
+
"logps/rejected": -428.9169921875,
|
| 314 |
+
"loss": 0.0592,
|
| 315 |
+
"rewards/accuracies": 0.762499988079071,
|
| 316 |
+
"rewards/chosen": -1.2154655456542969,
|
| 317 |
+
"rewards/margins": 1.0429514646530151,
|
| 318 |
+
"rewards/rejected": -2.2584171295166016,
|
| 319 |
+
"step": 200
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 0.23,
|
| 323 |
+
"eval_logits/chosen": -2.2729034423828125,
|
| 324 |
+
"eval_logits/rejected": -2.2396340370178223,
|
| 325 |
+
"eval_logps/chosen": -431.05743408203125,
|
| 326 |
+
"eval_logps/rejected": -490.1517639160156,
|
| 327 |
+
"eval_loss": 0.06622401624917984,
|
| 328 |
+
"eval_rewards/accuracies": 0.6796875,
|
| 329 |
+
"eval_rewards/chosen": -1.7401777505874634,
|
| 330 |
+
"eval_rewards/margins": 0.5878072381019592,
|
| 331 |
+
"eval_rewards/rejected": -2.3279850482940674,
|
| 332 |
+
"eval_runtime": 53.2522,
|
| 333 |
+
"eval_samples_per_second": 37.557,
|
| 334 |
+
"eval_steps_per_second": 0.601,
|
| 335 |
+
"step": 200
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.24,
|
| 339 |
+
"learning_rate": 4.7189210755018034e-07,
|
| 340 |
+
"logits/chosen": -2.2162270545959473,
|
| 341 |
+
"logits/rejected": -2.109973430633545,
|
| 342 |
+
"logps/chosen": -463.5928649902344,
|
| 343 |
+
"logps/rejected": -448.2275390625,
|
| 344 |
+
"loss": 0.0504,
|
| 345 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 346 |
+
"rewards/chosen": -1.341931939125061,
|
| 347 |
+
"rewards/margins": 1.1762769222259521,
|
| 348 |
+
"rewards/rejected": -2.5182089805603027,
|
| 349 |
+
"step": 210
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 0.25,
|
| 353 |
+
"learning_rate": 4.671627793504988e-07,
|
| 354 |
+
"logits/chosen": -2.222045421600342,
|
| 355 |
+
"logits/rejected": -2.127629280090332,
|
| 356 |
+
"logps/chosen": -493.30133056640625,
|
| 357 |
+
"logps/rejected": -467.193115234375,
|
| 358 |
+
"loss": 0.0508,
|
| 359 |
+
"rewards/accuracies": 0.793749988079071,
|
| 360 |
+
"rewards/chosen": -1.4817438125610352,
|
| 361 |
+
"rewards/margins": 1.220350980758667,
|
| 362 |
+
"rewards/rejected": -2.702094554901123,
|
| 363 |
+
"step": 220
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.26,
|
| 367 |
+
"learning_rate": 4.6209352273286095e-07,
|
| 368 |
+
"logits/chosen": -2.233363628387451,
|
| 369 |
+
"logits/rejected": -2.184459686279297,
|
| 370 |
+
"logps/chosen": -456.435302734375,
|
| 371 |
+
"logps/rejected": -444.67449951171875,
|
| 372 |
+
"loss": 0.0543,
|
| 373 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 374 |
+
"rewards/chosen": -1.346243143081665,
|
| 375 |
+
"rewards/margins": 1.1588122844696045,
|
| 376 |
+
"rewards/rejected": -2.5050556659698486,
|
| 377 |
+
"step": 230
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.27,
|
| 381 |
+
"learning_rate": 4.56692272686805e-07,
|
| 382 |
+
"logits/chosen": -2.227086067199707,
|
| 383 |
+
"logits/rejected": -2.1657800674438477,
|
| 384 |
+
"logps/chosen": -463.6180114746094,
|
| 385 |
+
"logps/rejected": -463.9268493652344,
|
| 386 |
+
"loss": 0.0517,
|
| 387 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 388 |
+
"rewards/chosen": -1.6306241750717163,
|
| 389 |
+
"rewards/margins": 1.0510774850845337,
|
| 390 |
+
"rewards/rejected": -2.68170166015625,
|
| 391 |
+
"step": 240
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.28,
|
| 395 |
+
"learning_rate": 4.5096748387656326e-07,
|
| 396 |
+
"logits/chosen": -2.2646913528442383,
|
| 397 |
+
"logits/rejected": -2.1847498416900635,
|
| 398 |
+
"logps/chosen": -526.8988647460938,
|
| 399 |
+
"logps/rejected": -455.5740661621094,
|
| 400 |
+
"loss": 0.049,
|
| 401 |
+
"rewards/accuracies": 0.762499988079071,
|
| 402 |
+
"rewards/chosen": -1.5727968215942383,
|
| 403 |
+
"rewards/margins": 1.1092311143875122,
|
| 404 |
+
"rewards/rejected": -2.68202805519104,
|
| 405 |
+
"step": 250
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.29,
|
| 409 |
+
"learning_rate": 4.4492811740683877e-07,
|
| 410 |
+
"logits/chosen": -2.272037982940674,
|
| 411 |
+
"logits/rejected": -2.1871819496154785,
|
| 412 |
+
"logps/chosen": -510.32061767578125,
|
| 413 |
+
"logps/rejected": -452.8402404785156,
|
| 414 |
+
"loss": 0.0492,
|
| 415 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 416 |
+
"rewards/chosen": -1.5913618803024292,
|
| 417 |
+
"rewards/margins": 1.0165178775787354,
|
| 418 |
+
"rewards/rejected": -2.607879400253296,
|
| 419 |
+
"step": 260
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.31,
|
| 423 |
+
"learning_rate": 4.3858362679584354e-07,
|
| 424 |
+
"logits/chosen": -2.17218017578125,
|
| 425 |
+
"logits/rejected": -2.114896774291992,
|
| 426 |
+
"logps/chosen": -487.0856018066406,
|
| 427 |
+
"logps/rejected": -454.93115234375,
|
| 428 |
+
"loss": 0.0458,
|
| 429 |
+
"rewards/accuracies": 0.737500011920929,
|
| 430 |
+
"rewards/chosen": -1.7943446636199951,
|
| 431 |
+
"rewards/margins": 0.9125874638557434,
|
| 432 |
+
"rewards/rejected": -2.706932544708252,
|
| 433 |
+
"step": 270
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.32,
|
| 437 |
+
"learning_rate": 4.3194394317755245e-07,
|
| 438 |
+
"logits/chosen": -2.1766180992126465,
|
| 439 |
+
"logits/rejected": -2.0868289470672607,
|
| 440 |
+
"logps/chosen": -574.8035888671875,
|
| 441 |
+
"logps/rejected": -496.2310485839844,
|
| 442 |
+
"loss": 0.0411,
|
| 443 |
+
"rewards/accuracies": 0.793749988079071,
|
| 444 |
+
"rewards/chosen": -2.0265350341796875,
|
| 445 |
+
"rewards/margins": 1.1947269439697266,
|
| 446 |
+
"rewards/rejected": -3.221261501312256,
|
| 447 |
+
"step": 280
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.33,
|
| 451 |
+
"learning_rate": 4.2501945975633914e-07,
|
| 452 |
+
"logits/chosen": -2.130429744720459,
|
| 453 |
+
"logits/rejected": -2.0812618732452393,
|
| 454 |
+
"logps/chosen": -565.4262084960938,
|
| 455 |
+
"logps/rejected": -521.2449951171875,
|
| 456 |
+
"loss": 0.0359,
|
| 457 |
+
"rewards/accuracies": 0.731249988079071,
|
| 458 |
+
"rewards/chosen": -2.1656954288482666,
|
| 459 |
+
"rewards/margins": 1.03809654712677,
|
| 460 |
+
"rewards/rejected": -3.203791856765747,
|
| 461 |
+
"step": 290
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.34,
|
| 465 |
+
"learning_rate": 4.1782101553832405e-07,
|
| 466 |
+
"logits/chosen": -2.1119463443756104,
|
| 467 |
+
"logits/rejected": -2.0517754554748535,
|
| 468 |
+
"logps/chosen": -525.1202392578125,
|
| 469 |
+
"logps/rejected": -516.9113159179688,
|
| 470 |
+
"loss": 0.0394,
|
| 471 |
+
"rewards/accuracies": 0.75,
|
| 472 |
+
"rewards/chosen": -2.1311917304992676,
|
| 473 |
+
"rewards/margins": 1.1441915035247803,
|
| 474 |
+
"rewards/rejected": -3.2753829956054688,
|
| 475 |
+
"step": 300
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 0.34,
|
| 479 |
+
"eval_logits/chosen": -2.138880968093872,
|
| 480 |
+
"eval_logits/rejected": -2.1100966930389404,
|
| 481 |
+
"eval_logps/chosen": -494.104736328125,
|
| 482 |
+
"eval_logps/rejected": -555.0247802734375,
|
| 483 |
+
"eval_loss": 0.04939676821231842,
|
| 484 |
+
"eval_rewards/accuracies": 0.6953125,
|
| 485 |
+
"eval_rewards/chosen": -2.37065052986145,
|
| 486 |
+
"eval_rewards/margins": 0.606063961982727,
|
| 487 |
+
"eval_rewards/rejected": -2.976714611053467,
|
| 488 |
+
"eval_runtime": 53.3026,
|
| 489 |
+
"eval_samples_per_second": 37.522,
|
| 490 |
+
"eval_steps_per_second": 0.6,
|
| 491 |
+
"step": 300
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.35,
|
| 495 |
+
"learning_rate": 4.103598783649029e-07,
|
| 496 |
+
"logits/chosen": -2.1108057498931885,
|
| 497 |
+
"logits/rejected": -2.0741026401519775,
|
| 498 |
+
"logps/chosen": -521.8655395507812,
|
| 499 |
+
"logps/rejected": -494.017578125,
|
| 500 |
+
"loss": 0.0399,
|
| 501 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 502 |
+
"rewards/chosen": -1.8981420993804932,
|
| 503 |
+
"rewards/margins": 0.9213889837265015,
|
| 504 |
+
"rewards/rejected": -2.819530963897705,
|
| 505 |
+
"step": 310
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 0.36,
|
| 509 |
+
"learning_rate": 4.026477272750119e-07,
|
| 510 |
+
"logits/chosen": -2.195026159286499,
|
| 511 |
+
"logits/rejected": -2.128283977508545,
|
| 512 |
+
"logps/chosen": -519.8826904296875,
|
| 513 |
+
"logps/rejected": -489.1722717285156,
|
| 514 |
+
"loss": 0.0407,
|
| 515 |
+
"rewards/accuracies": 0.78125,
|
| 516 |
+
"rewards/chosen": -1.8402801752090454,
|
| 517 |
+
"rewards/margins": 1.171037197113037,
|
| 518 |
+
"rewards/rejected": -3.011317491531372,
|
| 519 |
+
"step": 320
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.37,
|
| 523 |
+
"learning_rate": 3.9469663422373864e-07,
|
| 524 |
+
"logits/chosen": -2.1483047008514404,
|
| 525 |
+
"logits/rejected": -2.095182418823242,
|
| 526 |
+
"logps/chosen": -515.9031372070312,
|
| 527 |
+
"logps/rejected": -511.58416748046875,
|
| 528 |
+
"loss": 0.0362,
|
| 529 |
+
"rewards/accuracies": 0.793749988079071,
|
| 530 |
+
"rewards/chosen": -1.987370252609253,
|
| 531 |
+
"rewards/margins": 1.1986653804779053,
|
| 532 |
+
"rewards/rejected": -3.186035633087158,
|
| 533 |
+
"step": 330
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.38,
|
| 537 |
+
"learning_rate": 3.865190451858954e-07,
|
| 538 |
+
"logits/chosen": -2.1236190795898438,
|
| 539 |
+
"logits/rejected": -2.038846492767334,
|
| 540 |
+
"logps/chosen": -567.5887451171875,
|
| 541 |
+
"logps/rejected": -531.2391357421875,
|
| 542 |
+
"loss": 0.0378,
|
| 543 |
+
"rewards/accuracies": 0.762499988079071,
|
| 544 |
+
"rewards/chosen": -2.1050643920898438,
|
| 545 |
+
"rewards/margins": 1.3257628679275513,
|
| 546 |
+
"rewards/rejected": -3.4308273792266846,
|
| 547 |
+
"step": 340
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 0.4,
|
| 551 |
+
"learning_rate": 3.781277606741327e-07,
|
| 552 |
+
"logits/chosen": -2.073742628097534,
|
| 553 |
+
"logits/rejected": -2.0120556354522705,
|
| 554 |
+
"logps/chosen": -546.9391479492188,
|
| 555 |
+
"logps/rejected": -522.168701171875,
|
| 556 |
+
"loss": 0.0363,
|
| 557 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 558 |
+
"rewards/chosen": -1.975027084350586,
|
| 559 |
+
"rewards/margins": 1.3364970684051514,
|
| 560 |
+
"rewards/rejected": -3.311523914337158,
|
| 561 |
+
"step": 350
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.41,
|
| 565 |
+
"learning_rate": 3.6953591570208996e-07,
|
| 566 |
+
"logits/chosen": -2.0751125812530518,
|
| 567 |
+
"logits/rejected": -1.9913240671157837,
|
| 568 |
+
"logps/chosen": -532.4324951171875,
|
| 569 |
+
"logps/rejected": -570.5128173828125,
|
| 570 |
+
"loss": 0.0339,
|
| 571 |
+
"rewards/accuracies": 0.762499988079071,
|
| 572 |
+
"rewards/chosen": -2.149209976196289,
|
| 573 |
+
"rewards/margins": 1.5435600280761719,
|
| 574 |
+
"rewards/rejected": -3.692770004272461,
|
| 575 |
+
"step": 360
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 0.42,
|
| 579 |
+
"learning_rate": 3.607569592239452e-07,
|
| 580 |
+
"logits/chosen": -2.0640273094177246,
|
| 581 |
+
"logits/rejected": -1.9882128238677979,
|
| 582 |
+
"logps/chosen": -538.2806396484375,
|
| 583 |
+
"logps/rejected": -503.5306701660156,
|
| 584 |
+
"loss": 0.0419,
|
| 585 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 586 |
+
"rewards/chosen": -1.8893718719482422,
|
| 587 |
+
"rewards/margins": 1.269718050956726,
|
| 588 |
+
"rewards/rejected": -3.159090042114258,
|
| 589 |
+
"step": 370
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 0.43,
|
| 593 |
+
"learning_rate": 3.518046330825494e-07,
|
| 594 |
+
"logits/chosen": -2.1525139808654785,
|
| 595 |
+
"logits/rejected": -2.09255051612854,
|
| 596 |
+
"logps/chosen": -541.6168823242188,
|
| 597 |
+
"logps/rejected": -525.7049560546875,
|
| 598 |
+
"loss": 0.0488,
|
| 599 |
+
"rewards/accuracies": 0.793749988079071,
|
| 600 |
+
"rewards/chosen": -1.759751319885254,
|
| 601 |
+
"rewards/margins": 1.318419098854065,
|
| 602 |
+
"rewards/rejected": -3.0781702995300293,
|
| 603 |
+
"step": 380
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.44,
|
| 607 |
+
"learning_rate": 3.4269295049909713e-07,
|
| 608 |
+
"logits/chosen": -2.111579179763794,
|
| 609 |
+
"logits/rejected": -2.008594512939453,
|
| 610 |
+
"logps/chosen": -519.3077392578125,
|
| 611 |
+
"logps/rejected": -495.54736328125,
|
| 612 |
+
"loss": 0.0479,
|
| 613 |
+
"rewards/accuracies": 0.71875,
|
| 614 |
+
"rewards/chosen": -1.9454580545425415,
|
| 615 |
+
"rewards/margins": 1.1256051063537598,
|
| 616 |
+
"rewards/rejected": -3.0710630416870117,
|
| 617 |
+
"step": 390
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.45,
|
| 621 |
+
"learning_rate": 3.3343617413800453e-07,
|
| 622 |
+
"logits/chosen": -2.116332530975342,
|
| 623 |
+
"logits/rejected": -2.0108766555786133,
|
| 624 |
+
"logps/chosen": -560.86083984375,
|
| 625 |
+
"logps/rejected": -513.68408203125,
|
| 626 |
+
"loss": 0.0401,
|
| 627 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 628 |
+
"rewards/chosen": -1.8742320537567139,
|
| 629 |
+
"rewards/margins": 1.4719246625900269,
|
| 630 |
+
"rewards/rejected": -3.346156597137451,
|
| 631 |
+
"step": 400
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 0.45,
|
| 635 |
+
"eval_logits/chosen": -2.0799078941345215,
|
| 636 |
+
"eval_logits/rejected": -2.042891502380371,
|
| 637 |
+
"eval_logps/chosen": -499.7916259765625,
|
| 638 |
+
"eval_logps/rejected": -568.111572265625,
|
| 639 |
+
"eval_loss": 0.052323117852211,
|
| 640 |
+
"eval_rewards/accuracies": 0.703125,
|
| 641 |
+
"eval_rewards/chosen": -2.4275197982788086,
|
| 642 |
+
"eval_rewards/margins": 0.6800626516342163,
|
| 643 |
+
"eval_rewards/rejected": -3.1075825691223145,
|
| 644 |
+
"eval_runtime": 53.2431,
|
| 645 |
+
"eval_samples_per_second": 37.564,
|
| 646 |
+
"eval_steps_per_second": 0.601,
|
| 647 |
+
"step": 400
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.46,
|
| 651 |
+
"learning_rate": 3.2404879378132893e-07,
|
| 652 |
+
"logits/chosen": -2.1349825859069824,
|
| 653 |
+
"logits/rejected": -2.0403599739074707,
|
| 654 |
+
"logps/chosen": -573.2166137695312,
|
| 655 |
+
"logps/rejected": -546.21484375,
|
| 656 |
+
"loss": 0.0393,
|
| 657 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 658 |
+
"rewards/chosen": -1.9816631078720093,
|
| 659 |
+
"rewards/margins": 1.3866467475891113,
|
| 660 |
+
"rewards/rejected": -3.3683102130889893,
|
| 661 |
+
"step": 410
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.48,
|
| 665 |
+
"learning_rate": 3.1454550364767894e-07,
|
| 666 |
+
"logits/chosen": -2.0241355895996094,
|
| 667 |
+
"logits/rejected": -1.955384612083435,
|
| 668 |
+
"logps/chosen": -587.6964721679688,
|
| 669 |
+
"logps/rejected": -564.29833984375,
|
| 670 |
+
"loss": 0.0336,
|
| 671 |
+
"rewards/accuracies": 0.731249988079071,
|
| 672 |
+
"rewards/chosen": -2.4298739433288574,
|
| 673 |
+
"rewards/margins": 1.2927566766738892,
|
| 674 |
+
"rewards/rejected": -3.722630262374878,
|
| 675 |
+
"step": 420
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.49,
|
| 679 |
+
"learning_rate": 3.049411793911154e-07,
|
| 680 |
+
"logits/chosen": -2.0459389686584473,
|
| 681 |
+
"logits/rejected": -1.9516912698745728,
|
| 682 |
+
"logps/chosen": -587.4751586914062,
|
| 683 |
+
"logps/rejected": -571.4771728515625,
|
| 684 |
+
"loss": 0.0377,
|
| 685 |
+
"rewards/accuracies": 0.731249988079071,
|
| 686 |
+
"rewards/chosen": -2.2811179161071777,
|
| 687 |
+
"rewards/margins": 1.313342571258545,
|
| 688 |
+
"rewards/rejected": -3.5944607257843018,
|
| 689 |
+
"step": 430
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.5,
|
| 693 |
+
"learning_rate": 2.9525085481604914e-07,
|
| 694 |
+
"logits/chosen": -2.076460361480713,
|
| 695 |
+
"logits/rejected": -2.008845090866089,
|
| 696 |
+
"logps/chosen": -501.68743896484375,
|
| 697 |
+
"logps/rejected": -492.162841796875,
|
| 698 |
+
"loss": 0.0406,
|
| 699 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 700 |
+
"rewards/chosen": -1.8489151000976562,
|
| 701 |
+
"rewards/margins": 1.1866779327392578,
|
| 702 |
+
"rewards/rejected": -3.035592794418335,
|
| 703 |
+
"step": 440
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.51,
|
| 707 |
+
"learning_rate": 2.854896983445833e-07,
|
| 708 |
+
"logits/chosen": -2.0890183448791504,
|
| 709 |
+
"logits/rejected": -1.99982488155365,
|
| 710 |
+
"logps/chosen": -555.748291015625,
|
| 711 |
+
"logps/rejected": -527.487060546875,
|
| 712 |
+
"loss": 0.0446,
|
| 713 |
+
"rewards/accuracies": 0.706250011920929,
|
| 714 |
+
"rewards/chosen": -2.0640885829925537,
|
| 715 |
+
"rewards/margins": 1.2410598993301392,
|
| 716 |
+
"rewards/rejected": -3.3051486015319824,
|
| 717 |
+
"step": 450
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.52,
|
| 721 |
+
"learning_rate": 2.7567298927313654e-07,
|
| 722 |
+
"logits/chosen": -2.0581395626068115,
|
| 723 |
+
"logits/rejected": -1.9576694965362549,
|
| 724 |
+
"logps/chosen": -517.2969970703125,
|
| 725 |
+
"logps/rejected": -494.01580810546875,
|
| 726 |
+
"loss": 0.0444,
|
| 727 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 728 |
+
"rewards/chosen": -1.7231872081756592,
|
| 729 |
+
"rewards/margins": 1.3869811296463013,
|
| 730 |
+
"rewards/rejected": -3.110168695449829,
|
| 731 |
+
"step": 460
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.53,
|
| 735 |
+
"learning_rate": 2.658160938555123e-07,
|
| 736 |
+
"logits/chosen": -2.0949597358703613,
|
| 737 |
+
"logits/rejected": -1.978247046470642,
|
| 738 |
+
"logps/chosen": -530.58642578125,
|
| 739 |
+
"logps/rejected": -518.8834838867188,
|
| 740 |
+
"loss": 0.0408,
|
| 741 |
+
"rewards/accuracies": 0.75,
|
| 742 |
+
"rewards/chosen": -1.8325612545013428,
|
| 743 |
+
"rewards/margins": 1.4025018215179443,
|
| 744 |
+
"rewards/rejected": -3.235063076019287,
|
| 745 |
+
"step": 470
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.54,
|
| 749 |
+
"learning_rate": 2.559344412498532e-07,
|
| 750 |
+
"logits/chosen": -2.0444254875183105,
|
| 751 |
+
"logits/rejected": -1.958059310913086,
|
| 752 |
+
"logps/chosen": -538.9998779296875,
|
| 753 |
+
"logps/rejected": -531.4126586914062,
|
| 754 |
+
"loss": 0.0282,
|
| 755 |
+
"rewards/accuracies": 0.793749988079071,
|
| 756 |
+
"rewards/chosen": -1.9186712503433228,
|
| 757 |
+
"rewards/margins": 1.5285937786102295,
|
| 758 |
+
"rewards/rejected": -3.447265148162842,
|
| 759 |
+
"step": 480
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.55,
|
| 763 |
+
"learning_rate": 2.460434993671294e-07,
|
| 764 |
+
"logits/chosen": -2.060708522796631,
|
| 765 |
+
"logits/rejected": -1.9541358947753906,
|
| 766 |
+
"logps/chosen": -524.92431640625,
|
| 767 |
+
"logps/rejected": -490.29669189453125,
|
| 768 |
+
"loss": 0.0367,
|
| 769 |
+
"rewards/accuracies": 0.75,
|
| 770 |
+
"rewards/chosen": -1.956221342086792,
|
| 771 |
+
"rewards/margins": 1.1774486303329468,
|
| 772 |
+
"rewards/rejected": -3.1336700916290283,
|
| 773 |
+
"step": 490
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.57,
|
| 777 |
+
"learning_rate": 2.361587506589672e-07,
|
| 778 |
+
"logits/chosen": -2.0169599056243896,
|
| 779 |
+
"logits/rejected": -1.9417072534561157,
|
| 780 |
+
"logps/chosen": -512.1517333984375,
|
| 781 |
+
"logps/rejected": -506.2262268066406,
|
| 782 |
+
"loss": 0.0335,
|
| 783 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 784 |
+
"rewards/chosen": -1.9970229864120483,
|
| 785 |
+
"rewards/margins": 1.3175480365753174,
|
| 786 |
+
"rewards/rejected": -3.314570665359497,
|
| 787 |
+
"step": 500
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.57,
|
| 791 |
+
"eval_logits/chosen": -2.0455641746520996,
|
| 792 |
+
"eval_logits/rejected": -2.005706310272217,
|
| 793 |
+
"eval_logps/chosen": -497.6726989746094,
|
| 794 |
+
"eval_logps/rejected": -580.1129150390625,
|
| 795 |
+
"eval_loss": 0.04609997943043709,
|
| 796 |
+
"eval_rewards/accuracies": 0.71484375,
|
| 797 |
+
"eval_rewards/chosen": -2.4063305854797363,
|
| 798 |
+
"eval_rewards/margins": 0.8212659358978271,
|
| 799 |
+
"eval_rewards/rejected": -3.2275965213775635,
|
| 800 |
+
"eval_runtime": 53.2473,
|
| 801 |
+
"eval_samples_per_second": 37.561,
|
| 802 |
+
"eval_steps_per_second": 0.601,
|
| 803 |
+
"step": 500
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.58,
|
| 807 |
+
"learning_rate": 2.2629566788271613e-07,
|
| 808 |
+
"logits/chosen": -2.0371432304382324,
|
| 809 |
+
"logits/rejected": -1.9557580947875977,
|
| 810 |
+
"logps/chosen": -560.0842895507812,
|
| 811 |
+
"logps/rejected": -532.668212890625,
|
| 812 |
+
"loss": 0.038,
|
| 813 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 814 |
+
"rewards/chosen": -1.9359846115112305,
|
| 815 |
+
"rewards/margins": 1.3490617275238037,
|
| 816 |
+
"rewards/rejected": -3.285046339035034,
|
| 817 |
+
"step": 510
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.59,
|
| 821 |
+
"learning_rate": 2.1646968988169135e-07,
|
| 822 |
+
"logits/chosen": -2.011871814727783,
|
| 823 |
+
"logits/rejected": -1.9335308074951172,
|
| 824 |
+
"logps/chosen": -507.52337646484375,
|
| 825 |
+
"logps/rejected": -516.07958984375,
|
| 826 |
+
"loss": 0.0395,
|
| 827 |
+
"rewards/accuracies": 0.8125,
|
| 828 |
+
"rewards/chosen": -1.9623944759368896,
|
| 829 |
+
"rewards/margins": 1.2198023796081543,
|
| 830 |
+
"rewards/rejected": -3.182196855545044,
|
| 831 |
+
"step": 520
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.6,
|
| 835 |
+
"learning_rate": 2.0669619741850232e-07,
|
| 836 |
+
"logits/chosen": -1.9674116373062134,
|
| 837 |
+
"logits/rejected": -1.874136209487915,
|
| 838 |
+
"logps/chosen": -560.131591796875,
|
| 839 |
+
"logps/rejected": -497.582275390625,
|
| 840 |
+
"loss": 0.0333,
|
| 841 |
+
"rewards/accuracies": 0.75,
|
| 842 |
+
"rewards/chosen": -2.1217665672302246,
|
| 843 |
+
"rewards/margins": 1.2892667055130005,
|
| 844 |
+
"rewards/rejected": -3.4110336303710938,
|
| 845 |
+
"step": 530
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.61,
|
| 849 |
+
"learning_rate": 1.9699048909929518e-07,
|
| 850 |
+
"logits/chosen": -1.9731553792953491,
|
| 851 |
+
"logits/rejected": -1.9132474660873413,
|
| 852 |
+
"logps/chosen": -545.8017578125,
|
| 853 |
+
"logps/rejected": -487.0414123535156,
|
| 854 |
+
"loss": 0.0348,
|
| 855 |
+
"rewards/accuracies": 0.737500011920929,
|
| 856 |
+
"rewards/chosen": -2.1676135063171387,
|
| 857 |
+
"rewards/margins": 1.0356405973434448,
|
| 858 |
+
"rewards/rejected": -3.203253984451294,
|
| 859 |
+
"step": 540
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.62,
|
| 863 |
+
"learning_rate": 1.8736775742659732e-07,
|
| 864 |
+
"logits/chosen": -1.9223613739013672,
|
| 865 |
+
"logits/rejected": -1.8391857147216797,
|
| 866 |
+
"logps/chosen": -517.1666870117188,
|
| 867 |
+
"logps/rejected": -498.1993713378906,
|
| 868 |
+
"loss": 0.0342,
|
| 869 |
+
"rewards/accuracies": 0.762499988079071,
|
| 870 |
+
"rewards/chosen": -1.948054313659668,
|
| 871 |
+
"rewards/margins": 1.2562531232833862,
|
| 872 |
+
"rewards/rejected": -3.2043070793151855,
|
| 873 |
+
"step": 550
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.63,
|
| 877 |
+
"learning_rate": 1.7784306501824616e-07,
|
| 878 |
+
"logits/chosen": -1.9654014110565186,
|
| 879 |
+
"logits/rejected": -1.8999271392822266,
|
| 880 |
+
"logps/chosen": -574.9034423828125,
|
| 881 |
+
"logps/rejected": -553.96240234375,
|
| 882 |
+
"loss": 0.0336,
|
| 883 |
+
"rewards/accuracies": 0.768750011920929,
|
| 884 |
+
"rewards/chosen": -2.0827620029449463,
|
| 885 |
+
"rewards/margins": 1.3543148040771484,
|
| 886 |
+
"rewards/rejected": -3.4370765686035156,
|
| 887 |
+
"step": 560
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.65,
|
| 891 |
+
"learning_rate": 1.6843132102963025e-07,
|
| 892 |
+
"logits/chosen": -1.9901962280273438,
|
| 893 |
+
"logits/rejected": -1.8920910358428955,
|
| 894 |
+
"logps/chosen": -571.3399658203125,
|
| 895 |
+
"logps/rejected": -549.0781860351562,
|
| 896 |
+
"loss": 0.0364,
|
| 897 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 898 |
+
"rewards/chosen": -2.1259512901306152,
|
| 899 |
+
"rewards/margins": 1.4393842220306396,
|
| 900 |
+
"rewards/rejected": -3.565335512161255,
|
| 901 |
+
"step": 570
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.66,
|
| 905 |
+
"learning_rate": 1.591472578161458e-07,
|
| 906 |
+
"logits/chosen": -2.0036253929138184,
|
| 907 |
+
"logits/rejected": -1.9164316654205322,
|
| 908 |
+
"logps/chosen": -597.6680908203125,
|
| 909 |
+
"logps/rejected": -563.1033935546875,
|
| 910 |
+
"loss": 0.0314,
|
| 911 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 912 |
+
"rewards/chosen": -2.2230658531188965,
|
| 913 |
+
"rewards/margins": 1.4771394729614258,
|
| 914 |
+
"rewards/rejected": -3.7002053260803223,
|
| 915 |
+
"step": 580
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.67,
|
| 919 |
+
"learning_rate": 1.5000540787240274e-07,
|
| 920 |
+
"logits/chosen": -1.9651165008544922,
|
| 921 |
+
"logits/rejected": -1.9028050899505615,
|
| 922 |
+
"logps/chosen": -573.396728515625,
|
| 923 |
+
"logps/rejected": -580.0013427734375,
|
| 924 |
+
"loss": 0.0282,
|
| 925 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 926 |
+
"rewards/chosen": -2.4309842586517334,
|
| 927 |
+
"rewards/margins": 1.3270018100738525,
|
| 928 |
+
"rewards/rejected": -3.757986068725586,
|
| 929 |
+
"step": 590
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.68,
|
| 933 |
+
"learning_rate": 1.410200810842749e-07,
|
| 934 |
+
"logits/chosen": -1.9549171924591064,
|
| 935 |
+
"logits/rejected": -1.8664356470108032,
|
| 936 |
+
"logps/chosen": -546.6082153320312,
|
| 937 |
+
"logps/rejected": -557.6231689453125,
|
| 938 |
+
"loss": 0.0273,
|
| 939 |
+
"rewards/accuracies": 0.731249988079071,
|
| 940 |
+
"rewards/chosen": -2.497020959854126,
|
| 941 |
+
"rewards/margins": 1.3337533473968506,
|
| 942 |
+
"rewards/rejected": -3.8307743072509766,
|
| 943 |
+
"step": 600
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.68,
|
| 947 |
+
"eval_logits/chosen": -1.9557578563690186,
|
| 948 |
+
"eval_logits/rejected": -1.9161585569381714,
|
| 949 |
+
"eval_logps/chosen": -541.6861572265625,
|
| 950 |
+
"eval_logps/rejected": -628.8740844726562,
|
| 951 |
+
"eval_loss": 0.04087229445576668,
|
| 952 |
+
"eval_rewards/accuracies": 0.70703125,
|
| 953 |
+
"eval_rewards/chosen": -2.8464653491973877,
|
| 954 |
+
"eval_rewards/margins": 0.8687426447868347,
|
| 955 |
+
"eval_rewards/rejected": -3.715208053588867,
|
| 956 |
+
"eval_runtime": 53.2271,
|
| 957 |
+
"eval_samples_per_second": 37.575,
|
| 958 |
+
"eval_steps_per_second": 0.601,
|
| 959 |
+
"step": 600
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.69,
|
| 963 |
+
"learning_rate": 1.322053423294041e-07,
|
| 964 |
+
"logits/chosen": -1.9264421463012695,
|
| 965 |
+
"logits/rejected": -1.8596795797348022,
|
| 966 |
+
"logps/chosen": -555.9921875,
|
| 967 |
+
"logps/rejected": -528.2036743164062,
|
| 968 |
+
"loss": 0.0364,
|
| 969 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 970 |
+
"rewards/chosen": -2.3642823696136475,
|
| 971 |
+
"rewards/margins": 1.1525689363479614,
|
| 972 |
+
"rewards/rejected": -3.5168514251708984,
|
| 973 |
+
"step": 610
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 0.7,
|
| 977 |
+
"learning_rate": 1.2357498946121905e-07,
|
| 978 |
+
"logits/chosen": -2.047217845916748,
|
| 979 |
+
"logits/rejected": -1.9547252655029297,
|
| 980 |
+
"logps/chosen": -573.0296630859375,
|
| 981 |
+
"logps/rejected": -531.27880859375,
|
| 982 |
+
"loss": 0.04,
|
| 983 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 984 |
+
"rewards/chosen": -2.1570980548858643,
|
| 985 |
+
"rewards/margins": 1.236627459526062,
|
| 986 |
+
"rewards/rejected": -3.393725633621216,
|
| 987 |
+
"step": 620
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.71,
|
| 991 |
+
"learning_rate": 1.1514253171093161e-07,
|
| 992 |
+
"logits/chosen": -2.0023763179779053,
|
| 993 |
+
"logits/rejected": -1.9147300720214844,
|
| 994 |
+
"logps/chosen": -558.9900512695312,
|
| 995 |
+
"logps/rejected": -542.7515869140625,
|
| 996 |
+
"loss": 0.0424,
|
| 997 |
+
"rewards/accuracies": 0.75,
|
| 998 |
+
"rewards/chosen": -2.013390064239502,
|
| 999 |
+
"rewards/margins": 1.3009288311004639,
|
| 1000 |
+
"rewards/rejected": -3.314318895339966,
|
| 1001 |
+
"step": 630
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 0.72,
|
| 1005 |
+
"learning_rate": 1.0692116854131883e-07,
|
| 1006 |
+
"logits/chosen": -1.9406598806381226,
|
| 1007 |
+
"logits/rejected": -1.8927663564682007,
|
| 1008 |
+
"logps/chosen": -502.59344482421875,
|
| 1009 |
+
"logps/rejected": -507.50372314453125,
|
| 1010 |
+
"loss": 0.0437,
|
| 1011 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1012 |
+
"rewards/chosen": -2.104419231414795,
|
| 1013 |
+
"rewards/margins": 1.1473711729049683,
|
| 1014 |
+
"rewards/rejected": -3.251791000366211,
|
| 1015 |
+
"step": 640
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 0.74,
|
| 1019 |
+
"learning_rate": 9.89237689853889e-08,
|
| 1020 |
+
"logits/chosen": -1.9217647314071655,
|
| 1021 |
+
"logits/rejected": -1.840735673904419,
|
| 1022 |
+
"logps/chosen": -539.6429443359375,
|
| 1023 |
+
"logps/rejected": -520.7713623046875,
|
| 1024 |
+
"loss": 0.0348,
|
| 1025 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1026 |
+
"rewards/chosen": -2.0584945678710938,
|
| 1027 |
+
"rewards/margins": 1.4084278345108032,
|
| 1028 |
+
"rewards/rejected": -3.4669222831726074,
|
| 1029 |
+
"step": 650
|
| 1030 |
+
},
|
| 1031 |
{
|
| 1032 |
"epoch": 0.75,
|
| 1033 |
+
"learning_rate": 9.11628515022765e-08,
|
| 1034 |
+
"logits/chosen": -1.936431884765625,
|
| 1035 |
+
"logits/rejected": -1.8774102926254272,
|
| 1036 |
+
"logps/chosen": -507.42413330078125,
|
| 1037 |
+
"logps/rejected": -489.027099609375,
|
| 1038 |
+
"loss": 0.0321,
|
| 1039 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1040 |
+
"rewards/chosen": -1.9116840362548828,
|
| 1041 |
+
"rewards/margins": 1.3113796710968018,
|
| 1042 |
+
"rewards/rejected": -3.2230639457702637,
|
| 1043 |
+
"step": 660
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 0.76,
|
| 1047 |
+
"learning_rate": 8.365056438189486e-08,
|
| 1048 |
+
"logits/chosen": -1.9381834268569946,
|
| 1049 |
+
"logits/rejected": -1.8834428787231445,
|
| 1050 |
+
"logps/chosen": -545.0704345703125,
|
| 1051 |
+
"logps/rejected": -511.5694885253906,
|
| 1052 |
+
"loss": 0.0361,
|
| 1053 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1054 |
+
"rewards/chosen": -2.27673602104187,
|
| 1055 |
+
"rewards/margins": 1.0493533611297607,
|
| 1056 |
+
"rewards/rejected": -3.326089382171631,
|
| 1057 |
+
"step": 670
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 0.77,
|
| 1061 |
+
"learning_rate": 7.639866672902101e-08,
|
| 1062 |
+
"logits/chosen": -1.9358975887298584,
|
| 1063 |
+
"logits/rejected": -1.8624534606933594,
|
| 1064 |
+
"logps/chosen": -552.2891845703125,
|
| 1065 |
+
"logps/rejected": -548.515625,
|
| 1066 |
+
"loss": 0.0358,
|
| 1067 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1068 |
+
"rewards/chosen": -2.3348231315612793,
|
| 1069 |
+
"rewards/margins": 1.1951119899749756,
|
| 1070 |
+
"rewards/rejected": -3.529935121536255,
|
| 1071 |
+
"step": 680
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.78,
|
| 1075 |
+
"learning_rate": 6.941851005657851e-08,
|
| 1076 |
+
"logits/chosen": -1.9736747741699219,
|
| 1077 |
+
"logits/rejected": -1.8825995922088623,
|
| 1078 |
+
"logps/chosen": -592.0359497070312,
|
| 1079 |
+
"logps/rejected": -565.7269897460938,
|
| 1080 |
+
"loss": 0.0392,
|
| 1081 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1082 |
+
"rewards/chosen": -2.2588744163513184,
|
| 1083 |
+
"rewards/margins": 1.3530534505844116,
|
| 1084 |
+
"rewards/rejected": -3.6119277477264404,
|
| 1085 |
+
"step": 690
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.79,
|
| 1089 |
+
"learning_rate": 6.272102051693051e-08,
|
| 1090 |
+
"logits/chosen": -1.9536447525024414,
|
| 1091 |
+
"logits/rejected": -1.8848438262939453,
|
| 1092 |
+
"logps/chosen": -544.1442260742188,
|
| 1093 |
+
"logps/rejected": -547.0964965820312,
|
| 1094 |
+
"loss": 0.0377,
|
| 1095 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1096 |
+
"rewards/chosen": -2.0580058097839355,
|
| 1097 |
+
"rewards/margins": 1.3238760232925415,
|
| 1098 |
+
"rewards/rejected": -3.3818821907043457,
|
| 1099 |
+
"step": 700
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.79,
|
| 1103 |
+
"eval_logits/chosen": -1.967289686203003,
|
| 1104 |
+
"eval_logits/rejected": -1.9273663759231567,
|
| 1105 |
+
"eval_logps/chosen": -510.2101745605469,
|
| 1106 |
+
"eval_logps/rejected": -594.1712036132812,
|
| 1107 |
+
"eval_loss": 0.0495593398809433,
|
| 1108 |
+
"eval_rewards/accuracies": 0.72265625,
|
| 1109 |
+
"eval_rewards/chosen": -2.531705617904663,
|
| 1110 |
+
"eval_rewards/margins": 0.836473286151886,
|
| 1111 |
+
"eval_rewards/rejected": -3.3681788444519043,
|
| 1112 |
+
"eval_runtime": 53.2631,
|
| 1113 |
+
"eval_samples_per_second": 37.549,
|
| 1114 |
+
"eval_steps_per_second": 0.601,
|
| 1115 |
+
"step": 700
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.8,
|
| 1119 |
+
"learning_rate": 5.6316681798995844e-08,
|
| 1120 |
+
"logits/chosen": -1.9590046405792236,
|
| 1121 |
+
"logits/rejected": -1.8676058053970337,
|
| 1122 |
+
"logps/chosen": -530.419921875,
|
| 1123 |
+
"logps/rejected": -494.63720703125,
|
| 1124 |
+
"loss": 0.0395,
|
| 1125 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1126 |
+
"rewards/chosen": -1.950580358505249,
|
| 1127 |
+
"rewards/margins": 1.374487280845642,
|
| 1128 |
+
"rewards/rejected": -3.3250679969787598,
|
| 1129 |
+
"step": 710
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.81,
|
| 1133 |
+
"learning_rate": 5.0215518717961256e-08,
|
| 1134 |
+
"logits/chosen": -1.9477427005767822,
|
| 1135 |
+
"logits/rejected": -1.9216333627700806,
|
| 1136 |
+
"logps/chosen": -568.46337890625,
|
| 1137 |
+
"logps/rejected": -551.7440185546875,
|
| 1138 |
+
"loss": 0.0382,
|
| 1139 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1140 |
+
"rewards/chosen": -2.0898067951202393,
|
| 1141 |
+
"rewards/margins": 1.2489385604858398,
|
| 1142 |
+
"rewards/rejected": -3.3387451171875,
|
| 1143 |
+
"step": 720
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.83,
|
| 1147 |
+
"learning_rate": 4.4427081523275925e-08,
|
| 1148 |
+
"logits/chosen": -1.9811092615127563,
|
| 1149 |
+
"logits/rejected": -1.8924198150634766,
|
| 1150 |
+
"logps/chosen": -560.4685668945312,
|
| 1151 |
+
"logps/rejected": -545.5777587890625,
|
| 1152 |
+
"loss": 0.0384,
|
| 1153 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1154 |
+
"rewards/chosen": -2.166611433029175,
|
| 1155 |
+
"rewards/margins": 1.3478820323944092,
|
| 1156 |
+
"rewards/rejected": -3.514493465423584,
|
| 1157 |
+
"step": 730
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.84,
|
| 1161 |
+
"learning_rate": 3.896043094949061e-08,
|
| 1162 |
+
"logits/chosen": -1.952548623085022,
|
| 1163 |
+
"logits/rejected": -1.9003283977508545,
|
| 1164 |
+
"logps/chosen": -536.4952392578125,
|
| 1165 |
+
"logps/rejected": -554.130859375,
|
| 1166 |
+
"loss": 0.0368,
|
| 1167 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1168 |
+
"rewards/chosen": -2.2593836784362793,
|
| 1169 |
+
"rewards/margins": 1.1702790260314941,
|
| 1170 |
+
"rewards/rejected": -3.4296627044677734,
|
| 1171 |
+
"step": 740
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.85,
|
| 1175 |
+
"learning_rate": 3.3824124033343557e-08,
|
| 1176 |
+
"logits/chosen": -1.978271245956421,
|
| 1177 |
+
"logits/rejected": -1.8869798183441162,
|
| 1178 |
+
"logps/chosen": -595.8272094726562,
|
| 1179 |
+
"logps/rejected": -596.80908203125,
|
| 1180 |
+
"loss": 0.0356,
|
| 1181 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1182 |
+
"rewards/chosen": -2.2419838905334473,
|
| 1183 |
+
"rewards/margins": 1.5834529399871826,
|
| 1184 |
+
"rewards/rejected": -3.825437068939209,
|
| 1185 |
+
"step": 750
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.86,
|
| 1189 |
+
"learning_rate": 2.9026200719291904e-08,
|
| 1190 |
+
"logits/chosen": -1.9425359964370728,
|
| 1191 |
+
"logits/rejected": -1.8644979000091553,
|
| 1192 |
+
"logps/chosen": -541.2769775390625,
|
| 1193 |
+
"logps/rejected": -546.4044799804688,
|
| 1194 |
+
"loss": 0.0394,
|
| 1195 |
"rewards/accuracies": 0.699999988079071,
|
| 1196 |
+
"rewards/chosen": -2.3050169944763184,
|
| 1197 |
+
"rewards/margins": 1.2533900737762451,
|
| 1198 |
+
"rewards/rejected": -3.5584073066711426,
|
| 1199 |
+
"step": 760
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.87,
|
| 1203 |
+
"learning_rate": 2.4574171274456433e-08,
|
| 1204 |
+
"logits/chosen": -1.9859609603881836,
|
| 1205 |
+
"logits/rejected": -1.8982467651367188,
|
| 1206 |
+
"logps/chosen": -581.1209716796875,
|
| 1207 |
+
"logps/rejected": -553.73291015625,
|
| 1208 |
+
"loss": 0.036,
|
| 1209 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1210 |
+
"rewards/chosen": -1.9945024251937866,
|
| 1211 |
+
"rewards/margins": 1.5215613842010498,
|
| 1212 |
+
"rewards/rejected": -3.516064167022705,
|
| 1213 |
+
"step": 770
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.88,
|
| 1217 |
+
"learning_rate": 2.047500453267881e-08,
|
| 1218 |
+
"logits/chosen": -1.9515501260757446,
|
| 1219 |
+
"logits/rejected": -1.857346534729004,
|
| 1220 |
+
"logps/chosen": -565.3553466796875,
|
| 1221 |
+
"logps/rejected": -536.8668823242188,
|
| 1222 |
+
"loss": 0.0422,
|
| 1223 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1224 |
+
"rewards/chosen": -2.2416954040527344,
|
| 1225 |
+
"rewards/margins": 1.2734299898147583,
|
| 1226 |
+
"rewards/rejected": -3.515125274658203,
|
| 1227 |
+
"step": 780
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.89,
|
| 1231 |
+
"learning_rate": 1.673511698609292e-08,
|
| 1232 |
+
"logits/chosen": -1.9258711338043213,
|
| 1233 |
+
"logits/rejected": -1.8735402822494507,
|
| 1234 |
+
"logps/chosen": -532.5035400390625,
|
| 1235 |
+
"logps/rejected": -557.9683837890625,
|
| 1236 |
+
"loss": 0.0354,
|
| 1237 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1238 |
+
"rewards/chosen": -2.2362167835235596,
|
| 1239 |
+
"rewards/margins": 1.4313932657241821,
|
| 1240 |
+
"rewards/rejected": -3.6676101684570312,
|
| 1241 |
+
"step": 790
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.91,
|
| 1245 |
+
"learning_rate": 1.3360362741285769e-08,
|
| 1246 |
+
"logits/chosen": -1.9208987951278687,
|
| 1247 |
+
"logits/rejected": -1.8273556232452393,
|
| 1248 |
+
"logps/chosen": -566.8057861328125,
|
| 1249 |
+
"logps/rejected": -583.4982299804688,
|
| 1250 |
+
"loss": 0.0352,
|
| 1251 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1252 |
+
"rewards/chosen": -2.093482732772827,
|
| 1253 |
+
"rewards/margins": 1.81829035282135,
|
| 1254 |
+
"rewards/rejected": -3.911773681640625,
|
| 1255 |
+
"step": 800
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.91,
|
| 1259 |
+
"eval_logits/chosen": -1.9501041173934937,
|
| 1260 |
+
"eval_logits/rejected": -1.9090631008148193,
|
| 1261 |
+
"eval_logps/chosen": -521.0439453125,
|
| 1262 |
+
"eval_logps/rejected": -606.3505249023438,
|
| 1263 |
+
"eval_loss": 0.04646110162138939,
|
| 1264 |
+
"eval_rewards/accuracies": 0.72265625,
|
| 1265 |
+
"eval_rewards/chosen": -2.640043258666992,
|
| 1266 |
+
"eval_rewards/margins": 0.8499288558959961,
|
| 1267 |
+
"eval_rewards/rejected": -3.4899723529815674,
|
| 1268 |
+
"eval_runtime": 53.2044,
|
| 1269 |
+
"eval_samples_per_second": 37.591,
|
| 1270 |
+
"eval_steps_per_second": 0.601,
|
| 1271 |
+
"step": 800
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 0.92,
|
| 1275 |
+
"learning_rate": 1.0356024355769433e-08,
|
| 1276 |
+
"logits/chosen": -1.950596570968628,
|
| 1277 |
+
"logits/rejected": -1.882962942123413,
|
| 1278 |
+
"logps/chosen": -582.3919067382812,
|
| 1279 |
+
"logps/rejected": -592.8079223632812,
|
| 1280 |
+
"loss": 0.0361,
|
| 1281 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1282 |
+
"rewards/chosen": -2.399203062057495,
|
| 1283 |
+
"rewards/margins": 1.3655624389648438,
|
| 1284 |
+
"rewards/rejected": -3.7647652626037598,
|
| 1285 |
+
"step": 810
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.93,
|
| 1289 |
+
"learning_rate": 7.726804569108597e-09,
|
| 1290 |
+
"logits/chosen": -1.9748871326446533,
|
| 1291 |
+
"logits/rejected": -1.888811707496643,
|
| 1292 |
+
"logps/chosen": -553.1395263671875,
|
| 1293 |
+
"logps/rejected": -559.5152587890625,
|
| 1294 |
+
"loss": 0.0332,
|
| 1295 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1296 |
+
"rewards/chosen": -2.228854179382324,
|
| 1297 |
+
"rewards/margins": 1.4253013134002686,
|
| 1298 |
+
"rewards/rejected": -3.654155731201172,
|
| 1299 |
+
"step": 820
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
"epoch": 0.94,
|
| 1303 |
+
"learning_rate": 5.476818941645561e-09,
|
| 1304 |
+
"logits/chosen": -1.944598913192749,
|
| 1305 |
+
"logits/rejected": -1.8884025812149048,
|
| 1306 |
+
"logps/chosen": -540.9061279296875,
|
| 1307 |
+
"logps/rejected": -574.0789794921875,
|
| 1308 |
+
"loss": 0.036,
|
| 1309 |
+
"rewards/accuracies": 0.75,
|
| 1310 |
+
"rewards/chosen": -2.298133134841919,
|
| 1311 |
+
"rewards/margins": 1.5589947700500488,
|
| 1312 |
+
"rewards/rejected": -3.8571276664733887,
|
| 1313 |
+
"step": 830
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.95,
|
| 1317 |
+
"learning_rate": 3.609589412347347e-09,
|
| 1318 |
+
"logits/chosen": -1.995321273803711,
|
| 1319 |
+
"logits/rejected": -1.909641981124878,
|
| 1320 |
+
"logps/chosen": -570.28369140625,
|
| 1321 |
+
"logps/rejected": -541.8267211914062,
|
| 1322 |
+
"loss": 0.0347,
|
| 1323 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1324 |
+
"rewards/chosen": -2.329395055770874,
|
| 1325 |
+
"rewards/margins": 1.3247489929199219,
|
| 1326 |
+
"rewards/rejected": -3.654143810272217,
|
| 1327 |
+
"step": 840
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 0.96,
|
| 1331 |
+
"learning_rate": 2.1280387858572667e-09,
|
| 1332 |
+
"logits/chosen": -1.9664127826690674,
|
| 1333 |
+
"logits/rejected": -1.8978990316390991,
|
| 1334 |
+
"logps/chosen": -567.4363403320312,
|
| 1335 |
+
"logps/rejected": -551.3905639648438,
|
| 1336 |
+
"loss": 0.0347,
|
| 1337 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1338 |
+
"rewards/chosen": -2.342649459838867,
|
| 1339 |
+
"rewards/margins": 1.2894551753997803,
|
| 1340 |
+
"rewards/rejected": -3.6321043968200684,
|
| 1341 |
+
"step": 850
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 0.97,
|
| 1345 |
+
"learning_rate": 1.03448615738172e-09,
|
| 1346 |
+
"logits/chosen": -1.91958749294281,
|
| 1347 |
+
"logits/rejected": -1.8549039363861084,
|
| 1348 |
+
"logps/chosen": -635.5400390625,
|
| 1349 |
+
"logps/rejected": -613.4205322265625,
|
| 1350 |
+
"loss": 0.0354,
|
| 1351 |
+
"rewards/accuracies": 0.84375,
|
| 1352 |
+
"rewards/chosen": -2.2879905700683594,
|
| 1353 |
+
"rewards/margins": 1.6663434505462646,
|
| 1354 |
+
"rewards/rejected": -3.954333782196045,
|
| 1355 |
+
"step": 860
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 0.98,
|
| 1359 |
+
"learning_rate": 3.3064328257259575e-10,
|
| 1360 |
+
"logits/chosen": -1.9804208278656006,
|
| 1361 |
+
"logits/rejected": -1.891332983970642,
|
| 1362 |
+
"logps/chosen": -612.157958984375,
|
| 1363 |
+
"logps/rejected": -556.913818359375,
|
| 1364 |
+
"loss": 0.0366,
|
| 1365 |
+
"rewards/accuracies": 0.75,
|
| 1366 |
+
"rewards/chosen": -2.3392751216888428,
|
| 1367 |
+
"rewards/margins": 1.2889056205749512,
|
| 1368 |
+
"rewards/rejected": -3.628180980682373,
|
| 1369 |
+
"step": 870
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 1.0,
|
| 1373 |
+
"learning_rate": 1.7611898088715216e-11,
|
| 1374 |
+
"logits/chosen": -1.952614426612854,
|
| 1375 |
+
"logits/rejected": -1.846652626991272,
|
| 1376 |
+
"logps/chosen": -577.5618896484375,
|
| 1377 |
+
"logps/rejected": -596.773193359375,
|
| 1378 |
+
"loss": 0.0347,
|
| 1379 |
+
"rewards/accuracies": 0.8125,
|
| 1380 |
+
"rewards/chosen": -2.3046746253967285,
|
| 1381 |
+
"rewards/margins": 1.4788968563079834,
|
| 1382 |
+
"rewards/rejected": -3.783571720123291,
|
| 1383 |
+
"step": 880
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
"epoch": 1.0,
|
| 1387 |
+
"step": 883,
|
| 1388 |
"total_flos": 0.0,
|
| 1389 |
+
"train_loss": 0.06580274857555349,
|
| 1390 |
+
"train_runtime": 7965.4661,
|
| 1391 |
+
"train_samples_per_second": 14.19,
|
| 1392 |
+
"train_steps_per_second": 0.111
|
| 1393 |
}
|
| 1394 |
],
|
| 1395 |
"logging_steps": 10,
|
| 1396 |
+
"max_steps": 883,
|
| 1397 |
"num_train_epochs": 1,
|
| 1398 |
"save_steps": 100,
|
| 1399 |
"total_flos": 0.0,
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5944
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b81f2d5981de606997c389c9fb0580152916e09b6c995a6e72c4de6d634e539
|
| 3 |
size 5944
|