Model save
Browse files- README.md +78 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +992 -0
README.md
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- dpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: zephyr-7b-align-scan-6e-07-0.53-polynomial-2.0
|
| 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 |
+
# zephyr-7b-align-scan-6e-07-0.53-polynomial-2.0
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.8639
|
| 22 |
+
- Rewards/chosen: 0.6923
|
| 23 |
+
- Rewards/rejected: -0.5434
|
| 24 |
+
- Rewards/accuracies: 0.3393
|
| 25 |
+
- Rewards/margins: 1.2357
|
| 26 |
+
- Logps/rejected: -82.1536
|
| 27 |
+
- Logps/chosen: -73.1851
|
| 28 |
+
- Logits/rejected: -2.6910
|
| 29 |
+
- Logits/chosen: -2.7068
|
| 30 |
+
|
| 31 |
+
## Model description
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Intended uses & limitations
|
| 36 |
+
|
| 37 |
+
More information needed
|
| 38 |
+
|
| 39 |
+
## Training and evaluation data
|
| 40 |
+
|
| 41 |
+
More information needed
|
| 42 |
+
|
| 43 |
+
## Training procedure
|
| 44 |
+
|
| 45 |
+
### Training hyperparameters
|
| 46 |
+
|
| 47 |
+
The following hyperparameters were used during training:
|
| 48 |
+
- learning_rate: 6e-07
|
| 49 |
+
- train_batch_size: 8
|
| 50 |
+
- eval_batch_size: 8
|
| 51 |
+
- seed: 42
|
| 52 |
+
- distributed_type: multi-GPU
|
| 53 |
+
- num_devices: 4
|
| 54 |
+
- gradient_accumulation_steps: 2
|
| 55 |
+
- total_train_batch_size: 64
|
| 56 |
+
- total_eval_batch_size: 32
|
| 57 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 58 |
+
- lr_scheduler_type: polynomial
|
| 59 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 60 |
+
- num_epochs: 2
|
| 61 |
+
|
| 62 |
+
### Training results
|
| 63 |
+
|
| 64 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 65 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 66 |
+
| 0.7136 | 0.3484 | 100 | 0.7109 | 1.4011 | 0.8625 | 0.3512 | 0.5386 | -79.5010 | -71.8476 | -2.5458 | -2.5618 |
|
| 67 |
+
| 0.7461 | 0.6969 | 200 | 0.7643 | 1.0640 | 0.3687 | 0.3274 | 0.6952 | -80.4327 | -72.4838 | -2.5601 | -2.5759 |
|
| 68 |
+
| 0.3949 | 1.0453 | 300 | 0.7875 | 0.2070 | -0.6350 | 0.3472 | 0.8420 | -82.3265 | -74.1006 | -2.6135 | -2.6292 |
|
| 69 |
+
| 0.3838 | 1.3937 | 400 | 0.8714 | 0.4396 | -0.7042 | 0.3294 | 1.1438 | -82.4571 | -73.6618 | -2.6266 | -2.6422 |
|
| 70 |
+
| 0.371 | 1.7422 | 500 | 0.8639 | 0.6923 | -0.5434 | 0.3393 | 1.2357 | -82.1536 | -73.1851 | -2.6910 | -2.7068 |
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
### Framework versions
|
| 74 |
+
|
| 75 |
+
- Transformers 4.44.2
|
| 76 |
+
- Pytorch 2.4.0
|
| 77 |
+
- Datasets 2.21.0
|
| 78 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.5567006942287139,
|
| 5 |
+
"train_runtime": 6469.9642,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.669,
|
| 8 |
+
"train_steps_per_second": 0.089
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.44.2"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.5567006942287139,
|
| 5 |
+
"train_runtime": 6469.9642,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.669,
|
| 8 |
+
"train_steps_per_second": 0.089
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,992 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 574,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.003484320557491289,
|
| 13 |
+
"grad_norm": 306.63621441123377,
|
| 14 |
+
"learning_rate": 1.0344827586206896e-08,
|
| 15 |
+
"logits/chosen": -2.5345611572265625,
|
| 16 |
+
"logits/rejected": -2.581700563430786,
|
| 17 |
+
"logps/chosen": -60.002105712890625,
|
| 18 |
+
"logps/rejected": -99.98374938964844,
|
| 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": 0.03484320557491289,
|
| 28 |
+
"grad_norm": 287.10820765444424,
|
| 29 |
+
"learning_rate": 1.0344827586206897e-07,
|
| 30 |
+
"logits/chosen": -2.5633163452148438,
|
| 31 |
+
"logits/rejected": -2.562026023864746,
|
| 32 |
+
"logps/chosen": -59.65489196777344,
|
| 33 |
+
"logps/rejected": -73.39691925048828,
|
| 34 |
+
"loss": 0.6954,
|
| 35 |
+
"rewards/accuracies": 0.2152777761220932,
|
| 36 |
+
"rewards/chosen": 0.0025260683614760637,
|
| 37 |
+
"rewards/margins": 0.011007179506123066,
|
| 38 |
+
"rewards/rejected": -0.008481111377477646,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.06968641114982578,
|
| 43 |
+
"grad_norm": 361.11744918130063,
|
| 44 |
+
"learning_rate": 2.0689655172413793e-07,
|
| 45 |
+
"logits/chosen": -2.60577654838562,
|
| 46 |
+
"logits/rejected": -2.5645222663879395,
|
| 47 |
+
"logps/chosen": -104.05818939208984,
|
| 48 |
+
"logps/rejected": -94.88358306884766,
|
| 49 |
+
"loss": 0.6868,
|
| 50 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 51 |
+
"rewards/chosen": 0.0365118607878685,
|
| 52 |
+
"rewards/margins": 0.03092603012919426,
|
| 53 |
+
"rewards/rejected": 0.005585831124335527,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.10452961672473868,
|
| 58 |
+
"grad_norm": 362.0188620910404,
|
| 59 |
+
"learning_rate": 3.103448275862069e-07,
|
| 60 |
+
"logits/chosen": -2.593327045440674,
|
| 61 |
+
"logits/rejected": -2.573579788208008,
|
| 62 |
+
"logps/chosen": -82.2002944946289,
|
| 63 |
+
"logps/rejected": -91.45396423339844,
|
| 64 |
+
"loss": 0.6711,
|
| 65 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 66 |
+
"rewards/chosen": 0.15702371299266815,
|
| 67 |
+
"rewards/margins": 0.1113327145576477,
|
| 68 |
+
"rewards/rejected": 0.04569100961089134,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13937282229965156,
|
| 73 |
+
"grad_norm": 283.91931715411215,
|
| 74 |
+
"learning_rate": 4.1379310344827586e-07,
|
| 75 |
+
"logits/chosen": -2.4966464042663574,
|
| 76 |
+
"logits/rejected": -2.4948105812072754,
|
| 77 |
+
"logps/chosen": -77.7404556274414,
|
| 78 |
+
"logps/rejected": -73.33540344238281,
|
| 79 |
+
"loss": 0.6457,
|
| 80 |
+
"rewards/accuracies": 0.3125,
|
| 81 |
+
"rewards/chosen": 0.09431798756122589,
|
| 82 |
+
"rewards/margins": 0.27138587832450867,
|
| 83 |
+
"rewards/rejected": -0.17706790566444397,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.17421602787456447,
|
| 88 |
+
"grad_norm": 251.26626326645746,
|
| 89 |
+
"learning_rate": 5.172413793103448e-07,
|
| 90 |
+
"logits/chosen": -2.5199971199035645,
|
| 91 |
+
"logits/rejected": -2.5240330696105957,
|
| 92 |
+
"logps/chosen": -62.982337951660156,
|
| 93 |
+
"logps/rejected": -75.54759216308594,
|
| 94 |
+
"loss": 0.6705,
|
| 95 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 96 |
+
"rewards/chosen": 0.5071097612380981,
|
| 97 |
+
"rewards/margins": 0.21250581741333008,
|
| 98 |
+
"rewards/rejected": 0.2946038842201233,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.20905923344947736,
|
| 103 |
+
"grad_norm": 245.4785949332324,
|
| 104 |
+
"learning_rate": 5.98062015503876e-07,
|
| 105 |
+
"logits/chosen": -2.473501682281494,
|
| 106 |
+
"logits/rejected": -2.4677951335906982,
|
| 107 |
+
"logps/chosen": -70.63660430908203,
|
| 108 |
+
"logps/rejected": -66.41564178466797,
|
| 109 |
+
"loss": 0.6582,
|
| 110 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 111 |
+
"rewards/chosen": 1.3620513677597046,
|
| 112 |
+
"rewards/margins": 0.30580419301986694,
|
| 113 |
+
"rewards/rejected": 1.0562469959259033,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.24390243902439024,
|
| 118 |
+
"grad_norm": 268.1673884309266,
|
| 119 |
+
"learning_rate": 5.883720930232558e-07,
|
| 120 |
+
"logits/chosen": -2.48606538772583,
|
| 121 |
+
"logits/rejected": -2.4808874130249023,
|
| 122 |
+
"logps/chosen": -60.53791046142578,
|
| 123 |
+
"logps/rejected": -65.51335906982422,
|
| 124 |
+
"loss": 0.669,
|
| 125 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 126 |
+
"rewards/chosen": 1.9600938558578491,
|
| 127 |
+
"rewards/margins": 0.41606950759887695,
|
| 128 |
+
"rewards/rejected": 1.5440242290496826,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.2787456445993031,
|
| 133 |
+
"grad_norm": 266.9849201621109,
|
| 134 |
+
"learning_rate": 5.786821705426356e-07,
|
| 135 |
+
"logits/chosen": -2.44217586517334,
|
| 136 |
+
"logits/rejected": -2.432021379470825,
|
| 137 |
+
"logps/chosen": -71.77671813964844,
|
| 138 |
+
"logps/rejected": -74.41423797607422,
|
| 139 |
+
"loss": 0.6931,
|
| 140 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 141 |
+
"rewards/chosen": 2.0607221126556396,
|
| 142 |
+
"rewards/margins": 0.417347252368927,
|
| 143 |
+
"rewards/rejected": 1.643375039100647,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.313588850174216,
|
| 148 |
+
"grad_norm": 299.0911881932943,
|
| 149 |
+
"learning_rate": 5.689922480620155e-07,
|
| 150 |
+
"logits/chosen": -2.486508846282959,
|
| 151 |
+
"logits/rejected": -2.5009188652038574,
|
| 152 |
+
"logps/chosen": -62.32392120361328,
|
| 153 |
+
"logps/rejected": -67.05072784423828,
|
| 154 |
+
"loss": 0.745,
|
| 155 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 156 |
+
"rewards/chosen": 1.938929557800293,
|
| 157 |
+
"rewards/margins": 0.3153690993785858,
|
| 158 |
+
"rewards/rejected": 1.6235605478286743,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.34843205574912894,
|
| 163 |
+
"grad_norm": 316.56572577244293,
|
| 164 |
+
"learning_rate": 5.593023255813953e-07,
|
| 165 |
+
"logits/chosen": -2.4774889945983887,
|
| 166 |
+
"logits/rejected": -2.4782986640930176,
|
| 167 |
+
"logps/chosen": -71.84193420410156,
|
| 168 |
+
"logps/rejected": -78.91864013671875,
|
| 169 |
+
"loss": 0.7136,
|
| 170 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 171 |
+
"rewards/chosen": 1.8174957036972046,
|
| 172 |
+
"rewards/margins": 0.686999499797821,
|
| 173 |
+
"rewards/rejected": 1.1304961442947388,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.34843205574912894,
|
| 178 |
+
"eval_logits/chosen": -2.561777114868164,
|
| 179 |
+
"eval_logits/rejected": -2.545793294906616,
|
| 180 |
+
"eval_logps/chosen": -71.84760284423828,
|
| 181 |
+
"eval_logps/rejected": -79.50099182128906,
|
| 182 |
+
"eval_loss": 0.7108728885650635,
|
| 183 |
+
"eval_rewards/accuracies": 0.3511904776096344,
|
| 184 |
+
"eval_rewards/chosen": 1.4011276960372925,
|
| 185 |
+
"eval_rewards/margins": 0.5386245846748352,
|
| 186 |
+
"eval_rewards/rejected": 0.8625030517578125,
|
| 187 |
+
"eval_runtime": 113.5305,
|
| 188 |
+
"eval_samples_per_second": 17.616,
|
| 189 |
+
"eval_steps_per_second": 0.555,
|
| 190 |
+
"step": 100
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.3832752613240418,
|
| 194 |
+
"grad_norm": 345.3192427581726,
|
| 195 |
+
"learning_rate": 5.496124031007752e-07,
|
| 196 |
+
"logits/chosen": -2.500793933868408,
|
| 197 |
+
"logits/rejected": -2.4660115242004395,
|
| 198 |
+
"logps/chosen": -72.03620910644531,
|
| 199 |
+
"logps/rejected": -62.499176025390625,
|
| 200 |
+
"loss": 0.7594,
|
| 201 |
+
"rewards/accuracies": 0.28125,
|
| 202 |
+
"rewards/chosen": 0.6009193658828735,
|
| 203 |
+
"rewards/margins": 0.3298734426498413,
|
| 204 |
+
"rewards/rejected": 0.2710459232330322,
|
| 205 |
+
"step": 110
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.4181184668989547,
|
| 209 |
+
"grad_norm": 243.7872401824864,
|
| 210 |
+
"learning_rate": 5.399224806201551e-07,
|
| 211 |
+
"logits/chosen": -2.5411460399627686,
|
| 212 |
+
"logits/rejected": -2.510791778564453,
|
| 213 |
+
"logps/chosen": -76.94505310058594,
|
| 214 |
+
"logps/rejected": -67.3199462890625,
|
| 215 |
+
"loss": 0.704,
|
| 216 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 217 |
+
"rewards/chosen": 0.27375704050064087,
|
| 218 |
+
"rewards/margins": 0.6173890829086304,
|
| 219 |
+
"rewards/rejected": -0.3436321020126343,
|
| 220 |
+
"step": 120
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4529616724738676,
|
| 224 |
+
"grad_norm": 432.4880968925023,
|
| 225 |
+
"learning_rate": 5.302325581395349e-07,
|
| 226 |
+
"logits/chosen": -2.5748209953308105,
|
| 227 |
+
"logits/rejected": -2.5566792488098145,
|
| 228 |
+
"logps/chosen": -83.82911682128906,
|
| 229 |
+
"logps/rejected": -89.06166076660156,
|
| 230 |
+
"loss": 0.7703,
|
| 231 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 232 |
+
"rewards/chosen": 0.027380788698792458,
|
| 233 |
+
"rewards/margins": 0.8505465388298035,
|
| 234 |
+
"rewards/rejected": -0.823165774345398,
|
| 235 |
+
"step": 130
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.4878048780487805,
|
| 239 |
+
"grad_norm": 213.43504772096108,
|
| 240 |
+
"learning_rate": 5.205426356589147e-07,
|
| 241 |
+
"logits/chosen": -2.4663920402526855,
|
| 242 |
+
"logits/rejected": -2.456141233444214,
|
| 243 |
+
"logps/chosen": -79.90180969238281,
|
| 244 |
+
"logps/rejected": -71.20616149902344,
|
| 245 |
+
"loss": 0.7043,
|
| 246 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 247 |
+
"rewards/chosen": 0.6591196060180664,
|
| 248 |
+
"rewards/margins": 0.893332302570343,
|
| 249 |
+
"rewards/rejected": -0.23421280086040497,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.5226480836236934,
|
| 254 |
+
"grad_norm": 314.59295726808995,
|
| 255 |
+
"learning_rate": 5.108527131782946e-07,
|
| 256 |
+
"logits/chosen": -2.5461785793304443,
|
| 257 |
+
"logits/rejected": -2.5047733783721924,
|
| 258 |
+
"logps/chosen": -77.75875091552734,
|
| 259 |
+
"logps/rejected": -79.21798706054688,
|
| 260 |
+
"loss": 0.7534,
|
| 261 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 262 |
+
"rewards/chosen": 0.9145609140396118,
|
| 263 |
+
"rewards/margins": 0.7381798624992371,
|
| 264 |
+
"rewards/rejected": 0.17638106644153595,
|
| 265 |
+
"step": 150
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.5574912891986062,
|
| 269 |
+
"grad_norm": 288.3107962146333,
|
| 270 |
+
"learning_rate": 5.011627906976744e-07,
|
| 271 |
+
"logits/chosen": -2.5273547172546387,
|
| 272 |
+
"logits/rejected": -2.5466020107269287,
|
| 273 |
+
"logps/chosen": -62.79814910888672,
|
| 274 |
+
"logps/rejected": -71.2327880859375,
|
| 275 |
+
"loss": 0.7312,
|
| 276 |
+
"rewards/accuracies": 0.26249998807907104,
|
| 277 |
+
"rewards/chosen": 0.9151515960693359,
|
| 278 |
+
"rewards/margins": 0.5228849649429321,
|
| 279 |
+
"rewards/rejected": 0.3922666311264038,
|
| 280 |
+
"step": 160
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.5923344947735192,
|
| 284 |
+
"grad_norm": 324.81976956861786,
|
| 285 |
+
"learning_rate": 4.914728682170542e-07,
|
| 286 |
+
"logits/chosen": -2.5521976947784424,
|
| 287 |
+
"logits/rejected": -2.5384509563446045,
|
| 288 |
+
"logps/chosen": -66.97964477539062,
|
| 289 |
+
"logps/rejected": -75.58006286621094,
|
| 290 |
+
"loss": 0.7191,
|
| 291 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 292 |
+
"rewards/chosen": 1.0535058975219727,
|
| 293 |
+
"rewards/margins": 0.664495587348938,
|
| 294 |
+
"rewards/rejected": 0.3890102505683899,
|
| 295 |
+
"step": 170
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.627177700348432,
|
| 299 |
+
"grad_norm": 292.82490566453185,
|
| 300 |
+
"learning_rate": 4.817829457364341e-07,
|
| 301 |
+
"logits/chosen": -2.5870025157928467,
|
| 302 |
+
"logits/rejected": -2.5761828422546387,
|
| 303 |
+
"logps/chosen": -88.94615936279297,
|
| 304 |
+
"logps/rejected": -84.6361083984375,
|
| 305 |
+
"loss": 0.7825,
|
| 306 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 307 |
+
"rewards/chosen": 1.5428403615951538,
|
| 308 |
+
"rewards/margins": 0.5494336485862732,
|
| 309 |
+
"rewards/rejected": 0.9934068918228149,
|
| 310 |
+
"step": 180
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.662020905923345,
|
| 314 |
+
"grad_norm": 234.8773688372816,
|
| 315 |
+
"learning_rate": 4.7209302325581395e-07,
|
| 316 |
+
"logits/chosen": -2.5796515941619873,
|
| 317 |
+
"logits/rejected": -2.57136607170105,
|
| 318 |
+
"logps/chosen": -68.72258758544922,
|
| 319 |
+
"logps/rejected": -79.8985595703125,
|
| 320 |
+
"loss": 0.7347,
|
| 321 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 322 |
+
"rewards/chosen": 1.2343099117279053,
|
| 323 |
+
"rewards/margins": 0.37060627341270447,
|
| 324 |
+
"rewards/rejected": 0.8637038469314575,
|
| 325 |
+
"step": 190
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.6968641114982579,
|
| 329 |
+
"grad_norm": 439.89262521049136,
|
| 330 |
+
"learning_rate": 4.6240310077519373e-07,
|
| 331 |
+
"logits/chosen": -2.605844259262085,
|
| 332 |
+
"logits/rejected": -2.612717390060425,
|
| 333 |
+
"logps/chosen": -87.73692321777344,
|
| 334 |
+
"logps/rejected": -90.65494537353516,
|
| 335 |
+
"loss": 0.7461,
|
| 336 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 337 |
+
"rewards/chosen": 1.5472986698150635,
|
| 338 |
+
"rewards/margins": 0.8075531125068665,
|
| 339 |
+
"rewards/rejected": 0.739745557308197,
|
| 340 |
+
"step": 200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.6968641114982579,
|
| 344 |
+
"eval_logits/chosen": -2.5759267807006836,
|
| 345 |
+
"eval_logits/rejected": -2.5600757598876953,
|
| 346 |
+
"eval_logps/chosen": -72.48377227783203,
|
| 347 |
+
"eval_logps/rejected": -80.43267822265625,
|
| 348 |
+
"eval_loss": 0.7642679214477539,
|
| 349 |
+
"eval_rewards/accuracies": 0.3273809552192688,
|
| 350 |
+
"eval_rewards/chosen": 1.063955545425415,
|
| 351 |
+
"eval_rewards/margins": 0.6952447295188904,
|
| 352 |
+
"eval_rewards/rejected": 0.36871081590652466,
|
| 353 |
+
"eval_runtime": 113.5756,
|
| 354 |
+
"eval_samples_per_second": 17.609,
|
| 355 |
+
"eval_steps_per_second": 0.555,
|
| 356 |
+
"step": 200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.7317073170731707,
|
| 360 |
+
"grad_norm": 437.59995338780175,
|
| 361 |
+
"learning_rate": 4.527131782945735e-07,
|
| 362 |
+
"logits/chosen": -2.5814366340637207,
|
| 363 |
+
"logits/rejected": -2.556798219680786,
|
| 364 |
+
"logps/chosen": -67.51020812988281,
|
| 365 |
+
"logps/rejected": -63.342933654785156,
|
| 366 |
+
"loss": 0.7354,
|
| 367 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 368 |
+
"rewards/chosen": 0.9586971402168274,
|
| 369 |
+
"rewards/margins": 0.9261114001274109,
|
| 370 |
+
"rewards/rejected": 0.03258571773767471,
|
| 371 |
+
"step": 210
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.7665505226480837,
|
| 375 |
+
"grad_norm": 270.9450930053244,
|
| 376 |
+
"learning_rate": 4.4302325581395346e-07,
|
| 377 |
+
"logits/chosen": -2.6260294914245605,
|
| 378 |
+
"logits/rejected": -2.6077115535736084,
|
| 379 |
+
"logps/chosen": -71.33647155761719,
|
| 380 |
+
"logps/rejected": -70.29251861572266,
|
| 381 |
+
"loss": 0.7602,
|
| 382 |
+
"rewards/accuracies": 0.23125000298023224,
|
| 383 |
+
"rewards/chosen": 1.1245003938674927,
|
| 384 |
+
"rewards/margins": 0.3980458974838257,
|
| 385 |
+
"rewards/rejected": 0.7264544367790222,
|
| 386 |
+
"step": 220
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.8013937282229965,
|
| 390 |
+
"grad_norm": 365.5474496656497,
|
| 391 |
+
"learning_rate": 4.3333333333333335e-07,
|
| 392 |
+
"logits/chosen": -2.6431806087493896,
|
| 393 |
+
"logits/rejected": -2.623713493347168,
|
| 394 |
+
"logps/chosen": -86.96244812011719,
|
| 395 |
+
"logps/rejected": -87.527587890625,
|
| 396 |
+
"loss": 0.7878,
|
| 397 |
+
"rewards/accuracies": 0.375,
|
| 398 |
+
"rewards/chosen": 1.6102378368377686,
|
| 399 |
+
"rewards/margins": 1.2809720039367676,
|
| 400 |
+
"rewards/rejected": 0.329265832901001,
|
| 401 |
+
"step": 230
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.8362369337979094,
|
| 405 |
+
"grad_norm": 311.80252726932434,
|
| 406 |
+
"learning_rate": 4.2364341085271313e-07,
|
| 407 |
+
"logits/chosen": -2.643277406692505,
|
| 408 |
+
"logits/rejected": -2.609691619873047,
|
| 409 |
+
"logps/chosen": -83.11528015136719,
|
| 410 |
+
"logps/rejected": -77.69151306152344,
|
| 411 |
+
"loss": 0.8033,
|
| 412 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 413 |
+
"rewards/chosen": 1.9219558238983154,
|
| 414 |
+
"rewards/margins": 0.6263306736946106,
|
| 415 |
+
"rewards/rejected": 1.2956254482269287,
|
| 416 |
+
"step": 240
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.8710801393728222,
|
| 420 |
+
"grad_norm": 352.8980579572314,
|
| 421 |
+
"learning_rate": 4.13953488372093e-07,
|
| 422 |
+
"logits/chosen": -2.65216064453125,
|
| 423 |
+
"logits/rejected": -2.617506504058838,
|
| 424 |
+
"logps/chosen": -91.78350067138672,
|
| 425 |
+
"logps/rejected": -87.70478820800781,
|
| 426 |
+
"loss": 0.6512,
|
| 427 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 428 |
+
"rewards/chosen": 1.942857027053833,
|
| 429 |
+
"rewards/margins": 0.7549124956130981,
|
| 430 |
+
"rewards/rejected": 1.1879446506500244,
|
| 431 |
+
"step": 250
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.9059233449477352,
|
| 435 |
+
"grad_norm": 265.7817814419997,
|
| 436 |
+
"learning_rate": 4.0426356589147286e-07,
|
| 437 |
+
"logits/chosen": -2.5617775917053223,
|
| 438 |
+
"logits/rejected": -2.5759947299957275,
|
| 439 |
+
"logps/chosen": -56.67157745361328,
|
| 440 |
+
"logps/rejected": -64.38258361816406,
|
| 441 |
+
"loss": 0.7944,
|
| 442 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 443 |
+
"rewards/chosen": 1.26022469997406,
|
| 444 |
+
"rewards/margins": 0.6000041961669922,
|
| 445 |
+
"rewards/rejected": 0.6602205038070679,
|
| 446 |
+
"step": 260
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.9407665505226481,
|
| 450 |
+
"grad_norm": 513.5414604532187,
|
| 451 |
+
"learning_rate": 3.9457364341085264e-07,
|
| 452 |
+
"logits/chosen": -2.6499624252319336,
|
| 453 |
+
"logits/rejected": -2.650109052658081,
|
| 454 |
+
"logps/chosen": -66.84712219238281,
|
| 455 |
+
"logps/rejected": -82.05715942382812,
|
| 456 |
+
"loss": 0.7273,
|
| 457 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 458 |
+
"rewards/chosen": 1.0775161981582642,
|
| 459 |
+
"rewards/margins": 0.8259714841842651,
|
| 460 |
+
"rewards/rejected": 0.2515445053577423,
|
| 461 |
+
"step": 270
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.975609756097561,
|
| 465 |
+
"grad_norm": 338.03485816068525,
|
| 466 |
+
"learning_rate": 3.848837209302326e-07,
|
| 467 |
+
"logits/chosen": -2.550506591796875,
|
| 468 |
+
"logits/rejected": -2.529942512512207,
|
| 469 |
+
"logps/chosen": -65.82142639160156,
|
| 470 |
+
"logps/rejected": -70.44654846191406,
|
| 471 |
+
"loss": 0.6677,
|
| 472 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 473 |
+
"rewards/chosen": 1.0668643712997437,
|
| 474 |
+
"rewards/margins": 0.8612996339797974,
|
| 475 |
+
"rewards/rejected": 0.20556476712226868,
|
| 476 |
+
"step": 280
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.0104529616724738,
|
| 480 |
+
"grad_norm": 43.71804076421765,
|
| 481 |
+
"learning_rate": 3.7519379844961237e-07,
|
| 482 |
+
"logits/chosen": -2.5674736499786377,
|
| 483 |
+
"logits/rejected": -2.54020357131958,
|
| 484 |
+
"logps/chosen": -68.04539489746094,
|
| 485 |
+
"logps/rejected": -65.61439514160156,
|
| 486 |
+
"loss": 0.5817,
|
| 487 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 488 |
+
"rewards/chosen": 2.1318130493164062,
|
| 489 |
+
"rewards/margins": 2.503080368041992,
|
| 490 |
+
"rewards/rejected": -0.37126731872558594,
|
| 491 |
+
"step": 290
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.0452961672473868,
|
| 495 |
+
"grad_norm": 14.731335253505609,
|
| 496 |
+
"learning_rate": 3.6550387596899226e-07,
|
| 497 |
+
"logits/chosen": -2.6065874099731445,
|
| 498 |
+
"logits/rejected": -2.5927734375,
|
| 499 |
+
"logps/chosen": -59.9798698425293,
|
| 500 |
+
"logps/rejected": -77.5359115600586,
|
| 501 |
+
"loss": 0.3949,
|
| 502 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 503 |
+
"rewards/chosen": 3.481590747833252,
|
| 504 |
+
"rewards/margins": 7.915454864501953,
|
| 505 |
+
"rewards/rejected": -4.433863639831543,
|
| 506 |
+
"step": 300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.0452961672473868,
|
| 510 |
+
"eval_logits/chosen": -2.6291775703430176,
|
| 511 |
+
"eval_logits/rejected": -2.6134917736053467,
|
| 512 |
+
"eval_logps/chosen": -74.10063934326172,
|
| 513 |
+
"eval_logps/rejected": -82.3265380859375,
|
| 514 |
+
"eval_loss": 0.7874619364738464,
|
| 515 |
+
"eval_rewards/accuracies": 0.3472222089767456,
|
| 516 |
+
"eval_rewards/chosen": 0.20701055228710175,
|
| 517 |
+
"eval_rewards/margins": 0.8420494794845581,
|
| 518 |
+
"eval_rewards/rejected": -0.6350388526916504,
|
| 519 |
+
"eval_runtime": 113.6067,
|
| 520 |
+
"eval_samples_per_second": 17.605,
|
| 521 |
+
"eval_steps_per_second": 0.555,
|
| 522 |
+
"step": 300
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.0801393728222997,
|
| 526 |
+
"grad_norm": 4.834789646446964,
|
| 527 |
+
"learning_rate": 3.558139534883721e-07,
|
| 528 |
+
"logits/chosen": -2.578672409057617,
|
| 529 |
+
"logits/rejected": -2.5800602436065674,
|
| 530 |
+
"logps/chosen": -61.31939697265625,
|
| 531 |
+
"logps/rejected": -86.94207000732422,
|
| 532 |
+
"loss": 0.4017,
|
| 533 |
+
"rewards/accuracies": 0.4375,
|
| 534 |
+
"rewards/chosen": 3.34814715385437,
|
| 535 |
+
"rewards/margins": 9.546571731567383,
|
| 536 |
+
"rewards/rejected": -6.198423862457275,
|
| 537 |
+
"step": 310
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.1149825783972125,
|
| 541 |
+
"grad_norm": 99.9815974317529,
|
| 542 |
+
"learning_rate": 3.46124031007752e-07,
|
| 543 |
+
"logits/chosen": -2.6197047233581543,
|
| 544 |
+
"logits/rejected": -2.6059112548828125,
|
| 545 |
+
"logps/chosen": -68.153076171875,
|
| 546 |
+
"logps/rejected": -89.3502426147461,
|
| 547 |
+
"loss": 0.3832,
|
| 548 |
+
"rewards/accuracies": 0.46875,
|
| 549 |
+
"rewards/chosen": 3.07761812210083,
|
| 550 |
+
"rewards/margins": 8.979107856750488,
|
| 551 |
+
"rewards/rejected": -5.9014892578125,
|
| 552 |
+
"step": 320
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.1498257839721253,
|
| 556 |
+
"grad_norm": 65.16657207151003,
|
| 557 |
+
"learning_rate": 3.3643410852713177e-07,
|
| 558 |
+
"logits/chosen": -2.6080145835876465,
|
| 559 |
+
"logits/rejected": -2.584524631500244,
|
| 560 |
+
"logps/chosen": -77.22286224365234,
|
| 561 |
+
"logps/rejected": -88.59306335449219,
|
| 562 |
+
"loss": 0.3571,
|
| 563 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 564 |
+
"rewards/chosen": 3.6562671661376953,
|
| 565 |
+
"rewards/margins": 8.938417434692383,
|
| 566 |
+
"rewards/rejected": -5.282149791717529,
|
| 567 |
+
"step": 330
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.1846689895470384,
|
| 571 |
+
"grad_norm": 20.064441334453125,
|
| 572 |
+
"learning_rate": 3.267441860465116e-07,
|
| 573 |
+
"logits/chosen": -2.58263897895813,
|
| 574 |
+
"logits/rejected": -2.5866923332214355,
|
| 575 |
+
"logps/chosen": -76.30878448486328,
|
| 576 |
+
"logps/rejected": -106.47459411621094,
|
| 577 |
+
"loss": 0.3758,
|
| 578 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 579 |
+
"rewards/chosen": 3.9426658153533936,
|
| 580 |
+
"rewards/margins": 9.874300003051758,
|
| 581 |
+
"rewards/rejected": -5.931633949279785,
|
| 582 |
+
"step": 340
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.2195121951219512,
|
| 586 |
+
"grad_norm": 71.91053721699969,
|
| 587 |
+
"learning_rate": 3.170542635658915e-07,
|
| 588 |
+
"logits/chosen": -2.5975940227508545,
|
| 589 |
+
"logits/rejected": -2.565171003341675,
|
| 590 |
+
"logps/chosen": -61.74261474609375,
|
| 591 |
+
"logps/rejected": -74.67215728759766,
|
| 592 |
+
"loss": 0.3855,
|
| 593 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 594 |
+
"rewards/chosen": 3.632841110229492,
|
| 595 |
+
"rewards/margins": 8.399141311645508,
|
| 596 |
+
"rewards/rejected": -4.766300201416016,
|
| 597 |
+
"step": 350
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.254355400696864,
|
| 601 |
+
"grad_norm": 78.70073770358177,
|
| 602 |
+
"learning_rate": 3.073643410852713e-07,
|
| 603 |
+
"logits/chosen": -2.59340238571167,
|
| 604 |
+
"logits/rejected": -2.566377639770508,
|
| 605 |
+
"logps/chosen": -65.64441680908203,
|
| 606 |
+
"logps/rejected": -72.45703125,
|
| 607 |
+
"loss": 0.3676,
|
| 608 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 609 |
+
"rewards/chosen": 3.2890095710754395,
|
| 610 |
+
"rewards/margins": 6.889220237731934,
|
| 611 |
+
"rewards/rejected": -3.6002116203308105,
|
| 612 |
+
"step": 360
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.289198606271777,
|
| 616 |
+
"grad_norm": 42.412792949129724,
|
| 617 |
+
"learning_rate": 2.9767441860465116e-07,
|
| 618 |
+
"logits/chosen": -2.5649404525756836,
|
| 619 |
+
"logits/rejected": -2.5833497047424316,
|
| 620 |
+
"logps/chosen": -64.13664245605469,
|
| 621 |
+
"logps/rejected": -83.73193359375,
|
| 622 |
+
"loss": 0.399,
|
| 623 |
+
"rewards/accuracies": 0.46875,
|
| 624 |
+
"rewards/chosen": 3.6725857257843018,
|
| 625 |
+
"rewards/margins": 8.653525352478027,
|
| 626 |
+
"rewards/rejected": -4.980940818786621,
|
| 627 |
+
"step": 370
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 1.32404181184669,
|
| 631 |
+
"grad_norm": 89.8718675271705,
|
| 632 |
+
"learning_rate": 2.87984496124031e-07,
|
| 633 |
+
"logits/chosen": -2.576333999633789,
|
| 634 |
+
"logits/rejected": -2.5752272605895996,
|
| 635 |
+
"logps/chosen": -77.921875,
|
| 636 |
+
"logps/rejected": -98.92057800292969,
|
| 637 |
+
"loss": 0.3699,
|
| 638 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 639 |
+
"rewards/chosen": 4.744899749755859,
|
| 640 |
+
"rewards/margins": 11.345683097839355,
|
| 641 |
+
"rewards/rejected": -6.600783348083496,
|
| 642 |
+
"step": 380
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 1.3588850174216027,
|
| 646 |
+
"grad_norm": 53.88569450482403,
|
| 647 |
+
"learning_rate": 2.7829457364341084e-07,
|
| 648 |
+
"logits/chosen": -2.6578681468963623,
|
| 649 |
+
"logits/rejected": -2.6431689262390137,
|
| 650 |
+
"logps/chosen": -59.8434944152832,
|
| 651 |
+
"logps/rejected": -80.90740203857422,
|
| 652 |
+
"loss": 0.3745,
|
| 653 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 654 |
+
"rewards/chosen": 3.790491819381714,
|
| 655 |
+
"rewards/margins": 8.903260231018066,
|
| 656 |
+
"rewards/rejected": -5.112768650054932,
|
| 657 |
+
"step": 390
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.3937282229965158,
|
| 661 |
+
"grad_norm": 103.59808023860636,
|
| 662 |
+
"learning_rate": 2.686046511627907e-07,
|
| 663 |
+
"logits/chosen": -2.638619899749756,
|
| 664 |
+
"logits/rejected": -2.609290361404419,
|
| 665 |
+
"logps/chosen": -78.6235122680664,
|
| 666 |
+
"logps/rejected": -108.10832214355469,
|
| 667 |
+
"loss": 0.3838,
|
| 668 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 669 |
+
"rewards/chosen": 3.987473249435425,
|
| 670 |
+
"rewards/margins": 9.130887985229492,
|
| 671 |
+
"rewards/rejected": -5.1434149742126465,
|
| 672 |
+
"step": 400
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.3937282229965158,
|
| 676 |
+
"eval_logits/chosen": -2.6421802043914795,
|
| 677 |
+
"eval_logits/rejected": -2.626558542251587,
|
| 678 |
+
"eval_logps/chosen": -73.66181945800781,
|
| 679 |
+
"eval_logps/rejected": -82.45710754394531,
|
| 680 |
+
"eval_loss": 0.871368408203125,
|
| 681 |
+
"eval_rewards/accuracies": 0.329365074634552,
|
| 682 |
+
"eval_rewards/chosen": 0.4395846724510193,
|
| 683 |
+
"eval_rewards/margins": 1.1438220739364624,
|
| 684 |
+
"eval_rewards/rejected": -0.7042374610900879,
|
| 685 |
+
"eval_runtime": 113.5108,
|
| 686 |
+
"eval_samples_per_second": 17.619,
|
| 687 |
+
"eval_steps_per_second": 0.555,
|
| 688 |
+
"step": 400
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 1.4285714285714286,
|
| 692 |
+
"grad_norm": 0.4652582217397593,
|
| 693 |
+
"learning_rate": 2.589147286821705e-07,
|
| 694 |
+
"logits/chosen": -2.625115156173706,
|
| 695 |
+
"logits/rejected": -2.6144914627075195,
|
| 696 |
+
"logps/chosen": -74.08203887939453,
|
| 697 |
+
"logps/rejected": -89.65892791748047,
|
| 698 |
+
"loss": 0.5043,
|
| 699 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 700 |
+
"rewards/chosen": 4.521576881408691,
|
| 701 |
+
"rewards/margins": 9.441361427307129,
|
| 702 |
+
"rewards/rejected": -4.919784069061279,
|
| 703 |
+
"step": 410
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.4634146341463414,
|
| 707 |
+
"grad_norm": 14.542963331044218,
|
| 708 |
+
"learning_rate": 2.492248062015504e-07,
|
| 709 |
+
"logits/chosen": -2.6700663566589355,
|
| 710 |
+
"logits/rejected": -2.671051502227783,
|
| 711 |
+
"logps/chosen": -70.02223205566406,
|
| 712 |
+
"logps/rejected": -93.91789245605469,
|
| 713 |
+
"loss": 0.3962,
|
| 714 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 715 |
+
"rewards/chosen": 3.217597484588623,
|
| 716 |
+
"rewards/margins": 8.012170791625977,
|
| 717 |
+
"rewards/rejected": -4.794573783874512,
|
| 718 |
+
"step": 420
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.4982578397212545,
|
| 722 |
+
"grad_norm": 7.616609761831896,
|
| 723 |
+
"learning_rate": 2.3953488372093024e-07,
|
| 724 |
+
"logits/chosen": -2.6445345878601074,
|
| 725 |
+
"logits/rejected": -2.630586624145508,
|
| 726 |
+
"logps/chosen": -61.7302360534668,
|
| 727 |
+
"logps/rejected": -81.66353607177734,
|
| 728 |
+
"loss": 0.3812,
|
| 729 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 730 |
+
"rewards/chosen": 3.6740145683288574,
|
| 731 |
+
"rewards/margins": 8.53648567199707,
|
| 732 |
+
"rewards/rejected": -4.862471580505371,
|
| 733 |
+
"step": 430
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.533101045296167,
|
| 737 |
+
"grad_norm": 19.40227524240624,
|
| 738 |
+
"learning_rate": 2.2984496124031007e-07,
|
| 739 |
+
"logits/chosen": -2.639118194580078,
|
| 740 |
+
"logits/rejected": -2.648719310760498,
|
| 741 |
+
"logps/chosen": -61.67310333251953,
|
| 742 |
+
"logps/rejected": -81.26171875,
|
| 743 |
+
"loss": 0.3814,
|
| 744 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 745 |
+
"rewards/chosen": 4.401946067810059,
|
| 746 |
+
"rewards/margins": 8.5157470703125,
|
| 747 |
+
"rewards/rejected": -4.113801002502441,
|
| 748 |
+
"step": 440
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.5679442508710801,
|
| 752 |
+
"grad_norm": 11.176099808562324,
|
| 753 |
+
"learning_rate": 2.201550387596899e-07,
|
| 754 |
+
"logits/chosen": -2.7013635635375977,
|
| 755 |
+
"logits/rejected": -2.6666698455810547,
|
| 756 |
+
"logps/chosen": -81.5969009399414,
|
| 757 |
+
"logps/rejected": -96.87474060058594,
|
| 758 |
+
"loss": 0.3926,
|
| 759 |
+
"rewards/accuracies": 0.518750011920929,
|
| 760 |
+
"rewards/chosen": 4.899897575378418,
|
| 761 |
+
"rewards/margins": 10.683164596557617,
|
| 762 |
+
"rewards/rejected": -5.783267021179199,
|
| 763 |
+
"step": 450
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.6027874564459932,
|
| 767 |
+
"grad_norm": 123.87369972777346,
|
| 768 |
+
"learning_rate": 2.1046511627906974e-07,
|
| 769 |
+
"logits/chosen": -2.6888933181762695,
|
| 770 |
+
"logits/rejected": -2.6685612201690674,
|
| 771 |
+
"logps/chosen": -66.90978240966797,
|
| 772 |
+
"logps/rejected": -88.9155044555664,
|
| 773 |
+
"loss": 0.4047,
|
| 774 |
+
"rewards/accuracies": 0.5,
|
| 775 |
+
"rewards/chosen": 4.9129533767700195,
|
| 776 |
+
"rewards/margins": 11.225677490234375,
|
| 777 |
+
"rewards/rejected": -6.312723636627197,
|
| 778 |
+
"step": 460
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.6376306620209058,
|
| 782 |
+
"grad_norm": 3.43139199879915,
|
| 783 |
+
"learning_rate": 2.0077519379844966e-07,
|
| 784 |
+
"logits/chosen": -2.6837058067321777,
|
| 785 |
+
"logits/rejected": -2.674848794937134,
|
| 786 |
+
"logps/chosen": -54.61207962036133,
|
| 787 |
+
"logps/rejected": -79.80362701416016,
|
| 788 |
+
"loss": 0.3979,
|
| 789 |
+
"rewards/accuracies": 0.40625,
|
| 790 |
+
"rewards/chosen": 3.5458691120147705,
|
| 791 |
+
"rewards/margins": 7.8562331199646,
|
| 792 |
+
"rewards/rejected": -4.310364246368408,
|
| 793 |
+
"step": 470
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.6724738675958188,
|
| 797 |
+
"grad_norm": 12.439142544759255,
|
| 798 |
+
"learning_rate": 1.9108527131782944e-07,
|
| 799 |
+
"logits/chosen": -2.695263385772705,
|
| 800 |
+
"logits/rejected": -2.6781885623931885,
|
| 801 |
+
"logps/chosen": -48.82516860961914,
|
| 802 |
+
"logps/rejected": -57.3747444152832,
|
| 803 |
+
"loss": 0.4033,
|
| 804 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 805 |
+
"rewards/chosen": 2.775264024734497,
|
| 806 |
+
"rewards/margins": 5.633866786956787,
|
| 807 |
+
"rewards/rejected": -2.8586020469665527,
|
| 808 |
+
"step": 480
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 1.7073170731707317,
|
| 812 |
+
"grad_norm": 40.080099001713826,
|
| 813 |
+
"learning_rate": 1.8139534883720925e-07,
|
| 814 |
+
"logits/chosen": -2.6758790016174316,
|
| 815 |
+
"logits/rejected": -2.6650328636169434,
|
| 816 |
+
"logps/chosen": -66.64186096191406,
|
| 817 |
+
"logps/rejected": -78.38505554199219,
|
| 818 |
+
"loss": 0.4696,
|
| 819 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 820 |
+
"rewards/chosen": 3.4873664379119873,
|
| 821 |
+
"rewards/margins": 7.898496150970459,
|
| 822 |
+
"rewards/rejected": -4.411130428314209,
|
| 823 |
+
"step": 490
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.7421602787456445,
|
| 827 |
+
"grad_norm": 47.95492405495267,
|
| 828 |
+
"learning_rate": 1.7170542635658914e-07,
|
| 829 |
+
"logits/chosen": -2.593526601791382,
|
| 830 |
+
"logits/rejected": -2.5879125595092773,
|
| 831 |
+
"logps/chosen": -68.72013854980469,
|
| 832 |
+
"logps/rejected": -92.43871307373047,
|
| 833 |
+
"loss": 0.371,
|
| 834 |
+
"rewards/accuracies": 0.46875,
|
| 835 |
+
"rewards/chosen": 4.010916709899902,
|
| 836 |
+
"rewards/margins": 8.630084991455078,
|
| 837 |
+
"rewards/rejected": -4.619168281555176,
|
| 838 |
+
"step": 500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.7421602787456445,
|
| 842 |
+
"eval_logits/chosen": -2.7067761421203613,
|
| 843 |
+
"eval_logits/rejected": -2.691006898880005,
|
| 844 |
+
"eval_logps/chosen": -73.18505859375,
|
| 845 |
+
"eval_logps/rejected": -82.15363311767578,
|
| 846 |
+
"eval_loss": 0.863908588886261,
|
| 847 |
+
"eval_rewards/accuracies": 0.3392857015132904,
|
| 848 |
+
"eval_rewards/chosen": 0.6922710537910461,
|
| 849 |
+
"eval_rewards/margins": 1.2356635332107544,
|
| 850 |
+
"eval_rewards/rejected": -0.5433923602104187,
|
| 851 |
+
"eval_runtime": 113.7753,
|
| 852 |
+
"eval_samples_per_second": 17.579,
|
| 853 |
+
"eval_steps_per_second": 0.554,
|
| 854 |
+
"step": 500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.7770034843205575,
|
| 858 |
+
"grad_norm": 43.89133813372592,
|
| 859 |
+
"learning_rate": 1.6201550387596898e-07,
|
| 860 |
+
"logits/chosen": -2.665417432785034,
|
| 861 |
+
"logits/rejected": -2.647148609161377,
|
| 862 |
+
"logps/chosen": -63.23058319091797,
|
| 863 |
+
"logps/rejected": -77.56340026855469,
|
| 864 |
+
"loss": 0.3821,
|
| 865 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 866 |
+
"rewards/chosen": 3.9886889457702637,
|
| 867 |
+
"rewards/margins": 8.217727661132812,
|
| 868 |
+
"rewards/rejected": -4.229039192199707,
|
| 869 |
+
"step": 510
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.8118466898954704,
|
| 873 |
+
"grad_norm": 26.582330424497826,
|
| 874 |
+
"learning_rate": 1.523255813953488e-07,
|
| 875 |
+
"logits/chosen": -2.671867847442627,
|
| 876 |
+
"logits/rejected": -2.666949510574341,
|
| 877 |
+
"logps/chosen": -66.01771545410156,
|
| 878 |
+
"logps/rejected": -86.70332336425781,
|
| 879 |
+
"loss": 0.382,
|
| 880 |
+
"rewards/accuracies": 0.46875,
|
| 881 |
+
"rewards/chosen": 4.374355316162109,
|
| 882 |
+
"rewards/margins": 9.404329299926758,
|
| 883 |
+
"rewards/rejected": -5.029973983764648,
|
| 884 |
+
"step": 520
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.8466898954703832,
|
| 888 |
+
"grad_norm": 16.01608311357976,
|
| 889 |
+
"learning_rate": 1.426356589147287e-07,
|
| 890 |
+
"logits/chosen": -2.6440272331237793,
|
| 891 |
+
"logits/rejected": -2.6383345127105713,
|
| 892 |
+
"logps/chosen": -63.14166259765625,
|
| 893 |
+
"logps/rejected": -80.24067687988281,
|
| 894 |
+
"loss": 0.3916,
|
| 895 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 896 |
+
"rewards/chosen": 4.34307336807251,
|
| 897 |
+
"rewards/margins": 9.047937393188477,
|
| 898 |
+
"rewards/rejected": -4.704863548278809,
|
| 899 |
+
"step": 530
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.8815331010452963,
|
| 903 |
+
"grad_norm": 503.79684025145957,
|
| 904 |
+
"learning_rate": 1.3294573643410851e-07,
|
| 905 |
+
"logits/chosen": -2.6588096618652344,
|
| 906 |
+
"logits/rejected": -2.6732683181762695,
|
| 907 |
+
"logps/chosen": -53.71875,
|
| 908 |
+
"logps/rejected": -84.29718780517578,
|
| 909 |
+
"loss": 0.46,
|
| 910 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 911 |
+
"rewards/chosen": 5.3063740730285645,
|
| 912 |
+
"rewards/margins": 11.112980842590332,
|
| 913 |
+
"rewards/rejected": -5.806607246398926,
|
| 914 |
+
"step": 540
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.916376306620209,
|
| 918 |
+
"grad_norm": 1.5289785290366227,
|
| 919 |
+
"learning_rate": 1.2325581395348835e-07,
|
| 920 |
+
"logits/chosen": -2.636369228363037,
|
| 921 |
+
"logits/rejected": -2.6182668209075928,
|
| 922 |
+
"logps/chosen": -79.08727264404297,
|
| 923 |
+
"logps/rejected": -97.51543426513672,
|
| 924 |
+
"loss": 0.3689,
|
| 925 |
+
"rewards/accuracies": 0.5,
|
| 926 |
+
"rewards/chosen": 6.226182460784912,
|
| 927 |
+
"rewards/margins": 11.859712600708008,
|
| 928 |
+
"rewards/rejected": -5.633530139923096,
|
| 929 |
+
"step": 550
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 1.951219512195122,
|
| 933 |
+
"grad_norm": 96.76132190493311,
|
| 934 |
+
"learning_rate": 1.1356589147286824e-07,
|
| 935 |
+
"logits/chosen": -2.64457368850708,
|
| 936 |
+
"logits/rejected": -2.651458740234375,
|
| 937 |
+
"logps/chosen": -55.0025520324707,
|
| 938 |
+
"logps/rejected": -76.64137268066406,
|
| 939 |
+
"loss": 0.3985,
|
| 940 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 941 |
+
"rewards/chosen": 4.620251655578613,
|
| 942 |
+
"rewards/margins": 7.983218193054199,
|
| 943 |
+
"rewards/rejected": -3.362966537475586,
|
| 944 |
+
"step": 560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.986062717770035,
|
| 948 |
+
"grad_norm": 13.582049528683124,
|
| 949 |
+
"learning_rate": 1.0387596899224806e-07,
|
| 950 |
+
"logits/chosen": -2.7313754558563232,
|
| 951 |
+
"logits/rejected": -2.702322483062744,
|
| 952 |
+
"logps/chosen": -56.71698760986328,
|
| 953 |
+
"logps/rejected": -67.86329650878906,
|
| 954 |
+
"loss": 0.3789,
|
| 955 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 956 |
+
"rewards/chosen": 4.503007411956787,
|
| 957 |
+
"rewards/margins": 7.31237268447876,
|
| 958 |
+
"rewards/rejected": -2.8093647956848145,
|
| 959 |
+
"step": 570
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 2.0,
|
| 963 |
+
"step": 574,
|
| 964 |
+
"total_flos": 0.0,
|
| 965 |
+
"train_loss": 0.5567006942287139,
|
| 966 |
+
"train_runtime": 6469.9642,
|
| 967 |
+
"train_samples_per_second": 5.669,
|
| 968 |
+
"train_steps_per_second": 0.089
|
| 969 |
+
}
|
| 970 |
+
],
|
| 971 |
+
"logging_steps": 10,
|
| 972 |
+
"max_steps": 574,
|
| 973 |
+
"num_input_tokens_seen": 0,
|
| 974 |
+
"num_train_epochs": 2,
|
| 975 |
+
"save_steps": 100,
|
| 976 |
+
"stateful_callbacks": {
|
| 977 |
+
"TrainerControl": {
|
| 978 |
+
"args": {
|
| 979 |
+
"should_epoch_stop": false,
|
| 980 |
+
"should_evaluate": false,
|
| 981 |
+
"should_log": false,
|
| 982 |
+
"should_save": true,
|
| 983 |
+
"should_training_stop": true
|
| 984 |
+
},
|
| 985 |
+
"attributes": {}
|
| 986 |
+
}
|
| 987 |
+
},
|
| 988 |
+
"total_flos": 0.0,
|
| 989 |
+
"train_batch_size": 8,
|
| 990 |
+
"trial_name": null,
|
| 991 |
+
"trial_params": null
|
| 992 |
+
}
|