Model save
Browse files- README.md +60 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +945 -0
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: mistralai/Mistral-7B-Instruct-v0.2
|
| 9 |
+
model-index:
|
| 10 |
+
- name: mistral-7b-instruct-v0.2-dpo-multilingual-mix-1st-apr-final
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# mistral-7b-instruct-v0.2-dpo-multilingual-mix-1st-apr-final
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on the None dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-06
|
| 39 |
+
- train_batch_size: 16
|
| 40 |
+
- eval_batch_size: 16
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 64
|
| 45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
+
- lr_scheduler_type: cosine
|
| 47 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 48 |
+
- num_epochs: 1
|
| 49 |
+
|
| 50 |
+
### Training results
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
### Framework versions
|
| 55 |
+
|
| 56 |
+
- PEFT 0.7.1
|
| 57 |
+
- Transformers 4.39.0.dev0
|
| 58 |
+
- Pytorch 2.2.2
|
| 59 |
+
- Datasets 2.14.6
|
| 60 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 671150064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fbc9571cdf65fa3142af5d4e597a1c7e4bb94154f1d190b60173f783c4333ae1
|
| 3 |
size 671150064
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.150745850255092,
|
| 4 |
+
"train_runtime": 48820.8334,
|
| 5 |
+
"train_samples": 38447,
|
| 6 |
+
"train_samples_per_second": 0.788,
|
| 7 |
+
"train_steps_per_second": 0.012
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.150745850255092,
|
| 4 |
+
"train_runtime": 48820.8334,
|
| 5 |
+
"train_samples": 38447,
|
| 6 |
+
"train_samples_per_second": 0.788,
|
| 7 |
+
"train_steps_per_second": 0.012
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,945 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9987515605493134,
|
| 5 |
+
"eval_steps": 2000,
|
| 6 |
+
"global_step": 600,
|
| 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 |
+
"grad_norm": 4.760079834365155,
|
| 14 |
+
"learning_rate": 8.333333333333334e-08,
|
| 15 |
+
"logits/chosen": -1.1630980968475342,
|
| 16 |
+
"logits/rejected": -1.0201224088668823,
|
| 17 |
+
"logps/chosen": -350.4145812988281,
|
| 18 |
+
"logps/rejected": -220.30422973632812,
|
| 19 |
+
"loss": 0.6931,
|
| 20 |
+
"rewards/accuracies": 0.0,
|
| 21 |
+
"rewards/chosen": 0.0,
|
| 22 |
+
"rewards/margins": 0.0,
|
| 23 |
+
"rewards/rejected": 0.0,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.02,
|
| 28 |
+
"grad_norm": 3.6284772366587372,
|
| 29 |
+
"learning_rate": 8.333333333333333e-07,
|
| 30 |
+
"logits/chosen": -1.0121445655822754,
|
| 31 |
+
"logits/rejected": -0.9547010064125061,
|
| 32 |
+
"logps/chosen": -320.40802001953125,
|
| 33 |
+
"logps/rejected": -186.48419189453125,
|
| 34 |
+
"loss": 0.6924,
|
| 35 |
+
"rewards/accuracies": 0.4982638955116272,
|
| 36 |
+
"rewards/chosen": 0.0024558762088418007,
|
| 37 |
+
"rewards/margins": 0.0015082670142874122,
|
| 38 |
+
"rewards/rejected": 0.0009476091363467276,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.03,
|
| 43 |
+
"grad_norm": 4.24809146375121,
|
| 44 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 45 |
+
"logits/chosen": -0.9381664395332336,
|
| 46 |
+
"logits/rejected": -0.8647511601448059,
|
| 47 |
+
"logps/chosen": -328.6592102050781,
|
| 48 |
+
"logps/rejected": -212.4827880859375,
|
| 49 |
+
"loss": 0.6822,
|
| 50 |
+
"rewards/accuracies": 0.7953125238418579,
|
| 51 |
+
"rewards/chosen": 0.034552641212940216,
|
| 52 |
+
"rewards/margins": 0.022216904908418655,
|
| 53 |
+
"rewards/rejected": 0.012335737235844135,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.05,
|
| 58 |
+
"grad_norm": 3.9797438421104276,
|
| 59 |
+
"learning_rate": 2.5e-06,
|
| 60 |
+
"logits/chosen": -1.020849585533142,
|
| 61 |
+
"logits/rejected": -0.9427526593208313,
|
| 62 |
+
"logps/chosen": -329.3730163574219,
|
| 63 |
+
"logps/rejected": -215.0742645263672,
|
| 64 |
+
"loss": 0.6458,
|
| 65 |
+
"rewards/accuracies": 0.854687511920929,
|
| 66 |
+
"rewards/chosen": 0.1433950513601303,
|
| 67 |
+
"rewards/margins": 0.10005545616149902,
|
| 68 |
+
"rewards/rejected": 0.04333961382508278,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.07,
|
| 73 |
+
"grad_norm": 2.971987358375742,
|
| 74 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 75 |
+
"logits/chosen": -1.1010403633117676,
|
| 76 |
+
"logits/rejected": -1.0402967929840088,
|
| 77 |
+
"logps/chosen": -310.91778564453125,
|
| 78 |
+
"logps/rejected": -205.1094512939453,
|
| 79 |
+
"loss": 0.5801,
|
| 80 |
+
"rewards/accuracies": 0.885937511920929,
|
| 81 |
+
"rewards/chosen": 0.3175004720687866,
|
| 82 |
+
"rewards/margins": 0.25547632575035095,
|
| 83 |
+
"rewards/rejected": 0.06202414631843567,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.08,
|
| 88 |
+
"grad_norm": 2.284422715303357,
|
| 89 |
+
"learning_rate": 4.166666666666667e-06,
|
| 90 |
+
"logits/chosen": -1.176831841468811,
|
| 91 |
+
"logits/rejected": -1.1168550252914429,
|
| 92 |
+
"logps/chosen": -289.9400634765625,
|
| 93 |
+
"logps/rejected": -193.466552734375,
|
| 94 |
+
"loss": 0.5192,
|
| 95 |
+
"rewards/accuracies": 0.8765624761581421,
|
| 96 |
+
"rewards/chosen": 0.4821470379829407,
|
| 97 |
+
"rewards/margins": 0.42469802498817444,
|
| 98 |
+
"rewards/rejected": 0.057449012994766235,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.1,
|
| 103 |
+
"grad_norm": 2.693731441358458,
|
| 104 |
+
"learning_rate": 5e-06,
|
| 105 |
+
"logits/chosen": -1.1333694458007812,
|
| 106 |
+
"logits/rejected": -1.0512011051177979,
|
| 107 |
+
"logps/chosen": -268.7071838378906,
|
| 108 |
+
"logps/rejected": -216.21939086914062,
|
| 109 |
+
"loss": 0.4625,
|
| 110 |
+
"rewards/accuracies": 0.890625,
|
| 111 |
+
"rewards/chosen": 0.570508599281311,
|
| 112 |
+
"rewards/margins": 0.6040245890617371,
|
| 113 |
+
"rewards/rejected": -0.03351598605513573,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.12,
|
| 118 |
+
"grad_norm": 2.982413579425004,
|
| 119 |
+
"learning_rate": 4.995770395678171e-06,
|
| 120 |
+
"logits/chosen": -0.9235696792602539,
|
| 121 |
+
"logits/rejected": -0.8171814680099487,
|
| 122 |
+
"logps/chosen": -288.454833984375,
|
| 123 |
+
"logps/rejected": -243.9117431640625,
|
| 124 |
+
"loss": 0.3919,
|
| 125 |
+
"rewards/accuracies": 0.9359375238418579,
|
| 126 |
+
"rewards/chosen": 0.4965239465236664,
|
| 127 |
+
"rewards/margins": 0.8616136312484741,
|
| 128 |
+
"rewards/rejected": -0.36508968472480774,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.13,
|
| 133 |
+
"grad_norm": 3.0591291304636767,
|
| 134 |
+
"learning_rate": 4.983095894354858e-06,
|
| 135 |
+
"logits/chosen": -0.5841406583786011,
|
| 136 |
+
"logits/rejected": -0.5758659243583679,
|
| 137 |
+
"logps/chosen": -282.39080810546875,
|
| 138 |
+
"logps/rejected": -267.4037170410156,
|
| 139 |
+
"loss": 0.3708,
|
| 140 |
+
"rewards/accuracies": 0.9234374761581421,
|
| 141 |
+
"rewards/chosen": 0.3620058298110962,
|
| 142 |
+
"rewards/margins": 1.001431941986084,
|
| 143 |
+
"rewards/rejected": -0.639426052570343,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.15,
|
| 148 |
+
"grad_norm": 2.7852348655427495,
|
| 149 |
+
"learning_rate": 4.962019382530521e-06,
|
| 150 |
+
"logits/chosen": -0.3057808578014374,
|
| 151 |
+
"logits/rejected": -0.08402713388204575,
|
| 152 |
+
"logps/chosen": -304.71258544921875,
|
| 153 |
+
"logps/rejected": -306.744873046875,
|
| 154 |
+
"loss": 0.3332,
|
| 155 |
+
"rewards/accuracies": 0.934374988079071,
|
| 156 |
+
"rewards/chosen": 0.34849730134010315,
|
| 157 |
+
"rewards/margins": 1.234140157699585,
|
| 158 |
+
"rewards/rejected": -0.8856428861618042,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.17,
|
| 163 |
+
"grad_norm": 3.904931302273915,
|
| 164 |
+
"learning_rate": 4.93261217644956e-06,
|
| 165 |
+
"logits/chosen": -0.3401206433773041,
|
| 166 |
+
"logits/rejected": -0.006557087413966656,
|
| 167 |
+
"logps/chosen": -306.9206848144531,
|
| 168 |
+
"logps/rejected": -352.503173828125,
|
| 169 |
+
"loss": 0.284,
|
| 170 |
+
"rewards/accuracies": 0.9375,
|
| 171 |
+
"rewards/chosen": 0.13179777562618256,
|
| 172 |
+
"rewards/margins": 1.5720919370651245,
|
| 173 |
+
"rewards/rejected": -1.4402940273284912,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.18,
|
| 178 |
+
"grad_norm": 4.332999329515302,
|
| 179 |
+
"learning_rate": 4.894973780788722e-06,
|
| 180 |
+
"logits/chosen": -0.3165335953235626,
|
| 181 |
+
"logits/rejected": 0.0790834054350853,
|
| 182 |
+
"logps/chosen": -350.21405029296875,
|
| 183 |
+
"logps/rejected": -433.90643310546875,
|
| 184 |
+
"loss": 0.2218,
|
| 185 |
+
"rewards/accuracies": 0.9609375,
|
| 186 |
+
"rewards/chosen": -0.08845386654138565,
|
| 187 |
+
"rewards/margins": 2.1888351440429688,
|
| 188 |
+
"rewards/rejected": -2.2772889137268066,
|
| 189 |
+
"step": 110
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.2,
|
| 193 |
+
"grad_norm": 3.8287946526526464,
|
| 194 |
+
"learning_rate": 4.849231551964771e-06,
|
| 195 |
+
"logits/chosen": -0.34387272596359253,
|
| 196 |
+
"logits/rejected": -0.02132757380604744,
|
| 197 |
+
"logps/chosen": -342.6011047363281,
|
| 198 |
+
"logps/rejected": -551.7183837890625,
|
| 199 |
+
"loss": 0.1787,
|
| 200 |
+
"rewards/accuracies": 0.9468749761581421,
|
| 201 |
+
"rewards/chosen": -0.15126076340675354,
|
| 202 |
+
"rewards/margins": 3.2273590564727783,
|
| 203 |
+
"rewards/rejected": -3.378619432449341,
|
| 204 |
+
"step": 120
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.22,
|
| 208 |
+
"grad_norm": 3.7220399456471203,
|
| 209 |
+
"learning_rate": 4.7955402672006855e-06,
|
| 210 |
+
"logits/chosen": -0.4739972949028015,
|
| 211 |
+
"logits/rejected": -0.22087886929512024,
|
| 212 |
+
"logps/chosen": -355.5228271484375,
|
| 213 |
+
"logps/rejected": -637.9622192382812,
|
| 214 |
+
"loss": 0.1651,
|
| 215 |
+
"rewards/accuracies": 0.949999988079071,
|
| 216 |
+
"rewards/chosen": -0.20811741054058075,
|
| 217 |
+
"rewards/margins": 4.130118370056152,
|
| 218 |
+
"rewards/rejected": -4.338235378265381,
|
| 219 |
+
"step": 130
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.23,
|
| 223 |
+
"grad_norm": 2.216070201202239,
|
| 224 |
+
"learning_rate": 4.734081600808531e-06,
|
| 225 |
+
"logits/chosen": -0.5762065649032593,
|
| 226 |
+
"logits/rejected": -0.37428033351898193,
|
| 227 |
+
"logps/chosen": -342.99853515625,
|
| 228 |
+
"logps/rejected": -746.65234375,
|
| 229 |
+
"loss": 0.1198,
|
| 230 |
+
"rewards/accuracies": 0.973437488079071,
|
| 231 |
+
"rewards/chosen": -0.11533623933792114,
|
| 232 |
+
"rewards/margins": 5.1436309814453125,
|
| 233 |
+
"rewards/rejected": -5.258967399597168,
|
| 234 |
+
"step": 140
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.25,
|
| 238 |
+
"grad_norm": 4.452614644496966,
|
| 239 |
+
"learning_rate": 4.665063509461098e-06,
|
| 240 |
+
"logits/chosen": -0.43740949034690857,
|
| 241 |
+
"logits/rejected": -0.276010662317276,
|
| 242 |
+
"logps/chosen": -353.4092712402344,
|
| 243 |
+
"logps/rejected": -838.3703002929688,
|
| 244 |
+
"loss": 0.145,
|
| 245 |
+
"rewards/accuracies": 0.9671875238418579,
|
| 246 |
+
"rewards/chosen": -0.23360753059387207,
|
| 247 |
+
"rewards/margins": 5.936570167541504,
|
| 248 |
+
"rewards/rejected": -6.170177459716797,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.27,
|
| 253 |
+
"grad_norm": 6.909783574695525,
|
| 254 |
+
"learning_rate": 4.588719528532342e-06,
|
| 255 |
+
"logits/chosen": 0.02835695818066597,
|
| 256 |
+
"logits/rejected": 0.3119501769542694,
|
| 257 |
+
"logps/chosen": -353.52203369140625,
|
| 258 |
+
"logps/rejected": -654.4629516601562,
|
| 259 |
+
"loss": 0.1368,
|
| 260 |
+
"rewards/accuracies": 0.973437488079071,
|
| 261 |
+
"rewards/chosen": -0.25593429803848267,
|
| 262 |
+
"rewards/margins": 4.225825786590576,
|
| 263 |
+
"rewards/rejected": -4.481760501861572,
|
| 264 |
+
"step": 160
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.28,
|
| 268 |
+
"grad_norm": 1.7107339593657505,
|
| 269 |
+
"learning_rate": 4.50530798188761e-06,
|
| 270 |
+
"logits/chosen": -0.15743690729141235,
|
| 271 |
+
"logits/rejected": 0.23819151520729065,
|
| 272 |
+
"logps/chosen": -333.4959716796875,
|
| 273 |
+
"logps/rejected": -659.2352294921875,
|
| 274 |
+
"loss": 0.1267,
|
| 275 |
+
"rewards/accuracies": 0.965624988079071,
|
| 276 |
+
"rewards/chosen": -0.00615291204303503,
|
| 277 |
+
"rewards/margins": 4.596193790435791,
|
| 278 |
+
"rewards/rejected": -4.602346420288086,
|
| 279 |
+
"step": 170
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.3,
|
| 283 |
+
"grad_norm": 3.623547042771,
|
| 284 |
+
"learning_rate": 4.415111107797445e-06,
|
| 285 |
+
"logits/chosen": -0.31918713450431824,
|
| 286 |
+
"logits/rejected": 0.03602874279022217,
|
| 287 |
+
"logps/chosen": -330.2435302734375,
|
| 288 |
+
"logps/rejected": -757.0585327148438,
|
| 289 |
+
"loss": 0.1056,
|
| 290 |
+
"rewards/accuracies": 0.979687511920929,
|
| 291 |
+
"rewards/chosen": 0.007477378938347101,
|
| 292 |
+
"rewards/margins": 5.585347652435303,
|
| 293 |
+
"rewards/rejected": -5.577870845794678,
|
| 294 |
+
"step": 180
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.32,
|
| 298 |
+
"grad_norm": 2.1551725469519507,
|
| 299 |
+
"learning_rate": 4.318434103932622e-06,
|
| 300 |
+
"logits/chosen": -0.3329532742500305,
|
| 301 |
+
"logits/rejected": -0.041298139840364456,
|
| 302 |
+
"logps/chosen": -382.2855529785156,
|
| 303 |
+
"logps/rejected": -988.2845458984375,
|
| 304 |
+
"loss": 0.1055,
|
| 305 |
+
"rewards/accuracies": 0.9703124761581421,
|
| 306 |
+
"rewards/chosen": -0.40997394919395447,
|
| 307 |
+
"rewards/margins": 7.356125831604004,
|
| 308 |
+
"rewards/rejected": -7.766099452972412,
|
| 309 |
+
"step": 190
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.33,
|
| 313 |
+
"grad_norm": 2.5178815949434505,
|
| 314 |
+
"learning_rate": 4.215604094671835e-06,
|
| 315 |
+
"logits/chosen": -0.4541547894477844,
|
| 316 |
+
"logits/rejected": -0.1613123118877411,
|
| 317 |
+
"logps/chosen": -347.7196350097656,
|
| 318 |
+
"logps/rejected": -880.2561645507812,
|
| 319 |
+
"loss": 0.1013,
|
| 320 |
+
"rewards/accuracies": 0.9781249761581421,
|
| 321 |
+
"rewards/chosen": -0.13055315613746643,
|
| 322 |
+
"rewards/margins": 6.6189093589782715,
|
| 323 |
+
"rewards/rejected": -6.749462127685547,
|
| 324 |
+
"step": 200
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.35,
|
| 328 |
+
"grad_norm": 2.0381257493358285,
|
| 329 |
+
"learning_rate": 4.106969024216348e-06,
|
| 330 |
+
"logits/chosen": -0.5205026865005493,
|
| 331 |
+
"logits/rejected": -0.25759488344192505,
|
| 332 |
+
"logps/chosen": -393.9586486816406,
|
| 333 |
+
"logps/rejected": -1019.0203247070312,
|
| 334 |
+
"loss": 0.0999,
|
| 335 |
+
"rewards/accuracies": 0.9781249761581421,
|
| 336 |
+
"rewards/chosen": -0.4994255006313324,
|
| 337 |
+
"rewards/margins": 7.512589931488037,
|
| 338 |
+
"rewards/rejected": -8.012015342712402,
|
| 339 |
+
"step": 210
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.37,
|
| 343 |
+
"grad_norm": 6.3674783354384115,
|
| 344 |
+
"learning_rate": 3.992896479256966e-06,
|
| 345 |
+
"logits/chosen": -0.6750475168228149,
|
| 346 |
+
"logits/rejected": -0.3842785954475403,
|
| 347 |
+
"logps/chosen": -336.7120666503906,
|
| 348 |
+
"logps/rejected": -858.0791015625,
|
| 349 |
+
"loss": 0.1031,
|
| 350 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 351 |
+
"rewards/chosen": 0.017676908522844315,
|
| 352 |
+
"rewards/margins": 6.612088680267334,
|
| 353 |
+
"rewards/rejected": -6.594411373138428,
|
| 354 |
+
"step": 220
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.38,
|
| 358 |
+
"grad_norm": 2.419780694811269,
|
| 359 |
+
"learning_rate": 3.8737724451770155e-06,
|
| 360 |
+
"logits/chosen": -0.6326015591621399,
|
| 361 |
+
"logits/rejected": -0.40925782918930054,
|
| 362 |
+
"logps/chosen": -375.2301025390625,
|
| 363 |
+
"logps/rejected": -1089.742919921875,
|
| 364 |
+
"loss": 0.0811,
|
| 365 |
+
"rewards/accuracies": 0.984375,
|
| 366 |
+
"rewards/chosen": -0.42722994089126587,
|
| 367 |
+
"rewards/margins": 8.390230178833008,
|
| 368 |
+
"rewards/rejected": -8.817461013793945,
|
| 369 |
+
"step": 230
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.4,
|
| 373 |
+
"grad_norm": 17.310687719433773,
|
| 374 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 375 |
+
"logits/chosen": -0.5065186023712158,
|
| 376 |
+
"logits/rejected": -0.2878126800060272,
|
| 377 |
+
"logps/chosen": -386.92266845703125,
|
| 378 |
+
"logps/rejected": -1173.471435546875,
|
| 379 |
+
"loss": 0.0646,
|
| 380 |
+
"rewards/accuracies": 0.9921875,
|
| 381 |
+
"rewards/chosen": -0.5754821300506592,
|
| 382 |
+
"rewards/margins": 9.043745040893555,
|
| 383 |
+
"rewards/rejected": -9.619227409362793,
|
| 384 |
+
"step": 240
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.42,
|
| 388 |
+
"grad_norm": 1.566552488550366,
|
| 389 |
+
"learning_rate": 3.621997950501156e-06,
|
| 390 |
+
"logits/chosen": -0.23604285717010498,
|
| 391 |
+
"logits/rejected": 0.03810877352952957,
|
| 392 |
+
"logps/chosen": -380.695068359375,
|
| 393 |
+
"logps/rejected": -1131.797607421875,
|
| 394 |
+
"loss": 0.0778,
|
| 395 |
+
"rewards/accuracies": 0.9859374761581421,
|
| 396 |
+
"rewards/chosen": -0.5175285935401917,
|
| 397 |
+
"rewards/margins": 8.581127166748047,
|
| 398 |
+
"rewards/rejected": -9.098657608032227,
|
| 399 |
+
"step": 250
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.43,
|
| 403 |
+
"grad_norm": 2.2662266467083074,
|
| 404 |
+
"learning_rate": 3.4901994150978926e-06,
|
| 405 |
+
"logits/chosen": -0.14488890767097473,
|
| 406 |
+
"logits/rejected": 0.2611751854419708,
|
| 407 |
+
"logps/chosen": -305.7989196777344,
|
| 408 |
+
"logps/rejected": -813.2493896484375,
|
| 409 |
+
"loss": 0.1017,
|
| 410 |
+
"rewards/accuracies": 0.9765625,
|
| 411 |
+
"rewards/chosen": 0.1130492091178894,
|
| 412 |
+
"rewards/margins": 6.16524600982666,
|
| 413 |
+
"rewards/rejected": -6.052196502685547,
|
| 414 |
+
"step": 260
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.45,
|
| 418 |
+
"grad_norm": 2.023161637845827,
|
| 419 |
+
"learning_rate": 3.3550503583141726e-06,
|
| 420 |
+
"logits/chosen": -0.14377684891223907,
|
| 421 |
+
"logits/rejected": 0.22304537892341614,
|
| 422 |
+
"logps/chosen": -372.13165283203125,
|
| 423 |
+
"logps/rejected": -896.8603515625,
|
| 424 |
+
"loss": 0.0815,
|
| 425 |
+
"rewards/accuracies": 0.984375,
|
| 426 |
+
"rewards/chosen": -0.10399510711431503,
|
| 427 |
+
"rewards/margins": 6.930342197418213,
|
| 428 |
+
"rewards/rejected": -7.034337520599365,
|
| 429 |
+
"step": 270
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.47,
|
| 433 |
+
"grad_norm": 2.8073955958066197,
|
| 434 |
+
"learning_rate": 3.217008081777726e-06,
|
| 435 |
+
"logits/chosen": -0.10699782520532608,
|
| 436 |
+
"logits/rejected": 0.14858277142047882,
|
| 437 |
+
"logps/chosen": -438.87823486328125,
|
| 438 |
+
"logps/rejected": -1203.0933837890625,
|
| 439 |
+
"loss": 0.0873,
|
| 440 |
+
"rewards/accuracies": 0.9703124761581421,
|
| 441 |
+
"rewards/chosen": -1.0280810594558716,
|
| 442 |
+
"rewards/margins": 8.752517700195312,
|
| 443 |
+
"rewards/rejected": -9.780599594116211,
|
| 444 |
+
"step": 280
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.48,
|
| 448 |
+
"grad_norm": 4.347308476957998,
|
| 449 |
+
"learning_rate": 3.0765396768561005e-06,
|
| 450 |
+
"logits/chosen": -0.4454914927482605,
|
| 451 |
+
"logits/rejected": -0.15811693668365479,
|
| 452 |
+
"logps/chosen": -349.51531982421875,
|
| 453 |
+
"logps/rejected": -1020.2342529296875,
|
| 454 |
+
"loss": 0.0657,
|
| 455 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 456 |
+
"rewards/chosen": -0.13168838620185852,
|
| 457 |
+
"rewards/margins": 7.97658634185791,
|
| 458 |
+
"rewards/rejected": -8.108274459838867,
|
| 459 |
+
"step": 290
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.5,
|
| 463 |
+
"grad_norm": 4.923260234351252,
|
| 464 |
+
"learning_rate": 2.9341204441673267e-06,
|
| 465 |
+
"logits/chosen": -0.46575579047203064,
|
| 466 |
+
"logits/rejected": -0.18623068928718567,
|
| 467 |
+
"logps/chosen": -382.58807373046875,
|
| 468 |
+
"logps/rejected": -1112.0509033203125,
|
| 469 |
+
"loss": 0.0805,
|
| 470 |
+
"rewards/accuracies": 0.979687511920929,
|
| 471 |
+
"rewards/chosen": -0.5270034670829773,
|
| 472 |
+
"rewards/margins": 8.469846725463867,
|
| 473 |
+
"rewards/rejected": -8.996851921081543,
|
| 474 |
+
"step": 300
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.52,
|
| 478 |
+
"grad_norm": 4.035586599115476,
|
| 479 |
+
"learning_rate": 2.7902322853130758e-06,
|
| 480 |
+
"logits/chosen": -0.47277918457984924,
|
| 481 |
+
"logits/rejected": -0.22883549332618713,
|
| 482 |
+
"logps/chosen": -396.8818054199219,
|
| 483 |
+
"logps/rejected": -1298.666748046875,
|
| 484 |
+
"loss": 0.0646,
|
| 485 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 486 |
+
"rewards/chosen": -0.6961787343025208,
|
| 487 |
+
"rewards/margins": 10.189523696899414,
|
| 488 |
+
"rewards/rejected": -10.885702133178711,
|
| 489 |
+
"step": 310
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.53,
|
| 493 |
+
"grad_norm": 1.7894983493848236,
|
| 494 |
+
"learning_rate": 2.6453620722761897e-06,
|
| 495 |
+
"logits/chosen": -0.5364641547203064,
|
| 496 |
+
"logits/rejected": -0.28673312067985535,
|
| 497 |
+
"logps/chosen": -379.64337158203125,
|
| 498 |
+
"logps/rejected": -1212.439697265625,
|
| 499 |
+
"loss": 0.0705,
|
| 500 |
+
"rewards/accuracies": 0.984375,
|
| 501 |
+
"rewards/chosen": -0.35370904207229614,
|
| 502 |
+
"rewards/margins": 9.660139083862305,
|
| 503 |
+
"rewards/rejected": -10.013849258422852,
|
| 504 |
+
"step": 320
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.55,
|
| 508 |
+
"grad_norm": 3.3096114791299955,
|
| 509 |
+
"learning_rate": 2.5e-06,
|
| 510 |
+
"logits/chosen": -0.5816466212272644,
|
| 511 |
+
"logits/rejected": -0.35151442885398865,
|
| 512 |
+
"logps/chosen": -379.9997863769531,
|
| 513 |
+
"logps/rejected": -1187.98876953125,
|
| 514 |
+
"loss": 0.0533,
|
| 515 |
+
"rewards/accuracies": 0.987500011920929,
|
| 516 |
+
"rewards/chosen": -0.41831880807876587,
|
| 517 |
+
"rewards/margins": 9.45728874206543,
|
| 518 |
+
"rewards/rejected": -9.875606536865234,
|
| 519 |
+
"step": 330
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.57,
|
| 523 |
+
"grad_norm": 3.165697815332725,
|
| 524 |
+
"learning_rate": 2.3546379277238107e-06,
|
| 525 |
+
"logits/chosen": -0.47588858008384705,
|
| 526 |
+
"logits/rejected": -0.26532530784606934,
|
| 527 |
+
"logps/chosen": -454.8421325683594,
|
| 528 |
+
"logps/rejected": -1527.9405517578125,
|
| 529 |
+
"loss": 0.0613,
|
| 530 |
+
"rewards/accuracies": 0.987500011920929,
|
| 531 |
+
"rewards/chosen": -1.2038371562957764,
|
| 532 |
+
"rewards/margins": 11.936580657958984,
|
| 533 |
+
"rewards/rejected": -13.140419006347656,
|
| 534 |
+
"step": 340
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.58,
|
| 538 |
+
"grad_norm": 2.150182775021995,
|
| 539 |
+
"learning_rate": 2.2097677146869242e-06,
|
| 540 |
+
"logits/chosen": -0.5888150930404663,
|
| 541 |
+
"logits/rejected": -0.2789239287376404,
|
| 542 |
+
"logps/chosen": -349.5444030761719,
|
| 543 |
+
"logps/rejected": -1127.449951171875,
|
| 544 |
+
"loss": 0.0722,
|
| 545 |
+
"rewards/accuracies": 0.987500011920929,
|
| 546 |
+
"rewards/chosen": -0.23732244968414307,
|
| 547 |
+
"rewards/margins": 8.962444305419922,
|
| 548 |
+
"rewards/rejected": -9.199767112731934,
|
| 549 |
+
"step": 350
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.6,
|
| 553 |
+
"grad_norm": 2.370472414334315,
|
| 554 |
+
"learning_rate": 2.0658795558326745e-06,
|
| 555 |
+
"logits/chosen": -0.4766604006290436,
|
| 556 |
+
"logits/rejected": -0.22312171757221222,
|
| 557 |
+
"logps/chosen": -409.70135498046875,
|
| 558 |
+
"logps/rejected": -1264.747314453125,
|
| 559 |
+
"loss": 0.0719,
|
| 560 |
+
"rewards/accuracies": 0.979687511920929,
|
| 561 |
+
"rewards/chosen": -0.5473332405090332,
|
| 562 |
+
"rewards/margins": 9.985517501831055,
|
| 563 |
+
"rewards/rejected": -10.532853126525879,
|
| 564 |
+
"step": 360
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.62,
|
| 568 |
+
"grad_norm": 4.485106240623341,
|
| 569 |
+
"learning_rate": 1.9234603231439e-06,
|
| 570 |
+
"logits/chosen": -0.40441417694091797,
|
| 571 |
+
"logits/rejected": -0.13194730877876282,
|
| 572 |
+
"logps/chosen": -391.0859069824219,
|
| 573 |
+
"logps/rejected": -1242.066162109375,
|
| 574 |
+
"loss": 0.0632,
|
| 575 |
+
"rewards/accuracies": 0.9859374761581421,
|
| 576 |
+
"rewards/chosen": -0.6057429909706116,
|
| 577 |
+
"rewards/margins": 9.785425186157227,
|
| 578 |
+
"rewards/rejected": -10.391169548034668,
|
| 579 |
+
"step": 370
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.63,
|
| 583 |
+
"grad_norm": 1.3547912002584528,
|
| 584 |
+
"learning_rate": 1.7829919182222752e-06,
|
| 585 |
+
"logits/chosen": -0.3805684447288513,
|
| 586 |
+
"logits/rejected": -0.11748667806386948,
|
| 587 |
+
"logps/chosen": -420.901611328125,
|
| 588 |
+
"logps/rejected": -1328.623291015625,
|
| 589 |
+
"loss": 0.0609,
|
| 590 |
+
"rewards/accuracies": 0.9859374761581421,
|
| 591 |
+
"rewards/chosen": -0.752387523651123,
|
| 592 |
+
"rewards/margins": 10.4163236618042,
|
| 593 |
+
"rewards/rejected": -11.16871166229248,
|
| 594 |
+
"step": 380
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 0.65,
|
| 598 |
+
"grad_norm": 7.41839867309329,
|
| 599 |
+
"learning_rate": 1.6449496416858285e-06,
|
| 600 |
+
"logits/chosen": -0.439796507358551,
|
| 601 |
+
"logits/rejected": -0.18884414434432983,
|
| 602 |
+
"logps/chosen": -361.68634033203125,
|
| 603 |
+
"logps/rejected": -1187.6492919921875,
|
| 604 |
+
"loss": 0.093,
|
| 605 |
+
"rewards/accuracies": 0.9859374761581421,
|
| 606 |
+
"rewards/chosen": -0.2755175232887268,
|
| 607 |
+
"rewards/margins": 9.546114921569824,
|
| 608 |
+
"rewards/rejected": -9.82163143157959,
|
| 609 |
+
"step": 390
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.67,
|
| 613 |
+
"grad_norm": 1.9341085433225174,
|
| 614 |
+
"learning_rate": 1.509800584902108e-06,
|
| 615 |
+
"logits/chosen": -0.4470156133174896,
|
| 616 |
+
"logits/rejected": -0.15826158225536346,
|
| 617 |
+
"logps/chosen": -325.4810791015625,
|
| 618 |
+
"logps/rejected": -963.2032470703125,
|
| 619 |
+
"loss": 0.0697,
|
| 620 |
+
"rewards/accuracies": 0.9921875,
|
| 621 |
+
"rewards/chosen": 0.05155152827501297,
|
| 622 |
+
"rewards/margins": 7.570789337158203,
|
| 623 |
+
"rewards/rejected": -7.519238471984863,
|
| 624 |
+
"step": 400
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 0.68,
|
| 628 |
+
"grad_norm": 2.2887818701903795,
|
| 629 |
+
"learning_rate": 1.3780020494988447e-06,
|
| 630 |
+
"logits/chosen": -0.38168513774871826,
|
| 631 |
+
"logits/rejected": -0.13108135759830475,
|
| 632 |
+
"logps/chosen": -338.5289001464844,
|
| 633 |
+
"logps/rejected": -994.54345703125,
|
| 634 |
+
"loss": 0.0767,
|
| 635 |
+
"rewards/accuracies": 0.987500011920929,
|
| 636 |
+
"rewards/chosen": -0.13217389583587646,
|
| 637 |
+
"rewards/margins": 7.651003360748291,
|
| 638 |
+
"rewards/rejected": -7.783177375793457,
|
| 639 |
+
"step": 410
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.7,
|
| 643 |
+
"grad_norm": 10.334424423093157,
|
| 644 |
+
"learning_rate": 1.2500000000000007e-06,
|
| 645 |
+
"logits/chosen": -0.32895341515541077,
|
| 646 |
+
"logits/rejected": -0.11540427058935165,
|
| 647 |
+
"logps/chosen": -408.0703125,
|
| 648 |
+
"logps/rejected": -1260.2664794921875,
|
| 649 |
+
"loss": 0.0555,
|
| 650 |
+
"rewards/accuracies": 0.984375,
|
| 651 |
+
"rewards/chosen": -0.7070173025131226,
|
| 652 |
+
"rewards/margins": 9.677043914794922,
|
| 653 |
+
"rewards/rejected": -10.384060859680176,
|
| 654 |
+
"step": 420
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.72,
|
| 658 |
+
"grad_norm": 3.446631906756751,
|
| 659 |
+
"learning_rate": 1.1262275548229852e-06,
|
| 660 |
+
"logits/chosen": -0.45178350806236267,
|
| 661 |
+
"logits/rejected": -0.1636919528245926,
|
| 662 |
+
"logps/chosen": -414.2293395996094,
|
| 663 |
+
"logps/rejected": -1283.5029296875,
|
| 664 |
+
"loss": 0.0678,
|
| 665 |
+
"rewards/accuracies": 0.984375,
|
| 666 |
+
"rewards/chosen": -0.8513700366020203,
|
| 667 |
+
"rewards/margins": 9.903493881225586,
|
| 668 |
+
"rewards/rejected": -10.754863739013672,
|
| 669 |
+
"step": 430
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.73,
|
| 673 |
+
"grad_norm": 3.1936523282167912,
|
| 674 |
+
"learning_rate": 1.0071035207430352e-06,
|
| 675 |
+
"logits/chosen": -0.44514569640159607,
|
| 676 |
+
"logits/rejected": -0.20628270506858826,
|
| 677 |
+
"logps/chosen": -388.30224609375,
|
| 678 |
+
"logps/rejected": -1252.75146484375,
|
| 679 |
+
"loss": 0.048,
|
| 680 |
+
"rewards/accuracies": 0.9921875,
|
| 681 |
+
"rewards/chosen": -0.6398779153823853,
|
| 682 |
+
"rewards/margins": 9.881658554077148,
|
| 683 |
+
"rewards/rejected": -10.521535873413086,
|
| 684 |
+
"step": 440
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 0.75,
|
| 688 |
+
"grad_norm": 2.7973345573120514,
|
| 689 |
+
"learning_rate": 8.930309757836517e-07,
|
| 690 |
+
"logits/chosen": -0.5846482515335083,
|
| 691 |
+
"logits/rejected": -0.3184075355529785,
|
| 692 |
+
"logps/chosen": -380.7689514160156,
|
| 693 |
+
"logps/rejected": -1142.24365234375,
|
| 694 |
+
"loss": 0.0603,
|
| 695 |
+
"rewards/accuracies": 0.995312511920929,
|
| 696 |
+
"rewards/chosen": -0.4020051956176758,
|
| 697 |
+
"rewards/margins": 8.967304229736328,
|
| 698 |
+
"rewards/rejected": -9.36931037902832,
|
| 699 |
+
"step": 450
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.77,
|
| 703 |
+
"grad_norm": 1.7683111661015913,
|
| 704 |
+
"learning_rate": 7.843959053281663e-07,
|
| 705 |
+
"logits/chosen": -0.4719081521034241,
|
| 706 |
+
"logits/rejected": -0.248517706990242,
|
| 707 |
+
"logps/chosen": -379.7674560546875,
|
| 708 |
+
"logps/rejected": -1191.5079345703125,
|
| 709 |
+
"loss": 0.0567,
|
| 710 |
+
"rewards/accuracies": 0.9859374761581421,
|
| 711 |
+
"rewards/chosen": -0.3425825238227844,
|
| 712 |
+
"rewards/margins": 9.473920822143555,
|
| 713 |
+
"rewards/rejected": -9.816503524780273,
|
| 714 |
+
"step": 460
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.78,
|
| 718 |
+
"grad_norm": 2.619770880026216,
|
| 719 |
+
"learning_rate": 6.815658960673782e-07,
|
| 720 |
+
"logits/chosen": -0.5005173683166504,
|
| 721 |
+
"logits/rejected": -0.22156552970409393,
|
| 722 |
+
"logps/chosen": -389.5422668457031,
|
| 723 |
+
"logps/rejected": -1232.5107421875,
|
| 724 |
+
"loss": 0.0507,
|
| 725 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 726 |
+
"rewards/chosen": -0.6200243830680847,
|
| 727 |
+
"rewards/margins": 9.514973640441895,
|
| 728 |
+
"rewards/rejected": -10.134997367858887,
|
| 729 |
+
"step": 470
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.8,
|
| 733 |
+
"grad_norm": 2.206190169965742,
|
| 734 |
+
"learning_rate": 5.848888922025553e-07,
|
| 735 |
+
"logits/chosen": -0.46861904859542847,
|
| 736 |
+
"logits/rejected": -0.18907694518566132,
|
| 737 |
+
"logps/chosen": -388.5583801269531,
|
| 738 |
+
"logps/rejected": -1196.972900390625,
|
| 739 |
+
"loss": 0.0648,
|
| 740 |
+
"rewards/accuracies": 0.9921875,
|
| 741 |
+
"rewards/chosen": -0.40207749605178833,
|
| 742 |
+
"rewards/margins": 9.50928783416748,
|
| 743 |
+
"rewards/rejected": -9.911364555358887,
|
| 744 |
+
"step": 480
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.82,
|
| 748 |
+
"grad_norm": 5.22635215684865,
|
| 749 |
+
"learning_rate": 4.946920181123904e-07,
|
| 750 |
+
"logits/chosen": -0.49871087074279785,
|
| 751 |
+
"logits/rejected": -0.23706772923469543,
|
| 752 |
+
"logps/chosen": -378.83209228515625,
|
| 753 |
+
"logps/rejected": -1230.994140625,
|
| 754 |
+
"loss": 0.0594,
|
| 755 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 756 |
+
"rewards/chosen": -0.36057573556900024,
|
| 757 |
+
"rewards/margins": 9.88486099243164,
|
| 758 |
+
"rewards/rejected": -10.245436668395996,
|
| 759 |
+
"step": 490
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.83,
|
| 763 |
+
"grad_norm": 4.399177165103731,
|
| 764 |
+
"learning_rate": 4.1128047146765936e-07,
|
| 765 |
+
"logits/chosen": -0.4957138001918793,
|
| 766 |
+
"logits/rejected": -0.22013764083385468,
|
| 767 |
+
"logps/chosen": -358.6220703125,
|
| 768 |
+
"logps/rejected": -1129.539794921875,
|
| 769 |
+
"loss": 0.0699,
|
| 770 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 771 |
+
"rewards/chosen": -0.2314276248216629,
|
| 772 |
+
"rewards/margins": 9.046243667602539,
|
| 773 |
+
"rewards/rejected": -9.277670860290527,
|
| 774 |
+
"step": 500
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.85,
|
| 778 |
+
"grad_norm": 2.4578322691260586,
|
| 779 |
+
"learning_rate": 3.3493649053890325e-07,
|
| 780 |
+
"logits/chosen": -0.3916376233100891,
|
| 781 |
+
"logits/rejected": -0.13147909939289093,
|
| 782 |
+
"logps/chosen": -376.29583740234375,
|
| 783 |
+
"logps/rejected": -1182.824462890625,
|
| 784 |
+
"loss": 0.0487,
|
| 785 |
+
"rewards/accuracies": 0.995312511920929,
|
| 786 |
+
"rewards/chosen": -0.43332648277282715,
|
| 787 |
+
"rewards/margins": 9.195769309997559,
|
| 788 |
+
"rewards/rejected": -9.629096031188965,
|
| 789 |
+
"step": 510
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.87,
|
| 793 |
+
"grad_norm": 2.130490613601597,
|
| 794 |
+
"learning_rate": 2.6591839919146963e-07,
|
| 795 |
+
"logits/chosen": -0.36289340257644653,
|
| 796 |
+
"logits/rejected": -0.13988874852657318,
|
| 797 |
+
"logps/chosen": -386.994873046875,
|
| 798 |
+
"logps/rejected": -1234.4404296875,
|
| 799 |
+
"loss": 0.0544,
|
| 800 |
+
"rewards/accuracies": 0.9859374761581421,
|
| 801 |
+
"rewards/chosen": -0.5290244221687317,
|
| 802 |
+
"rewards/margins": 9.592443466186523,
|
| 803 |
+
"rewards/rejected": -10.121468544006348,
|
| 804 |
+
"step": 520
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 0.88,
|
| 808 |
+
"grad_norm": 2.5265891239165703,
|
| 809 |
+
"learning_rate": 2.044597327993153e-07,
|
| 810 |
+
"logits/chosen": -0.44066888093948364,
|
| 811 |
+
"logits/rejected": -0.1672184318304062,
|
| 812 |
+
"logps/chosen": -404.02349853515625,
|
| 813 |
+
"logps/rejected": -1213.29638671875,
|
| 814 |
+
"loss": 0.0571,
|
| 815 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 816 |
+
"rewards/chosen": -0.50049889087677,
|
| 817 |
+
"rewards/margins": 9.578344345092773,
|
| 818 |
+
"rewards/rejected": -10.07884407043457,
|
| 819 |
+
"step": 530
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.9,
|
| 823 |
+
"grad_norm": 1.3929764183361966,
|
| 824 |
+
"learning_rate": 1.507684480352292e-07,
|
| 825 |
+
"logits/chosen": -0.4457983374595642,
|
| 826 |
+
"logits/rejected": -0.20653533935546875,
|
| 827 |
+
"logps/chosen": -387.2595520019531,
|
| 828 |
+
"logps/rejected": -1257.850341796875,
|
| 829 |
+
"loss": 0.0582,
|
| 830 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 831 |
+
"rewards/chosen": -0.5553634762763977,
|
| 832 |
+
"rewards/margins": 9.88083267211914,
|
| 833 |
+
"rewards/rejected": -10.436195373535156,
|
| 834 |
+
"step": 540
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 0.92,
|
| 838 |
+
"grad_norm": 1.8197547536284149,
|
| 839 |
+
"learning_rate": 1.0502621921127776e-07,
|
| 840 |
+
"logits/chosen": -0.4422330856323242,
|
| 841 |
+
"logits/rejected": -0.1755351722240448,
|
| 842 |
+
"logps/chosen": -414.64715576171875,
|
| 843 |
+
"logps/rejected": -1265.412353515625,
|
| 844 |
+
"loss": 0.052,
|
| 845 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 846 |
+
"rewards/chosen": -0.6048385500907898,
|
| 847 |
+
"rewards/margins": 9.867313385009766,
|
| 848 |
+
"rewards/rejected": -10.472152709960938,
|
| 849 |
+
"step": 550
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.93,
|
| 853 |
+
"grad_norm": 3.101887331185102,
|
| 854 |
+
"learning_rate": 6.738782355044048e-08,
|
| 855 |
+
"logits/chosen": -0.4752727448940277,
|
| 856 |
+
"logits/rejected": -0.20606884360313416,
|
| 857 |
+
"logps/chosen": -390.14984130859375,
|
| 858 |
+
"logps/rejected": -1334.9105224609375,
|
| 859 |
+
"loss": 0.0647,
|
| 860 |
+
"rewards/accuracies": 0.979687511920929,
|
| 861 |
+
"rewards/chosen": -0.6308324933052063,
|
| 862 |
+
"rewards/margins": 10.629243850708008,
|
| 863 |
+
"rewards/rejected": -11.260076522827148,
|
| 864 |
+
"step": 560
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.95,
|
| 868 |
+
"grad_norm": 3.441898096294271,
|
| 869 |
+
"learning_rate": 3.798061746947995e-08,
|
| 870 |
+
"logits/chosen": -0.480734646320343,
|
| 871 |
+
"logits/rejected": -0.21511860191822052,
|
| 872 |
+
"logps/chosen": -387.4568786621094,
|
| 873 |
+
"logps/rejected": -1270.74853515625,
|
| 874 |
+
"loss": 0.0527,
|
| 875 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 876 |
+
"rewards/chosen": -0.7548898458480835,
|
| 877 |
+
"rewards/margins": 9.84511947631836,
|
| 878 |
+
"rewards/rejected": -10.600008964538574,
|
| 879 |
+
"step": 570
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.97,
|
| 883 |
+
"grad_norm": 2.2741744504193826,
|
| 884 |
+
"learning_rate": 1.6904105645142443e-08,
|
| 885 |
+
"logits/chosen": -0.44097834825515747,
|
| 886 |
+
"logits/rejected": -0.2139013707637787,
|
| 887 |
+
"logps/chosen": -391.39453125,
|
| 888 |
+
"logps/rejected": -1291.123779296875,
|
| 889 |
+
"loss": 0.0469,
|
| 890 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 891 |
+
"rewards/chosen": -0.5810464024543762,
|
| 892 |
+
"rewards/margins": 10.310724258422852,
|
| 893 |
+
"rewards/rejected": -10.891772270202637,
|
| 894 |
+
"step": 580
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.98,
|
| 898 |
+
"grad_norm": 2.909048473319547,
|
| 899 |
+
"learning_rate": 4.229604321829561e-09,
|
| 900 |
+
"logits/chosen": -0.48271116614341736,
|
| 901 |
+
"logits/rejected": -0.22559651732444763,
|
| 902 |
+
"logps/chosen": -406.5685119628906,
|
| 903 |
+
"logps/rejected": -1254.4326171875,
|
| 904 |
+
"loss": 0.0513,
|
| 905 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 906 |
+
"rewards/chosen": -0.6645079851150513,
|
| 907 |
+
"rewards/margins": 9.709487915039062,
|
| 908 |
+
"rewards/rejected": -10.373995780944824,
|
| 909 |
+
"step": 590
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 1.0,
|
| 913 |
+
"grad_norm": 2.7152944727872392,
|
| 914 |
+
"learning_rate": 0.0,
|
| 915 |
+
"logits/chosen": -0.49268728494644165,
|
| 916 |
+
"logits/rejected": -0.23753933608531952,
|
| 917 |
+
"logps/chosen": -411.12884521484375,
|
| 918 |
+
"logps/rejected": -1246.8544921875,
|
| 919 |
+
"loss": 0.0566,
|
| 920 |
+
"rewards/accuracies": 0.9921875,
|
| 921 |
+
"rewards/chosen": -0.7164795994758606,
|
| 922 |
+
"rewards/margins": 9.6405668258667,
|
| 923 |
+
"rewards/rejected": -10.357046127319336,
|
| 924 |
+
"step": 600
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 1.0,
|
| 928 |
+
"step": 600,
|
| 929 |
+
"total_flos": 0.0,
|
| 930 |
+
"train_loss": 0.150745850255092,
|
| 931 |
+
"train_runtime": 48820.8334,
|
| 932 |
+
"train_samples_per_second": 0.788,
|
| 933 |
+
"train_steps_per_second": 0.012
|
| 934 |
+
}
|
| 935 |
+
],
|
| 936 |
+
"logging_steps": 10,
|
| 937 |
+
"max_steps": 600,
|
| 938 |
+
"num_input_tokens_seen": 0,
|
| 939 |
+
"num_train_epochs": 1,
|
| 940 |
+
"save_steps": 100,
|
| 941 |
+
"total_flos": 0.0,
|
| 942 |
+
"train_batch_size": 16,
|
| 943 |
+
"trial_name": null,
|
| 944 |
+
"trial_params": null
|
| 945 |
+
}
|