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-2e-07-0.5-linear-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-2e-07-0.5-linear-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: 0.6465
|
| 22 |
+
- Rewards/chosen: 0.3072
|
| 23 |
+
- Rewards/rejected: -0.2969
|
| 24 |
+
- Rewards/accuracies: 0.3710
|
| 25 |
+
- Rewards/margins: 0.6040
|
| 26 |
+
- Logps/rejected: -81.7221
|
| 27 |
+
- Logps/chosen: -73.8769
|
| 28 |
+
- Logits/rejected: -2.5543
|
| 29 |
+
- Logits/chosen: -2.5709
|
| 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: 2e-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: linear
|
| 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 |
+
| 0.6562 | 0.3484 | 100 | 0.6328 | 0.6802 | 0.3841 | 0.3552 | 0.2961 | -80.3602 | -73.1308 | -2.5405 | -2.5565 |
|
| 67 |
+
| 0.6601 | 0.6969 | 200 | 0.6410 | 0.2989 | -0.0897 | 0.3452 | 0.3887 | -81.3078 | -73.8934 | -2.5169 | -2.5332 |
|
| 68 |
+
| 0.4195 | 1.0453 | 300 | 0.6371 | 0.6242 | 0.1593 | 0.3532 | 0.4648 | -80.8097 | -73.2429 | -2.5193 | -2.5354 |
|
| 69 |
+
| 0.3956 | 1.3937 | 400 | 0.6460 | 0.4324 | -0.1472 | 0.3631 | 0.5796 | -81.4227 | -73.6264 | -2.5378 | -2.5541 |
|
| 70 |
+
| 0.3945 | 1.7422 | 500 | 0.6465 | 0.3072 | -0.2969 | 0.3710 | 0.6040 | -81.7221 | -73.8769 | -2.5543 | -2.5709 |
|
| 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": 0.5264799162901237,
|
| 5 |
+
"train_runtime": 6499.6728,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.643,
|
| 8 |
+
"train_steps_per_second": 0.088
|
| 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": 0.5264799162901237,
|
| 5 |
+
"train_runtime": 6499.6728,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.643,
|
| 8 |
+
"train_steps_per_second": 0.088
|
| 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": 289.1508698329696,
|
| 14 |
+
"learning_rate": 3.4482758620689654e-09,
|
| 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": 277.24379710378327,
|
| 29 |
+
"learning_rate": 3.448275862068966e-08,
|
| 30 |
+
"logits/chosen": -2.5635900497436523,
|
| 31 |
+
"logits/rejected": -2.562225818634033,
|
| 32 |
+
"logps/chosen": -59.68327331542969,
|
| 33 |
+
"logps/rejected": -73.39383697509766,
|
| 34 |
+
"loss": 0.6974,
|
| 35 |
+
"rewards/accuracies": 0.1875,
|
| 36 |
+
"rewards/chosen": -0.01181174349039793,
|
| 37 |
+
"rewards/margins": -0.0053515201434493065,
|
| 38 |
+
"rewards/rejected": -0.006460222881287336,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.06968641114982578,
|
| 43 |
+
"grad_norm": 357.5434709312531,
|
| 44 |
+
"learning_rate": 6.896551724137931e-08,
|
| 45 |
+
"logits/chosen": -2.6049184799194336,
|
| 46 |
+
"logits/rejected": -2.5635530948638916,
|
| 47 |
+
"logps/chosen": -104.09075927734375,
|
| 48 |
+
"logps/rejected": -94.9400634765625,
|
| 49 |
+
"loss": 0.6899,
|
| 50 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 51 |
+
"rewards/chosen": 0.01815590262413025,
|
| 52 |
+
"rewards/margins": 0.04113120958209038,
|
| 53 |
+
"rewards/rejected": -0.02297530695796013,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.10452961672473868,
|
| 58 |
+
"grad_norm": 337.4139606266084,
|
| 59 |
+
"learning_rate": 1.0344827586206897e-07,
|
| 60 |
+
"logits/chosen": -2.5913662910461426,
|
| 61 |
+
"logits/rejected": -2.5713112354278564,
|
| 62 |
+
"logps/chosen": -82.475341796875,
|
| 63 |
+
"logps/rejected": -91.52882385253906,
|
| 64 |
+
"loss": 0.6923,
|
| 65 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 66 |
+
"rewards/chosen": 0.01061153132468462,
|
| 67 |
+
"rewards/margins": 0.004934412427246571,
|
| 68 |
+
"rewards/rejected": 0.005677118897438049,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13937282229965156,
|
| 73 |
+
"grad_norm": 282.9755513914394,
|
| 74 |
+
"learning_rate": 1.3793103448275863e-07,
|
| 75 |
+
"logits/chosen": -2.4979004859924316,
|
| 76 |
+
"logits/rejected": -2.495636463165283,
|
| 77 |
+
"logps/chosen": -77.93832397460938,
|
| 78 |
+
"logps/rejected": -73.0629653930664,
|
| 79 |
+
"loss": 0.6791,
|
| 80 |
+
"rewards/accuracies": 0.20624999701976776,
|
| 81 |
+
"rewards/chosen": -0.00995340384542942,
|
| 82 |
+
"rewards/margins": 0.020863836631178856,
|
| 83 |
+
"rewards/rejected": -0.030817240476608276,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.17421602787456447,
|
| 88 |
+
"grad_norm": 245.53657846887563,
|
| 89 |
+
"learning_rate": 1.7241379310344825e-07,
|
| 90 |
+
"logits/chosen": -2.533763885498047,
|
| 91 |
+
"logits/rejected": -2.537358522415161,
|
| 92 |
+
"logps/chosen": -63.84333419799805,
|
| 93 |
+
"logps/rejected": -76.09790802001953,
|
| 94 |
+
"loss": 0.6717,
|
| 95 |
+
"rewards/accuracies": 0.2562499940395355,
|
| 96 |
+
"rewards/chosen": 0.047906167805194855,
|
| 97 |
+
"rewards/margins": 0.04513678327202797,
|
| 98 |
+
"rewards/rejected": 0.002769380807876587,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.20905923344947736,
|
| 103 |
+
"grad_norm": 247.7041408944673,
|
| 104 |
+
"learning_rate": 1.9922480620155037e-07,
|
| 105 |
+
"logits/chosen": -2.506993532180786,
|
| 106 |
+
"logits/rejected": -2.5004165172576904,
|
| 107 |
+
"logps/chosen": -72.5146255493164,
|
| 108 |
+
"logps/rejected": -67.91732025146484,
|
| 109 |
+
"loss": 0.6647,
|
| 110 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 111 |
+
"rewards/chosen": 0.34594517946243286,
|
| 112 |
+
"rewards/margins": 0.1003229171037674,
|
| 113 |
+
"rewards/rejected": 0.24562230706214905,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.24390243902439024,
|
| 118 |
+
"grad_norm": 249.16248968780815,
|
| 119 |
+
"learning_rate": 1.9534883720930232e-07,
|
| 120 |
+
"logits/chosen": -2.528493881225586,
|
| 121 |
+
"logits/rejected": -2.5241379737854004,
|
| 122 |
+
"logps/chosen": -62.91923904418945,
|
| 123 |
+
"logps/rejected": -67.45994567871094,
|
| 124 |
+
"loss": 0.6534,
|
| 125 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 126 |
+
"rewards/chosen": 0.6584776043891907,
|
| 127 |
+
"rewards/margins": 0.17514197528362274,
|
| 128 |
+
"rewards/rejected": 0.4833356738090515,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.2787456445993031,
|
| 133 |
+
"grad_norm": 307.9296174602441,
|
| 134 |
+
"learning_rate": 1.9147286821705426e-07,
|
| 135 |
+
"logits/chosen": -2.4762024879455566,
|
| 136 |
+
"logits/rejected": -2.4663474559783936,
|
| 137 |
+
"logps/chosen": -74.05433654785156,
|
| 138 |
+
"logps/rejected": -76.48423767089844,
|
| 139 |
+
"loss": 0.6613,
|
| 140 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 141 |
+
"rewards/chosen": 0.805270791053772,
|
| 142 |
+
"rewards/margins": 0.2899132966995239,
|
| 143 |
+
"rewards/rejected": 0.5153574347496033,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.313588850174216,
|
| 148 |
+
"grad_norm": 236.73851878615983,
|
| 149 |
+
"learning_rate": 1.8759689922480618e-07,
|
| 150 |
+
"logits/chosen": -2.4865849018096924,
|
| 151 |
+
"logits/rejected": -2.5006980895996094,
|
| 152 |
+
"logps/chosen": -64.770263671875,
|
| 153 |
+
"logps/rejected": -69.27301788330078,
|
| 154 |
+
"loss": 0.6551,
|
| 155 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 156 |
+
"rewards/chosen": 0.6060104966163635,
|
| 157 |
+
"rewards/margins": 0.18549713492393494,
|
| 158 |
+
"rewards/rejected": 0.420513391494751,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.34843205574912894,
|
| 163 |
+
"grad_norm": 302.9071215603059,
|
| 164 |
+
"learning_rate": 1.8372093023255813e-07,
|
| 165 |
+
"logits/chosen": -2.4684031009674072,
|
| 166 |
+
"logits/rejected": -2.4694905281066895,
|
| 167 |
+
"logps/chosen": -73.97297668457031,
|
| 168 |
+
"logps/rejected": -80.28447723388672,
|
| 169 |
+
"loss": 0.6562,
|
| 170 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 171 |
+
"rewards/chosen": 0.649095356464386,
|
| 172 |
+
"rewards/margins": 0.26550906896591187,
|
| 173 |
+
"rewards/rejected": 0.3835863173007965,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.34843205574912894,
|
| 178 |
+
"eval_logits/chosen": -2.5565242767333984,
|
| 179 |
+
"eval_logits/rejected": -2.540536880493164,
|
| 180 |
+
"eval_logps/chosen": -73.13079833984375,
|
| 181 |
+
"eval_logps/rejected": -80.36016845703125,
|
| 182 |
+
"eval_loss": 0.6328476667404175,
|
| 183 |
+
"eval_rewards/accuracies": 0.3551587164402008,
|
| 184 |
+
"eval_rewards/chosen": 0.6802141070365906,
|
| 185 |
+
"eval_rewards/margins": 0.296122670173645,
|
| 186 |
+
"eval_rewards/rejected": 0.3840915858745575,
|
| 187 |
+
"eval_runtime": 113.5393,
|
| 188 |
+
"eval_samples_per_second": 17.615,
|
| 189 |
+
"eval_steps_per_second": 0.555,
|
| 190 |
+
"step": 100
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.3832752613240418,
|
| 194 |
+
"grad_norm": 321.32489311642775,
|
| 195 |
+
"learning_rate": 1.7984496124031007e-07,
|
| 196 |
+
"logits/chosen": -2.497744560241699,
|
| 197 |
+
"logits/rejected": -2.4591681957244873,
|
| 198 |
+
"logps/chosen": -72.22078704833984,
|
| 199 |
+
"logps/rejected": -62.5744514465332,
|
| 200 |
+
"loss": 0.6334,
|
| 201 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 202 |
+
"rewards/chosen": 0.474620521068573,
|
| 203 |
+
"rewards/margins": 0.256552129983902,
|
| 204 |
+
"rewards/rejected": 0.21806840598583221,
|
| 205 |
+
"step": 110
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.4181184668989547,
|
| 209 |
+
"grad_norm": 224.1076153559292,
|
| 210 |
+
"learning_rate": 1.7596899224806202e-07,
|
| 211 |
+
"logits/chosen": -2.520296573638916,
|
| 212 |
+
"logits/rejected": -2.488671064376831,
|
| 213 |
+
"logps/chosen": -76.48248291015625,
|
| 214 |
+
"logps/rejected": -66.29643249511719,
|
| 215 |
+
"loss": 0.619,
|
| 216 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 217 |
+
"rewards/chosen": 0.48954588174819946,
|
| 218 |
+
"rewards/margins": 0.30197009444236755,
|
| 219 |
+
"rewards/rejected": 0.1875758022069931,
|
| 220 |
+
"step": 120
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4529616724738676,
|
| 224 |
+
"grad_norm": 339.42643355240523,
|
| 225 |
+
"learning_rate": 1.7209302325581394e-07,
|
| 226 |
+
"logits/chosen": -2.5603480339050293,
|
| 227 |
+
"logits/rejected": -2.5410096645355225,
|
| 228 |
+
"logps/chosen": -83.0821533203125,
|
| 229 |
+
"logps/rejected": -87.61650085449219,
|
| 230 |
+
"loss": 0.6341,
|
| 231 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 232 |
+
"rewards/chosen": 0.3993155062198639,
|
| 233 |
+
"rewards/margins": 0.4533071517944336,
|
| 234 |
+
"rewards/rejected": -0.053991664201021194,
|
| 235 |
+
"step": 130
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.4878048780487805,
|
| 239 |
+
"grad_norm": 307.50496227541765,
|
| 240 |
+
"learning_rate": 1.6821705426356588e-07,
|
| 241 |
+
"logits/chosen": -2.459794282913208,
|
| 242 |
+
"logits/rejected": -2.447605609893799,
|
| 243 |
+
"logps/chosen": -80.36363220214844,
|
| 244 |
+
"logps/rejected": -70.87696838378906,
|
| 245 |
+
"loss": 0.616,
|
| 246 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 247 |
+
"rewards/chosen": 0.3909004330635071,
|
| 248 |
+
"rewards/margins": 0.4472619891166687,
|
| 249 |
+
"rewards/rejected": -0.05636156350374222,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.5226480836236934,
|
| 254 |
+
"grad_norm": 305.3753178775468,
|
| 255 |
+
"learning_rate": 1.6434108527131783e-07,
|
| 256 |
+
"logits/chosen": -2.520991086959839,
|
| 257 |
+
"logits/rejected": -2.47597074508667,
|
| 258 |
+
"logps/chosen": -78.84916687011719,
|
| 259 |
+
"logps/rejected": -79.61804962158203,
|
| 260 |
+
"loss": 0.6534,
|
| 261 |
+
"rewards/accuracies": 0.29374998807907104,
|
| 262 |
+
"rewards/chosen": 0.31758052110671997,
|
| 263 |
+
"rewards/margins": 0.35121217370033264,
|
| 264 |
+
"rewards/rejected": -0.03363170847296715,
|
| 265 |
+
"step": 150
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.5574912891986062,
|
| 269 |
+
"grad_norm": 267.54053358304463,
|
| 270 |
+
"learning_rate": 1.6046511627906975e-07,
|
| 271 |
+
"logits/chosen": -2.4771218299865723,
|
| 272 |
+
"logits/rejected": -2.498429298400879,
|
| 273 |
+
"logps/chosen": -63.93762969970703,
|
| 274 |
+
"logps/rejected": -71.9593734741211,
|
| 275 |
+
"loss": 0.6535,
|
| 276 |
+
"rewards/accuracies": 0.28125,
|
| 277 |
+
"rewards/chosen": 0.29360949993133545,
|
| 278 |
+
"rewards/margins": 0.2868410050868988,
|
| 279 |
+
"rewards/rejected": 0.0067685008980333805,
|
| 280 |
+
"step": 160
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.5923344947735192,
|
| 284 |
+
"grad_norm": 274.1930461210022,
|
| 285 |
+
"learning_rate": 1.565891472868217e-07,
|
| 286 |
+
"logits/chosen": -2.4762015342712402,
|
| 287 |
+
"logits/rejected": -2.461243152618408,
|
| 288 |
+
"logps/chosen": -68.60774230957031,
|
| 289 |
+
"logps/rejected": -76.81765747070312,
|
| 290 |
+
"loss": 0.6071,
|
| 291 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 292 |
+
"rewards/chosen": 0.17982278764247894,
|
| 293 |
+
"rewards/margins": 0.43163204193115234,
|
| 294 |
+
"rewards/rejected": -0.251809298992157,
|
| 295 |
+
"step": 170
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.627177700348432,
|
| 299 |
+
"grad_norm": 274.26901021006046,
|
| 300 |
+
"learning_rate": 1.5271317829457364e-07,
|
| 301 |
+
"logits/chosen": -2.5262434482574463,
|
| 302 |
+
"logits/rejected": -2.5153231620788574,
|
| 303 |
+
"logps/chosen": -91.26216125488281,
|
| 304 |
+
"logps/rejected": -86.61434936523438,
|
| 305 |
+
"loss": 0.6537,
|
| 306 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 307 |
+
"rewards/chosen": 0.29750674962997437,
|
| 308 |
+
"rewards/margins": 0.3494582176208496,
|
| 309 |
+
"rewards/rejected": -0.05195152759552002,
|
| 310 |
+
"step": 180
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.662020905923345,
|
| 314 |
+
"grad_norm": 207.64618032691078,
|
| 315 |
+
"learning_rate": 1.4883720930232558e-07,
|
| 316 |
+
"logits/chosen": -2.523658037185669,
|
| 317 |
+
"logits/rejected": -2.513747453689575,
|
| 318 |
+
"logps/chosen": -70.37347412109375,
|
| 319 |
+
"logps/rejected": -81.337646484375,
|
| 320 |
+
"loss": 0.6272,
|
| 321 |
+
"rewards/accuracies": 0.3125,
|
| 322 |
+
"rewards/chosen": 0.33900195360183716,
|
| 323 |
+
"rewards/margins": 0.24373552203178406,
|
| 324 |
+
"rewards/rejected": 0.0952664315700531,
|
| 325 |
+
"step": 190
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.6968641114982579,
|
| 329 |
+
"grad_norm": 364.4684926175276,
|
| 330 |
+
"learning_rate": 1.449612403100775e-07,
|
| 331 |
+
"logits/chosen": -2.5569560527801514,
|
| 332 |
+
"logits/rejected": -2.564101219177246,
|
| 333 |
+
"logps/chosen": -89.63420104980469,
|
| 334 |
+
"logps/rejected": -91.90037536621094,
|
| 335 |
+
"loss": 0.6601,
|
| 336 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 337 |
+
"rewards/chosen": 0.5110756754875183,
|
| 338 |
+
"rewards/margins": 0.4359167516231537,
|
| 339 |
+
"rewards/rejected": 0.07515887171030045,
|
| 340 |
+
"step": 200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.6968641114982579,
|
| 344 |
+
"eval_logits/chosen": -2.5332281589508057,
|
| 345 |
+
"eval_logits/rejected": -2.516913414001465,
|
| 346 |
+
"eval_logps/chosen": -73.89338684082031,
|
| 347 |
+
"eval_logps/rejected": -81.30782318115234,
|
| 348 |
+
"eval_loss": 0.6409674882888794,
|
| 349 |
+
"eval_rewards/accuracies": 0.3452380895614624,
|
| 350 |
+
"eval_rewards/chosen": 0.29892128705978394,
|
| 351 |
+
"eval_rewards/margins": 0.38865822553634644,
|
| 352 |
+
"eval_rewards/rejected": -0.08973691612482071,
|
| 353 |
+
"eval_runtime": 113.5083,
|
| 354 |
+
"eval_samples_per_second": 17.62,
|
| 355 |
+
"eval_steps_per_second": 0.555,
|
| 356 |
+
"step": 200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.7317073170731707,
|
| 360 |
+
"grad_norm": 418.8344130807906,
|
| 361 |
+
"learning_rate": 1.4108527131782945e-07,
|
| 362 |
+
"logits/chosen": -2.5373802185058594,
|
| 363 |
+
"logits/rejected": -2.510119915008545,
|
| 364 |
+
"logps/chosen": -68.75162506103516,
|
| 365 |
+
"logps/rejected": -63.84955978393555,
|
| 366 |
+
"loss": 0.6439,
|
| 367 |
+
"rewards/accuracies": 0.39375001192092896,
|
| 368 |
+
"rewards/chosen": 0.2837230861186981,
|
| 369 |
+
"rewards/margins": 0.5062949657440186,
|
| 370 |
+
"rewards/rejected": -0.22257189452648163,
|
| 371 |
+
"step": 210
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.7665505226480837,
|
| 375 |
+
"grad_norm": 227.78900163940662,
|
| 376 |
+
"learning_rate": 1.372093023255814e-07,
|
| 377 |
+
"logits/chosen": -2.5648019313812256,
|
| 378 |
+
"logits/rejected": -2.5423765182495117,
|
| 379 |
+
"logps/chosen": -72.84169006347656,
|
| 380 |
+
"logps/rejected": -71.51380920410156,
|
| 381 |
+
"loss": 0.644,
|
| 382 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 383 |
+
"rewards/chosen": 0.30824023485183716,
|
| 384 |
+
"rewards/margins": 0.23355619609355927,
|
| 385 |
+
"rewards/rejected": 0.0746840387582779,
|
| 386 |
+
"step": 220
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.8013937282229965,
|
| 390 |
+
"grad_norm": 280.8923204773476,
|
| 391 |
+
"learning_rate": 1.333333333333333e-07,
|
| 392 |
+
"logits/chosen": -2.56278920173645,
|
| 393 |
+
"logits/rejected": -2.5390899181365967,
|
| 394 |
+
"logps/chosen": -89.15130615234375,
|
| 395 |
+
"logps/rejected": -88.63081359863281,
|
| 396 |
+
"loss": 0.6323,
|
| 397 |
+
"rewards/accuracies": 0.40625,
|
| 398 |
+
"rewards/chosen": 0.4246680736541748,
|
| 399 |
+
"rewards/margins": 0.6656498908996582,
|
| 400 |
+
"rewards/rejected": -0.24098177254199982,
|
| 401 |
+
"step": 230
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.8362369337979094,
|
| 405 |
+
"grad_norm": 250.23928935872348,
|
| 406 |
+
"learning_rate": 1.2945736434108528e-07,
|
| 407 |
+
"logits/chosen": -2.5548999309539795,
|
| 408 |
+
"logits/rejected": -2.5180442333221436,
|
| 409 |
+
"logps/chosen": -86.33263397216797,
|
| 410 |
+
"logps/rejected": -80.53160095214844,
|
| 411 |
+
"loss": 0.6558,
|
| 412 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 413 |
+
"rewards/chosen": 0.20448681712150574,
|
| 414 |
+
"rewards/margins": 0.4022372364997864,
|
| 415 |
+
"rewards/rejected": -0.19775035977363586,
|
| 416 |
+
"step": 240
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.8710801393728222,
|
| 420 |
+
"grad_norm": 281.88116685137084,
|
| 421 |
+
"learning_rate": 1.255813953488372e-07,
|
| 422 |
+
"logits/chosen": -2.5691909790039062,
|
| 423 |
+
"logits/rejected": -2.53218412399292,
|
| 424 |
+
"logps/chosen": -95.07816314697266,
|
| 425 |
+
"logps/rejected": -90.45535278320312,
|
| 426 |
+
"loss": 0.5834,
|
| 427 |
+
"rewards/accuracies": 0.40625,
|
| 428 |
+
"rewards/chosen": 0.1855524331331253,
|
| 429 |
+
"rewards/margins": 0.44013509154319763,
|
| 430 |
+
"rewards/rejected": -0.25458264350891113,
|
| 431 |
+
"step": 250
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.9059233449477352,
|
| 435 |
+
"grad_norm": 217.05487038029386,
|
| 436 |
+
"learning_rate": 1.2170542635658915e-07,
|
| 437 |
+
"logits/chosen": -2.4733526706695557,
|
| 438 |
+
"logits/rejected": -2.4898457527160645,
|
| 439 |
+
"logps/chosen": -58.55949783325195,
|
| 440 |
+
"logps/rejected": -65.79776763916016,
|
| 441 |
+
"loss": 0.6288,
|
| 442 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 443 |
+
"rewards/chosen": 0.24493035674095154,
|
| 444 |
+
"rewards/margins": 0.329669326543808,
|
| 445 |
+
"rewards/rejected": -0.08473895490169525,
|
| 446 |
+
"step": 260
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.9407665505226481,
|
| 450 |
+
"grad_norm": 271.80312819553575,
|
| 451 |
+
"learning_rate": 1.1782945736434109e-07,
|
| 452 |
+
"logits/chosen": -2.5729527473449707,
|
| 453 |
+
"logits/rejected": -2.5736825466156006,
|
| 454 |
+
"logps/chosen": -68.29542541503906,
|
| 455 |
+
"logps/rejected": -83.07144927978516,
|
| 456 |
+
"loss": 0.6039,
|
| 457 |
+
"rewards/accuracies": 0.34375,
|
| 458 |
+
"rewards/chosen": 0.2923750579357147,
|
| 459 |
+
"rewards/margins": 0.562214732170105,
|
| 460 |
+
"rewards/rejected": -0.26983970403671265,
|
| 461 |
+
"step": 270
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.975609756097561,
|
| 465 |
+
"grad_norm": 238.59512203733584,
|
| 466 |
+
"learning_rate": 1.1395348837209302e-07,
|
| 467 |
+
"logits/chosen": -2.4834656715393066,
|
| 468 |
+
"logits/rejected": -2.458590030670166,
|
| 469 |
+
"logps/chosen": -67.38432312011719,
|
| 470 |
+
"logps/rejected": -71.17386627197266,
|
| 471 |
+
"loss": 0.6017,
|
| 472 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 473 |
+
"rewards/chosen": 0.22502359747886658,
|
| 474 |
+
"rewards/margins": 0.39475032687187195,
|
| 475 |
+
"rewards/rejected": -0.16972672939300537,
|
| 476 |
+
"step": 280
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.0104529616724738,
|
| 480 |
+
"grad_norm": 76.34654611961486,
|
| 481 |
+
"learning_rate": 1.1007751937984495e-07,
|
| 482 |
+
"logits/chosen": -2.4974358081817627,
|
| 483 |
+
"logits/rejected": -2.4653944969177246,
|
| 484 |
+
"logps/chosen": -70.81477355957031,
|
| 485 |
+
"logps/rejected": -65.4640121459961,
|
| 486 |
+
"loss": 0.5461,
|
| 487 |
+
"rewards/accuracies": 0.3812499940395355,
|
| 488 |
+
"rewards/chosen": 0.6264529228210449,
|
| 489 |
+
"rewards/margins": 0.9015191197395325,
|
| 490 |
+
"rewards/rejected": -0.27506622672080994,
|
| 491 |
+
"step": 290
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.0452961672473868,
|
| 495 |
+
"grad_norm": 39.056755757958314,
|
| 496 |
+
"learning_rate": 1.062015503875969e-07,
|
| 497 |
+
"logits/chosen": -2.51816987991333,
|
| 498 |
+
"logits/rejected": -2.5042262077331543,
|
| 499 |
+
"logps/chosen": -64.20988464355469,
|
| 500 |
+
"logps/rejected": -70.89491271972656,
|
| 501 |
+
"loss": 0.4195,
|
| 502 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 503 |
+
"rewards/chosen": 1.1695148944854736,
|
| 504 |
+
"rewards/margins": 2.031907320022583,
|
| 505 |
+
"rewards/rejected": -0.862392246723175,
|
| 506 |
+
"step": 300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.0452961672473868,
|
| 510 |
+
"eval_logits/chosen": -2.5354163646698,
|
| 511 |
+
"eval_logits/rejected": -2.519293785095215,
|
| 512 |
+
"eval_logps/chosen": -73.24293518066406,
|
| 513 |
+
"eval_logps/rejected": -80.80972290039062,
|
| 514 |
+
"eval_loss": 0.6371246576309204,
|
| 515 |
+
"eval_rewards/accuracies": 0.3531745970249176,
|
| 516 |
+
"eval_rewards/chosen": 0.6241527199745178,
|
| 517 |
+
"eval_rewards/margins": 0.4648365378379822,
|
| 518 |
+
"eval_rewards/rejected": 0.15931615233421326,
|
| 519 |
+
"eval_runtime": 113.4957,
|
| 520 |
+
"eval_samples_per_second": 17.622,
|
| 521 |
+
"eval_steps_per_second": 0.555,
|
| 522 |
+
"step": 300
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.0801393728222997,
|
| 526 |
+
"grad_norm": 33.225277343615126,
|
| 527 |
+
"learning_rate": 1.0232558139534883e-07,
|
| 528 |
+
"logits/chosen": -2.4929442405700684,
|
| 529 |
+
"logits/rejected": -2.498056650161743,
|
| 530 |
+
"logps/chosen": -65.01827239990234,
|
| 531 |
+
"logps/rejected": -77.01332092285156,
|
| 532 |
+
"loss": 0.4196,
|
| 533 |
+
"rewards/accuracies": 0.4375,
|
| 534 |
+
"rewards/chosen": 1.309188723564148,
|
| 535 |
+
"rewards/margins": 2.192389965057373,
|
| 536 |
+
"rewards/rejected": -0.8832012414932251,
|
| 537 |
+
"step": 310
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.1149825783972125,
|
| 541 |
+
"grad_norm": 73.59283724981043,
|
| 542 |
+
"learning_rate": 9.844961240310076e-08,
|
| 543 |
+
"logits/chosen": -2.544821262359619,
|
| 544 |
+
"logits/rejected": -2.5316672325134277,
|
| 545 |
+
"logps/chosen": -71.5367431640625,
|
| 546 |
+
"logps/rejected": -79.99516296386719,
|
| 547 |
+
"loss": 0.4155,
|
| 548 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 549 |
+
"rewards/chosen": 1.2115765810012817,
|
| 550 |
+
"rewards/margins": 2.1014885902404785,
|
| 551 |
+
"rewards/rejected": -0.8899120092391968,
|
| 552 |
+
"step": 320
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.1498257839721253,
|
| 556 |
+
"grad_norm": 61.862973963070324,
|
| 557 |
+
"learning_rate": 9.457364341085271e-08,
|
| 558 |
+
"logits/chosen": -2.5284900665283203,
|
| 559 |
+
"logits/rejected": -2.502081871032715,
|
| 560 |
+
"logps/chosen": -81.13185119628906,
|
| 561 |
+
"logps/rejected": -80.43305969238281,
|
| 562 |
+
"loss": 0.3845,
|
| 563 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 564 |
+
"rewards/chosen": 1.4948151111602783,
|
| 565 |
+
"rewards/margins": 2.3979744911193848,
|
| 566 |
+
"rewards/rejected": -0.9031593203544617,
|
| 567 |
+
"step": 330
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.1846689895470384,
|
| 571 |
+
"grad_norm": 38.944715898756485,
|
| 572 |
+
"learning_rate": 9.069767441860464e-08,
|
| 573 |
+
"logits/chosen": -2.5041089057922363,
|
| 574 |
+
"logits/rejected": -2.510469436645508,
|
| 575 |
+
"logps/chosen": -80.71260833740234,
|
| 576 |
+
"logps/rejected": -97.7466049194336,
|
| 577 |
+
"loss": 0.4026,
|
| 578 |
+
"rewards/accuracies": 0.550000011920929,
|
| 579 |
+
"rewards/chosen": 1.5175840854644775,
|
| 580 |
+
"rewards/margins": 2.749465227127075,
|
| 581 |
+
"rewards/rejected": -1.2318809032440186,
|
| 582 |
+
"step": 340
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.2195121951219512,
|
| 586 |
+
"grad_norm": 64.94626084684288,
|
| 587 |
+
"learning_rate": 8.682170542635659e-08,
|
| 588 |
+
"logits/chosen": -2.5372118949890137,
|
| 589 |
+
"logits/rejected": -2.503645896911621,
|
| 590 |
+
"logps/chosen": -66.34137725830078,
|
| 591 |
+
"logps/rejected": -67.35131072998047,
|
| 592 |
+
"loss": 0.4149,
|
| 593 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 594 |
+
"rewards/chosen": 1.1278231143951416,
|
| 595 |
+
"rewards/margins": 1.9639180898666382,
|
| 596 |
+
"rewards/rejected": -0.8360951542854309,
|
| 597 |
+
"step": 350
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.254355400696864,
|
| 601 |
+
"grad_norm": 72.87191418772136,
|
| 602 |
+
"learning_rate": 8.294573643410853e-08,
|
| 603 |
+
"logits/chosen": -2.529754877090454,
|
| 604 |
+
"logits/rejected": -2.4995744228363037,
|
| 605 |
+
"logps/chosen": -69.45609283447266,
|
| 606 |
+
"logps/rejected": -67.36495971679688,
|
| 607 |
+
"loss": 0.3984,
|
| 608 |
+
"rewards/accuracies": 0.4375,
|
| 609 |
+
"rewards/chosen": 1.1970014572143555,
|
| 610 |
+
"rewards/margins": 2.0473990440368652,
|
| 611 |
+
"rewards/rejected": -0.8503974080085754,
|
| 612 |
+
"step": 360
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.289198606271777,
|
| 616 |
+
"grad_norm": 78.06787405031129,
|
| 617 |
+
"learning_rate": 7.906976744186046e-08,
|
| 618 |
+
"logits/chosen": -2.498607635498047,
|
| 619 |
+
"logits/rejected": -2.5195744037628174,
|
| 620 |
+
"logps/chosen": -68.30110168457031,
|
| 621 |
+
"logps/rejected": -75.86735534667969,
|
| 622 |
+
"loss": 0.418,
|
| 623 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 624 |
+
"rewards/chosen": 1.382476568222046,
|
| 625 |
+
"rewards/margins": 2.1491870880126953,
|
| 626 |
+
"rewards/rejected": -0.7667104005813599,
|
| 627 |
+
"step": 370
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 1.32404181184669,
|
| 631 |
+
"grad_norm": 35.524129566350084,
|
| 632 |
+
"learning_rate": 7.51937984496124e-08,
|
| 633 |
+
"logits/chosen": -2.507997989654541,
|
| 634 |
+
"logits/rejected": -2.5067758560180664,
|
| 635 |
+
"logps/chosen": -82.98981475830078,
|
| 636 |
+
"logps/rejected": -88.70475769042969,
|
| 637 |
+
"loss": 0.3909,
|
| 638 |
+
"rewards/accuracies": 0.550000011920929,
|
| 639 |
+
"rewards/chosen": 1.942350149154663,
|
| 640 |
+
"rewards/margins": 3.0615932941436768,
|
| 641 |
+
"rewards/rejected": -1.1192431449890137,
|
| 642 |
+
"step": 380
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 1.3588850174216027,
|
| 646 |
+
"grad_norm": 71.79408635042864,
|
| 647 |
+
"learning_rate": 7.131782945736434e-08,
|
| 648 |
+
"logits/chosen": -2.5821242332458496,
|
| 649 |
+
"logits/rejected": -2.5696043968200684,
|
| 650 |
+
"logps/chosen": -64.58931732177734,
|
| 651 |
+
"logps/rejected": -73.18416595458984,
|
| 652 |
+
"loss": 0.3949,
|
| 653 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 654 |
+
"rewards/chosen": 1.203021764755249,
|
| 655 |
+
"rewards/margins": 2.1647653579711914,
|
| 656 |
+
"rewards/rejected": -0.9617435336112976,
|
| 657 |
+
"step": 390
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.3937282229965158,
|
| 661 |
+
"grad_norm": 40.565767402769495,
|
| 662 |
+
"learning_rate": 6.744186046511627e-08,
|
| 663 |
+
"logits/chosen": -2.553471088409424,
|
| 664 |
+
"logits/rejected": -2.5233709812164307,
|
| 665 |
+
"logps/chosen": -83.41809844970703,
|
| 666 |
+
"logps/rejected": -100.60682678222656,
|
| 667 |
+
"loss": 0.3956,
|
| 668 |
+
"rewards/accuracies": 0.518750011920929,
|
| 669 |
+
"rewards/chosen": 1.3644764423370361,
|
| 670 |
+
"rewards/margins": 2.466003179550171,
|
| 671 |
+
"rewards/rejected": -1.1015267372131348,
|
| 672 |
+
"step": 400
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.3937282229965158,
|
| 676 |
+
"eval_logits/chosen": -2.5541367530822754,
|
| 677 |
+
"eval_logits/rejected": -2.5378055572509766,
|
| 678 |
+
"eval_logps/chosen": -73.62639617919922,
|
| 679 |
+
"eval_logps/rejected": -81.42265319824219,
|
| 680 |
+
"eval_loss": 0.645966649055481,
|
| 681 |
+
"eval_rewards/accuracies": 0.363095223903656,
|
| 682 |
+
"eval_rewards/chosen": 0.43241986632347107,
|
| 683 |
+
"eval_rewards/margins": 0.5795699954032898,
|
| 684 |
+
"eval_rewards/rejected": -0.14715011417865753,
|
| 685 |
+
"eval_runtime": 113.5638,
|
| 686 |
+
"eval_samples_per_second": 17.611,
|
| 687 |
+
"eval_steps_per_second": 0.555,
|
| 688 |
+
"step": 400
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 1.4285714285714286,
|
| 692 |
+
"grad_norm": 26.444115538892063,
|
| 693 |
+
"learning_rate": 6.356589147286822e-08,
|
| 694 |
+
"logits/chosen": -2.534757614135742,
|
| 695 |
+
"logits/rejected": -2.5250518321990967,
|
| 696 |
+
"logps/chosen": -79.89534759521484,
|
| 697 |
+
"logps/rejected": -81.98127746582031,
|
| 698 |
+
"loss": 0.3911,
|
| 699 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 700 |
+
"rewards/chosen": 1.35897958278656,
|
| 701 |
+
"rewards/margins": 2.1614632606506348,
|
| 702 |
+
"rewards/rejected": -0.8024836778640747,
|
| 703 |
+
"step": 410
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.4634146341463414,
|
| 707 |
+
"grad_norm": 50.792504489879086,
|
| 708 |
+
"learning_rate": 5.968992248062015e-08,
|
| 709 |
+
"logits/chosen": -2.5849547386169434,
|
| 710 |
+
"logits/rejected": -2.5869174003601074,
|
| 711 |
+
"logps/chosen": -73.73815155029297,
|
| 712 |
+
"logps/rejected": -87.13761901855469,
|
| 713 |
+
"loss": 0.4085,
|
| 714 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 715 |
+
"rewards/chosen": 1.1775071620941162,
|
| 716 |
+
"rewards/margins": 2.3105506896972656,
|
| 717 |
+
"rewards/rejected": -1.1330437660217285,
|
| 718 |
+
"step": 420
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.4982578397212545,
|
| 722 |
+
"grad_norm": 62.104061513696834,
|
| 723 |
+
"learning_rate": 5.5813953488372087e-08,
|
| 724 |
+
"logits/chosen": -2.5392727851867676,
|
| 725 |
+
"logits/rejected": -2.528538465499878,
|
| 726 |
+
"logps/chosen": -66.37205505371094,
|
| 727 |
+
"logps/rejected": -74.73827362060547,
|
| 728 |
+
"loss": 0.4034,
|
| 729 |
+
"rewards/accuracies": 0.46875,
|
| 730 |
+
"rewards/chosen": 1.1451423168182373,
|
| 731 |
+
"rewards/margins": 2.2697441577911377,
|
| 732 |
+
"rewards/rejected": -1.12460196018219,
|
| 733 |
+
"step": 430
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.533101045296167,
|
| 737 |
+
"grad_norm": 99.60660196952468,
|
| 738 |
+
"learning_rate": 5.193798449612403e-08,
|
| 739 |
+
"logits/chosen": -2.525665283203125,
|
| 740 |
+
"logits/rejected": -2.5361111164093018,
|
| 741 |
+
"logps/chosen": -67.78907775878906,
|
| 742 |
+
"logps/rejected": -75.18824005126953,
|
| 743 |
+
"loss": 0.4097,
|
| 744 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 745 |
+
"rewards/chosen": 1.094794511795044,
|
| 746 |
+
"rewards/margins": 1.9390010833740234,
|
| 747 |
+
"rewards/rejected": -0.8442065119743347,
|
| 748 |
+
"step": 440
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.5679442508710801,
|
| 752 |
+
"grad_norm": 52.21837049460921,
|
| 753 |
+
"learning_rate": 4.806201550387597e-08,
|
| 754 |
+
"logits/chosen": -2.573503255844116,
|
| 755 |
+
"logits/rejected": -2.538872718811035,
|
| 756 |
+
"logps/chosen": -88.26780700683594,
|
| 757 |
+
"logps/rejected": -89.10135650634766,
|
| 758 |
+
"loss": 0.4059,
|
| 759 |
+
"rewards/accuracies": 0.518750011920929,
|
| 760 |
+
"rewards/chosen": 1.2870919704437256,
|
| 761 |
+
"rewards/margins": 2.856309413909912,
|
| 762 |
+
"rewards/rejected": -1.5692174434661865,
|
| 763 |
+
"step": 450
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.6027874564459932,
|
| 767 |
+
"grad_norm": 32.28937168577243,
|
| 768 |
+
"learning_rate": 4.418604651162791e-08,
|
| 769 |
+
"logits/chosen": -2.5357863903045654,
|
| 770 |
+
"logits/rejected": -2.5155162811279297,
|
| 771 |
+
"logps/chosen": -73.77734375,
|
| 772 |
+
"logps/rejected": -79.86143493652344,
|
| 773 |
+
"loss": 0.42,
|
| 774 |
+
"rewards/accuracies": 0.5,
|
| 775 |
+
"rewards/chosen": 1.2010835409164429,
|
| 776 |
+
"rewards/margins": 2.6294467449188232,
|
| 777 |
+
"rewards/rejected": -1.4283630847930908,
|
| 778 |
+
"step": 460
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.6376306620209058,
|
| 782 |
+
"grad_norm": 38.87268593820624,
|
| 783 |
+
"learning_rate": 4.031007751937984e-08,
|
| 784 |
+
"logits/chosen": -2.5288286209106445,
|
| 785 |
+
"logits/rejected": -2.5216517448425293,
|
| 786 |
+
"logps/chosen": -59.2172966003418,
|
| 787 |
+
"logps/rejected": -73.46331024169922,
|
| 788 |
+
"loss": 0.422,
|
| 789 |
+
"rewards/accuracies": 0.39375001192092896,
|
| 790 |
+
"rewards/chosen": 1.042550802230835,
|
| 791 |
+
"rewards/margins": 1.938773512840271,
|
| 792 |
+
"rewards/rejected": -0.8962229490280151,
|
| 793 |
+
"step": 470
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.6724738675958188,
|
| 797 |
+
"grad_norm": 86.29055693743794,
|
| 798 |
+
"learning_rate": 3.643410852713178e-08,
|
| 799 |
+
"logits/chosen": -2.5448553562164307,
|
| 800 |
+
"logits/rejected": -2.5282416343688965,
|
| 801 |
+
"logps/chosen": -52.4884147644043,
|
| 802 |
+
"logps/rejected": -53.635406494140625,
|
| 803 |
+
"loss": 0.4111,
|
| 804 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 805 |
+
"rewards/chosen": 0.7865492105484009,
|
| 806 |
+
"rewards/margins": 1.6136754751205444,
|
| 807 |
+
"rewards/rejected": -0.8271263837814331,
|
| 808 |
+
"step": 480
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 1.7073170731707317,
|
| 812 |
+
"grad_norm": 101.66995149882506,
|
| 813 |
+
"learning_rate": 3.2558139534883724e-08,
|
| 814 |
+
"logits/chosen": -2.5356173515319824,
|
| 815 |
+
"logits/rejected": -2.5221338272094727,
|
| 816 |
+
"logps/chosen": -71.38683319091797,
|
| 817 |
+
"logps/rejected": -72.329345703125,
|
| 818 |
+
"loss": 0.4539,
|
| 819 |
+
"rewards/accuracies": 0.39375001192092896,
|
| 820 |
+
"rewards/chosen": 0.9174805879592896,
|
| 821 |
+
"rewards/margins": 2.05107045173645,
|
| 822 |
+
"rewards/rejected": -1.1335899829864502,
|
| 823 |
+
"step": 490
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.7421602787456445,
|
| 827 |
+
"grad_norm": 62.46391948863256,
|
| 828 |
+
"learning_rate": 2.868217054263566e-08,
|
| 829 |
+
"logits/chosen": -2.4681618213653564,
|
| 830 |
+
"logits/rejected": -2.4625699520111084,
|
| 831 |
+
"logps/chosen": -73.69934844970703,
|
| 832 |
+
"logps/rejected": -86.02604675292969,
|
| 833 |
+
"loss": 0.3945,
|
| 834 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 835 |
+
"rewards/chosen": 1.2942867279052734,
|
| 836 |
+
"rewards/margins": 2.4456610679626465,
|
| 837 |
+
"rewards/rejected": -1.1513745784759521,
|
| 838 |
+
"step": 500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.7421602787456445,
|
| 842 |
+
"eval_logits/chosen": -2.5708680152893066,
|
| 843 |
+
"eval_logits/rejected": -2.5542969703674316,
|
| 844 |
+
"eval_logps/chosen": -73.87686920166016,
|
| 845 |
+
"eval_logps/rejected": -81.72207641601562,
|
| 846 |
+
"eval_loss": 0.6465447545051575,
|
| 847 |
+
"eval_rewards/accuracies": 0.3710317313671112,
|
| 848 |
+
"eval_rewards/chosen": 0.3071807324886322,
|
| 849 |
+
"eval_rewards/margins": 0.6040430665016174,
|
| 850 |
+
"eval_rewards/rejected": -0.29686233401298523,
|
| 851 |
+
"eval_runtime": 114.1539,
|
| 852 |
+
"eval_samples_per_second": 17.52,
|
| 853 |
+
"eval_steps_per_second": 0.552,
|
| 854 |
+
"step": 500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.7770034843205575,
|
| 858 |
+
"grad_norm": 45.13747528668081,
|
| 859 |
+
"learning_rate": 2.4806201550387595e-08,
|
| 860 |
+
"logits/chosen": -2.5347788333892822,
|
| 861 |
+
"logits/rejected": -2.5145461559295654,
|
| 862 |
+
"logps/chosen": -68.51091003417969,
|
| 863 |
+
"logps/rejected": -71.65690612792969,
|
| 864 |
+
"loss": 0.3946,
|
| 865 |
+
"rewards/accuracies": 0.46875,
|
| 866 |
+
"rewards/chosen": 1.1227554082870483,
|
| 867 |
+
"rewards/margins": 2.1591665744781494,
|
| 868 |
+
"rewards/rejected": -1.036410927772522,
|
| 869 |
+
"step": 510
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.8118466898954704,
|
| 873 |
+
"grad_norm": 69.84605729348092,
|
| 874 |
+
"learning_rate": 2.0930232558139537e-08,
|
| 875 |
+
"logits/chosen": -2.549290418624878,
|
| 876 |
+
"logits/rejected": -2.5436315536499023,
|
| 877 |
+
"logps/chosen": -72.00182342529297,
|
| 878 |
+
"logps/rejected": -79.60043334960938,
|
| 879 |
+
"loss": 0.4047,
|
| 880 |
+
"rewards/accuracies": 0.46875,
|
| 881 |
+
"rewards/chosen": 1.1346954107284546,
|
| 882 |
+
"rewards/margins": 2.3285088539123535,
|
| 883 |
+
"rewards/rejected": -1.1938132047653198,
|
| 884 |
+
"step": 520
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.8466898954703832,
|
| 888 |
+
"grad_norm": 47.800583904519996,
|
| 889 |
+
"learning_rate": 1.7054263565891472e-08,
|
| 890 |
+
"logits/chosen": -2.526092767715454,
|
| 891 |
+
"logits/rejected": -2.5203917026519775,
|
| 892 |
+
"logps/chosen": -68.9186782836914,
|
| 893 |
+
"logps/rejected": -73.74215698242188,
|
| 894 |
+
"loss": 0.4153,
|
| 895 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 896 |
+
"rewards/chosen": 1.208732008934021,
|
| 897 |
+
"rewards/margins": 2.3980166912078857,
|
| 898 |
+
"rewards/rejected": -1.1892845630645752,
|
| 899 |
+
"step": 530
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.8815331010452963,
|
| 903 |
+
"grad_norm": 503.5623137913587,
|
| 904 |
+
"learning_rate": 1.3178294573643412e-08,
|
| 905 |
+
"logits/chosen": -2.5318820476531982,
|
| 906 |
+
"logits/rejected": -2.5522282123565674,
|
| 907 |
+
"logps/chosen": -61.30474090576172,
|
| 908 |
+
"logps/rejected": -75.95423889160156,
|
| 909 |
+
"loss": 0.426,
|
| 910 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 911 |
+
"rewards/chosen": 1.213020920753479,
|
| 912 |
+
"rewards/margins": 2.5194857120513916,
|
| 913 |
+
"rewards/rejected": -1.3064649105072021,
|
| 914 |
+
"step": 540
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.916376306620209,
|
| 918 |
+
"grad_norm": 39.90012660671112,
|
| 919 |
+
"learning_rate": 9.302325581395349e-09,
|
| 920 |
+
"logits/chosen": -2.523984909057617,
|
| 921 |
+
"logits/rejected": -2.5038905143737793,
|
| 922 |
+
"logps/chosen": -87.86249542236328,
|
| 923 |
+
"logps/rejected": -89.35391235351562,
|
| 924 |
+
"loss": 0.4004,
|
| 925 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 926 |
+
"rewards/chosen": 1.4861438274383545,
|
| 927 |
+
"rewards/margins": 2.720036506652832,
|
| 928 |
+
"rewards/rejected": -1.233892560005188,
|
| 929 |
+
"step": 550
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 1.951219512195122,
|
| 933 |
+
"grad_norm": 18.05863362962586,
|
| 934 |
+
"learning_rate": 5.4263565891472866e-09,
|
| 935 |
+
"logits/chosen": -2.5144786834716797,
|
| 936 |
+
"logits/rejected": -2.5275046825408936,
|
| 937 |
+
"logps/chosen": -61.50837326049805,
|
| 938 |
+
"logps/rejected": -72.43628692626953,
|
| 939 |
+
"loss": 0.424,
|
| 940 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 941 |
+
"rewards/chosen": 1.105821132659912,
|
| 942 |
+
"rewards/margins": 2.175888776779175,
|
| 943 |
+
"rewards/rejected": -1.0700676441192627,
|
| 944 |
+
"step": 560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.986062717770035,
|
| 948 |
+
"grad_norm": 49.2882801078698,
|
| 949 |
+
"learning_rate": 1.5503875968992247e-09,
|
| 950 |
+
"logits/chosen": -2.589893102645874,
|
| 951 |
+
"logits/rejected": -2.5597004890441895,
|
| 952 |
+
"logps/chosen": -63.0964241027832,
|
| 953 |
+
"logps/rejected": -64.02296447753906,
|
| 954 |
+
"loss": 0.4131,
|
| 955 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 956 |
+
"rewards/chosen": 1.058401346206665,
|
| 957 |
+
"rewards/margins": 1.7885780334472656,
|
| 958 |
+
"rewards/rejected": -0.7301766872406006,
|
| 959 |
+
"step": 570
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 2.0,
|
| 963 |
+
"step": 574,
|
| 964 |
+
"total_flos": 0.0,
|
| 965 |
+
"train_loss": 0.5264799162901237,
|
| 966 |
+
"train_runtime": 6499.6728,
|
| 967 |
+
"train_samples_per_second": 5.643,
|
| 968 |
+
"train_steps_per_second": 0.088
|
| 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 |
+
}
|