Model save
Browse files- README.md +81 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +1475 -0
README.md
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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-7e-07-0.45-cosine-3.0
|
| 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-7e-07-0.45-cosine-3.0
|
| 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: 0.9344
|
| 22 |
+
- Rewards/chosen: -0.3373
|
| 23 |
+
- Rewards/rejected: -2.0235
|
| 24 |
+
- Rewards/accuracies: 0.3452
|
| 25 |
+
- Rewards/margins: 1.6862
|
| 26 |
+
- Logps/rejected: -85.6249
|
| 27 |
+
- Logps/chosen: -75.2407
|
| 28 |
+
- Logits/rejected: -2.6727
|
| 29 |
+
- Logits/chosen: -2.6886
|
| 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: 7e-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: cosine
|
| 59 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 60 |
+
- num_epochs: 3
|
| 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 |
+
| 0.6948 | 0.3484 | 100 | 0.6997 | 0.9816 | 0.4942 | 0.3452 | 0.4873 | -80.0301 | -72.3100 | -2.5413 | -2.5577 |
|
| 67 |
+
| 0.7373 | 0.6969 | 200 | 0.7720 | 1.2732 | 0.5117 | 0.3294 | 0.7615 | -79.9912 | -71.6619 | -2.5716 | -2.5870 |
|
| 68 |
+
| 0.4002 | 1.0453 | 300 | 0.8163 | 0.4524 | -0.4497 | 0.3472 | 0.9021 | -82.1276 | -73.4859 | -2.6256 | -2.6409 |
|
| 69 |
+
| 0.3982 | 1.3937 | 400 | 0.8872 | 1.2165 | 0.0680 | 0.3313 | 1.1485 | -80.9772 | -71.7879 | -2.7106 | -2.7265 |
|
| 70 |
+
| 0.389 | 1.7422 | 500 | 0.9107 | 0.3181 | -0.9594 | 0.3353 | 1.2775 | -83.2604 | -73.7844 | -2.7188 | -2.7346 |
|
| 71 |
+
| 0.3707 | 2.0906 | 600 | 0.8992 | 0.6908 | -0.7854 | 0.3472 | 1.4762 | -82.8736 | -72.9561 | -2.6904 | -2.7065 |
|
| 72 |
+
| 0.3672 | 2.4390 | 700 | 0.9354 | -0.5110 | -2.2396 | 0.3492 | 1.7285 | -86.1051 | -75.6269 | -2.6662 | -2.6823 |
|
| 73 |
+
| 0.3596 | 2.7875 | 800 | 0.9344 | -0.3373 | -2.0235 | 0.3452 | 1.6862 | -85.6249 | -75.2407 | -2.6727 | -2.6886 |
|
| 74 |
+
|
| 75 |
+
|
| 76 |
+
### Framework versions
|
| 77 |
+
|
| 78 |
+
- Transformers 4.44.2
|
| 79 |
+
- Pytorch 2.4.0
|
| 80 |
+
- Datasets 2.21.0
|
| 81 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.49774221859604084,
|
| 5 |
+
"train_runtime": 9884.8529,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.566,
|
| 8 |
+
"train_steps_per_second": 0.087
|
| 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": 3.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.49774221859604084,
|
| 5 |
+
"train_runtime": 9884.8529,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.566,
|
| 8 |
+
"train_steps_per_second": 0.087
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1475 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 861,
|
| 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": 260.1595222158553,
|
| 14 |
+
"learning_rate": 8.045977011494253e-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": 236.39725577198064,
|
| 29 |
+
"learning_rate": 8.045977011494252e-08,
|
| 30 |
+
"logits/chosen": -2.5637922286987305,
|
| 31 |
+
"logits/rejected": -2.5625600814819336,
|
| 32 |
+
"logps/chosen": -59.67917251586914,
|
| 33 |
+
"logps/rejected": -73.37747955322266,
|
| 34 |
+
"loss": 0.6959,
|
| 35 |
+
"rewards/accuracies": 0.1875,
|
| 36 |
+
"rewards/chosen": -0.008780181407928467,
|
| 37 |
+
"rewards/margins": -0.010324436239898205,
|
| 38 |
+
"rewards/rejected": 0.0015442547155544162,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.06968641114982578,
|
| 43 |
+
"grad_norm": 308.42103352579716,
|
| 44 |
+
"learning_rate": 1.6091954022988505e-07,
|
| 45 |
+
"logits/chosen": -2.6053006649017334,
|
| 46 |
+
"logits/rejected": -2.5642189979553223,
|
| 47 |
+
"logps/chosen": -104.08708190917969,
|
| 48 |
+
"logps/rejected": -94.90479278564453,
|
| 49 |
+
"loss": 0.6931,
|
| 50 |
+
"rewards/accuracies": 0.34375,
|
| 51 |
+
"rewards/chosen": 0.017988674342632294,
|
| 52 |
+
"rewards/margins": 0.022790271788835526,
|
| 53 |
+
"rewards/rejected": -0.004801597446203232,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.10452961672473868,
|
| 58 |
+
"grad_norm": 300.68038725792655,
|
| 59 |
+
"learning_rate": 2.413793103448276e-07,
|
| 60 |
+
"logits/chosen": -2.5924947261810303,
|
| 61 |
+
"logits/rejected": -2.5725739002227783,
|
| 62 |
+
"logps/chosen": -82.33592987060547,
|
| 63 |
+
"logps/rejected": -91.51930236816406,
|
| 64 |
+
"loss": 0.6759,
|
| 65 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 66 |
+
"rewards/chosen": 0.0722854882478714,
|
| 67 |
+
"rewards/margins": 0.06289331614971161,
|
| 68 |
+
"rewards/rejected": 0.009392165578901768,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13937282229965156,
|
| 73 |
+
"grad_norm": 238.79409703480542,
|
| 74 |
+
"learning_rate": 3.218390804597701e-07,
|
| 75 |
+
"logits/chosen": -2.4986560344696045,
|
| 76 |
+
"logits/rejected": -2.49660062789917,
|
| 77 |
+
"logps/chosen": -77.81121826171875,
|
| 78 |
+
"logps/rejected": -73.2593002319336,
|
| 79 |
+
"loss": 0.654,
|
| 80 |
+
"rewards/accuracies": 0.28125,
|
| 81 |
+
"rewards/chosen": 0.04823530465364456,
|
| 82 |
+
"rewards/margins": 0.16432270407676697,
|
| 83 |
+
"rewards/rejected": -0.1160874143242836,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.17421602787456447,
|
| 88 |
+
"grad_norm": 217.0577300579145,
|
| 89 |
+
"learning_rate": 4.022988505747126e-07,
|
| 90 |
+
"logits/chosen": -2.5254292488098145,
|
| 91 |
+
"logits/rejected": -2.52939772605896,
|
| 92 |
+
"logps/chosen": -63.20558547973633,
|
| 93 |
+
"logps/rejected": -75.68934631347656,
|
| 94 |
+
"loss": 0.6651,
|
| 95 |
+
"rewards/accuracies": 0.26249998807907104,
|
| 96 |
+
"rewards/chosen": 0.33010634779930115,
|
| 97 |
+
"rewards/margins": 0.14375844597816467,
|
| 98 |
+
"rewards/rejected": 0.18634793162345886,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.20905923344947736,
|
| 103 |
+
"grad_norm": 222.80981857126162,
|
| 104 |
+
"learning_rate": 4.827586206896552e-07,
|
| 105 |
+
"logits/chosen": -2.4844985008239746,
|
| 106 |
+
"logits/rejected": -2.478571653366089,
|
| 107 |
+
"logps/chosen": -70.86429595947266,
|
| 108 |
+
"logps/rejected": -66.64778137207031,
|
| 109 |
+
"loss": 0.6498,
|
| 110 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 111 |
+
"rewards/chosen": 1.0539968013763428,
|
| 112 |
+
"rewards/margins": 0.2616458535194397,
|
| 113 |
+
"rewards/rejected": 0.7923508882522583,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.24390243902439024,
|
| 118 |
+
"grad_norm": 243.92358905965492,
|
| 119 |
+
"learning_rate": 5.632183908045977e-07,
|
| 120 |
+
"logits/chosen": -2.4932572841644287,
|
| 121 |
+
"logits/rejected": -2.4880998134613037,
|
| 122 |
+
"logps/chosen": -61.057044982910156,
|
| 123 |
+
"logps/rejected": -66.05777740478516,
|
| 124 |
+
"loss": 0.6558,
|
| 125 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 126 |
+
"rewards/chosen": 1.4306201934814453,
|
| 127 |
+
"rewards/margins": 0.36464259028434753,
|
| 128 |
+
"rewards/rejected": 1.0659778118133545,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.2787456445993031,
|
| 133 |
+
"grad_norm": 250.0266197573928,
|
| 134 |
+
"learning_rate": 6.436781609195402e-07,
|
| 135 |
+
"logits/chosen": -2.4400105476379395,
|
| 136 |
+
"logits/rejected": -2.430251121520996,
|
| 137 |
+
"logps/chosen": -72.05781555175781,
|
| 138 |
+
"logps/rejected": -74.76571655273438,
|
| 139 |
+
"loss": 0.6758,
|
| 140 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 141 |
+
"rewards/chosen": 1.623175024986267,
|
| 142 |
+
"rewards/margins": 0.38602036237716675,
|
| 143 |
+
"rewards/rejected": 1.2371546030044556,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.313588850174216,
|
| 148 |
+
"grad_norm": 256.3370381835667,
|
| 149 |
+
"learning_rate": 6.999740526496426e-07,
|
| 150 |
+
"logits/chosen": -2.489266872406006,
|
| 151 |
+
"logits/rejected": -2.503830909729004,
|
| 152 |
+
"logps/chosen": -62.3740234375,
|
| 153 |
+
"logps/rejected": -67.14833068847656,
|
| 154 |
+
"loss": 0.7197,
|
| 155 |
+
"rewards/accuracies": 0.28125,
|
| 156 |
+
"rewards/chosen": 1.6237154006958008,
|
| 157 |
+
"rewards/margins": 0.28914302587509155,
|
| 158 |
+
"rewards/rejected": 1.334572196006775,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.34843205574912894,
|
| 163 |
+
"grad_norm": 260.27523546187444,
|
| 164 |
+
"learning_rate": 6.995128734390302e-07,
|
| 165 |
+
"logits/chosen": -2.478929042816162,
|
| 166 |
+
"logits/rejected": -2.4793598651885986,
|
| 167 |
+
"logps/chosen": -71.95938873291016,
|
| 168 |
+
"logps/rejected": -79.0130615234375,
|
| 169 |
+
"loss": 0.6948,
|
| 170 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 171 |
+
"rewards/chosen": 1.4903005361557007,
|
| 172 |
+
"rewards/margins": 0.5729378461837769,
|
| 173 |
+
"rewards/rejected": 0.9173626899719238,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.34843205574912894,
|
| 178 |
+
"eval_logits/chosen": -2.5576956272125244,
|
| 179 |
+
"eval_logits/rejected": -2.541304588317871,
|
| 180 |
+
"eval_logps/chosen": -72.30996704101562,
|
| 181 |
+
"eval_logps/rejected": -80.03009033203125,
|
| 182 |
+
"eval_loss": 0.6997446417808533,
|
| 183 |
+
"eval_rewards/accuracies": 0.3452380895614624,
|
| 184 |
+
"eval_rewards/chosen": 0.9815704226493835,
|
| 185 |
+
"eval_rewards/margins": 0.48734748363494873,
|
| 186 |
+
"eval_rewards/rejected": 0.4942229390144348,
|
| 187 |
+
"eval_runtime": 114.1155,
|
| 188 |
+
"eval_samples_per_second": 17.526,
|
| 189 |
+
"eval_steps_per_second": 0.552,
|
| 190 |
+
"step": 100
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.3832752613240418,
|
| 194 |
+
"grad_norm": 269.9879110557674,
|
| 195 |
+
"learning_rate": 6.984759608935431e-07,
|
| 196 |
+
"logits/chosen": -2.495579957962036,
|
| 197 |
+
"logits/rejected": -2.460557460784912,
|
| 198 |
+
"logps/chosen": -72.69947814941406,
|
| 199 |
+
"logps/rejected": -63.181190490722656,
|
| 200 |
+
"loss": 0.7231,
|
| 201 |
+
"rewards/accuracies": 0.26875001192092896,
|
| 202 |
+
"rewards/chosen": 0.21174518764019012,
|
| 203 |
+
"rewards/margins": 0.2885180115699768,
|
| 204 |
+
"rewards/rejected": -0.0767727941274643,
|
| 205 |
+
"step": 110
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.4181184668989547,
|
| 209 |
+
"grad_norm": 215.87968567956005,
|
| 210 |
+
"learning_rate": 6.96865023062192e-07,
|
| 211 |
+
"logits/chosen": -2.5388295650482178,
|
| 212 |
+
"logits/rejected": -2.5087077617645264,
|
| 213 |
+
"logps/chosen": -77.22286224365234,
|
| 214 |
+
"logps/rejected": -67.7669448852539,
|
| 215 |
+
"loss": 0.6929,
|
| 216 |
+
"rewards/accuracies": 0.3125,
|
| 217 |
+
"rewards/chosen": 0.10741810500621796,
|
| 218 |
+
"rewards/margins": 0.6003297567367554,
|
| 219 |
+
"rewards/rejected": -0.4929116666316986,
|
| 220 |
+
"step": 120
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4529616724738676,
|
| 224 |
+
"grad_norm": 364.597210128526,
|
| 225 |
+
"learning_rate": 6.946827135542728e-07,
|
| 226 |
+
"logits/chosen": -2.575723171234131,
|
| 227 |
+
"logits/rejected": -2.5577731132507324,
|
| 228 |
+
"logps/chosen": -83.23577117919922,
|
| 229 |
+
"logps/rejected": -88.67012023925781,
|
| 230 |
+
"loss": 0.7704,
|
| 231 |
+
"rewards/accuracies": 0.34375,
|
| 232 |
+
"rewards/chosen": 0.2902565896511078,
|
| 233 |
+
"rewards/margins": 0.8129722476005554,
|
| 234 |
+
"rewards/rejected": -0.52271568775177,
|
| 235 |
+
"step": 130
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.4878048780487805,
|
| 239 |
+
"grad_norm": 192.41269674377438,
|
| 240 |
+
"learning_rate": 6.919326271682209e-07,
|
| 241 |
+
"logits/chosen": -2.478255271911621,
|
| 242 |
+
"logits/rejected": -2.4691178798675537,
|
| 243 |
+
"logps/chosen": -78.7398681640625,
|
| 244 |
+
"logps/rejected": -70.11216735839844,
|
| 245 |
+
"loss": 0.6961,
|
| 246 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 247 |
+
"rewards/chosen": 1.082502007484436,
|
| 248 |
+
"rewards/margins": 0.7890673875808716,
|
| 249 |
+
"rewards/rejected": 0.29343467950820923,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.5226480836236934,
|
| 254 |
+
"grad_norm": 266.7719593983313,
|
| 255 |
+
"learning_rate": 6.886192939700987e-07,
|
| 256 |
+
"logits/chosen": -2.5632805824279785,
|
| 257 |
+
"logits/rejected": -2.5232510566711426,
|
| 258 |
+
"logps/chosen": -76.60926818847656,
|
| 259 |
+
"logps/rejected": -78.16521453857422,
|
| 260 |
+
"loss": 0.7487,
|
| 261 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 262 |
+
"rewards/chosen": 1.2937796115875244,
|
| 263 |
+
"rewards/margins": 0.6702733635902405,
|
| 264 |
+
"rewards/rejected": 0.6235060691833496,
|
| 265 |
+
"step": 150
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.5574912891986062,
|
| 269 |
+
"grad_norm": 236.73335221722851,
|
| 270 |
+
"learning_rate": 6.84748171831466e-07,
|
| 271 |
+
"logits/chosen": -2.5542993545532227,
|
| 272 |
+
"logits/rejected": -2.5730948448181152,
|
| 273 |
+
"logps/chosen": -61.49628829956055,
|
| 274 |
+
"logps/rejected": -69.87091064453125,
|
| 275 |
+
"loss": 0.7401,
|
| 276 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 277 |
+
"rewards/chosen": 1.3628534078598022,
|
| 278 |
+
"rewards/margins": 0.4169517159461975,
|
| 279 |
+
"rewards/rejected": 0.9459015727043152,
|
| 280 |
+
"step": 160
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.5923344947735192,
|
| 284 |
+
"grad_norm": 280.18013018164186,
|
| 285 |
+
"learning_rate": 6.803256374389282e-07,
|
| 286 |
+
"logits/chosen": -2.571448564529419,
|
| 287 |
+
"logits/rejected": -2.558177947998047,
|
| 288 |
+
"logps/chosen": -65.29753112792969,
|
| 289 |
+
"logps/rejected": -73.9312744140625,
|
| 290 |
+
"loss": 0.7499,
|
| 291 |
+
"rewards/accuracies": 0.28125,
|
| 292 |
+
"rewards/chosen": 1.65143620967865,
|
| 293 |
+
"rewards/margins": 0.579193651676178,
|
| 294 |
+
"rewards/rejected": 1.0722427368164062,
|
| 295 |
+
"step": 170
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.627177700348432,
|
| 299 |
+
"grad_norm": 257.9946756555396,
|
| 300 |
+
"learning_rate": 6.753589757901721e-07,
|
| 301 |
+
"logits/chosen": -2.601243495941162,
|
| 302 |
+
"logits/rejected": -2.5908031463623047,
|
| 303 |
+
"logps/chosen": -87.05482482910156,
|
| 304 |
+
"logps/rejected": -82.89573669433594,
|
| 305 |
+
"loss": 0.7906,
|
| 306 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 307 |
+
"rewards/chosen": 2.16105318069458,
|
| 308 |
+
"rewards/margins": 0.5344269871711731,
|
| 309 |
+
"rewards/rejected": 1.6266262531280518,
|
| 310 |
+
"step": 180
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.662020905923345,
|
| 314 |
+
"grad_norm": 200.6147691015439,
|
| 315 |
+
"learning_rate": 6.69856368193789e-07,
|
| 316 |
+
"logits/chosen": -2.6006789207458496,
|
| 317 |
+
"logits/rejected": -2.5925745964050293,
|
| 318 |
+
"logps/chosen": -67.43627166748047,
|
| 319 |
+
"logps/rejected": -78.66804504394531,
|
| 320 |
+
"loss": 0.7368,
|
| 321 |
+
"rewards/accuracies": 0.28125,
|
| 322 |
+
"rewards/chosen": 1.626844048500061,
|
| 323 |
+
"rewards/margins": 0.3397817015647888,
|
| 324 |
+
"rewards/rejected": 1.2870622873306274,
|
| 325 |
+
"step": 190
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.6968641114982579,
|
| 329 |
+
"grad_norm": 346.9455585144427,
|
| 330 |
+
"learning_rate": 6.63826878792655e-07,
|
| 331 |
+
"logits/chosen": -2.6272549629211426,
|
| 332 |
+
"logits/rejected": -2.633775234222412,
|
| 333 |
+
"logps/chosen": -86.80072021484375,
|
| 334 |
+
"logps/rejected": -90.06497192382812,
|
| 335 |
+
"loss": 0.7373,
|
| 336 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 337 |
+
"rewards/chosen": 1.7350423336029053,
|
| 338 |
+
"rewards/margins": 0.8414648175239563,
|
| 339 |
+
"rewards/rejected": 0.8935775756835938,
|
| 340 |
+
"step": 200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.6968641114982579,
|
| 344 |
+
"eval_logits/chosen": -2.5869662761688232,
|
| 345 |
+
"eval_logits/rejected": -2.571585178375244,
|
| 346 |
+
"eval_logps/chosen": -71.66189575195312,
|
| 347 |
+
"eval_logps/rejected": -79.99118041992188,
|
| 348 |
+
"eval_loss": 0.7720171213150024,
|
| 349 |
+
"eval_rewards/accuracies": 0.329365074634552,
|
| 350 |
+
"eval_rewards/chosen": 1.2732017040252686,
|
| 351 |
+
"eval_rewards/margins": 0.7614731192588806,
|
| 352 |
+
"eval_rewards/rejected": 0.5117285847663879,
|
| 353 |
+
"eval_runtime": 113.8896,
|
| 354 |
+
"eval_samples_per_second": 17.561,
|
| 355 |
+
"eval_steps_per_second": 0.553,
|
| 356 |
+
"step": 200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.7317073170731707,
|
| 360 |
+
"grad_norm": 371.1201689930978,
|
| 361 |
+
"learning_rate": 6.572804396330676e-07,
|
| 362 |
+
"logits/chosen": -2.590230941772461,
|
| 363 |
+
"logits/rejected": -2.5670275688171387,
|
| 364 |
+
"logps/chosen": -66.80162811279297,
|
| 365 |
+
"logps/rejected": -63.04413604736328,
|
| 366 |
+
"loss": 0.7579,
|
| 367 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 368 |
+
"rewards/chosen": 1.132849931716919,
|
| 369 |
+
"rewards/margins": 0.9707193374633789,
|
| 370 |
+
"rewards/rejected": 0.16213065385818481,
|
| 371 |
+
"step": 210
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.7665505226480837,
|
| 375 |
+
"grad_norm": 233.2613327041566,
|
| 376 |
+
"learning_rate": 6.502278343042315e-07,
|
| 377 |
+
"logits/chosen": -2.6396641731262207,
|
| 378 |
+
"logits/rejected": -2.6218152046203613,
|
| 379 |
+
"logps/chosen": -70.56974029541016,
|
| 380 |
+
"logps/rejected": -69.75597381591797,
|
| 381 |
+
"loss": 0.7735,
|
| 382 |
+
"rewards/accuracies": 0.24375000596046448,
|
| 383 |
+
"rewards/chosen": 1.2997958660125732,
|
| 384 |
+
"rewards/margins": 0.4415515959262848,
|
| 385 |
+
"rewards/rejected": 0.8582441210746765,
|
| 386 |
+
"step": 220
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.8013937282229965,
|
| 390 |
+
"grad_norm": 326.1156990352116,
|
| 391 |
+
"learning_rate": 6.42680680175045e-07,
|
| 392 |
+
"logits/chosen": -2.6760098934173584,
|
| 393 |
+
"logits/rejected": -2.65986967086792,
|
| 394 |
+
"logps/chosen": -86.16526794433594,
|
| 395 |
+
"logps/rejected": -87.11499786376953,
|
| 396 |
+
"loss": 0.8168,
|
| 397 |
+
"rewards/accuracies": 0.3687500059604645,
|
| 398 |
+
"rewards/chosen": 1.7259118556976318,
|
| 399 |
+
"rewards/margins": 1.2606786489486694,
|
| 400 |
+
"rewards/rejected": 0.4652332663536072,
|
| 401 |
+
"step": 230
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.8362369337979094,
|
| 405 |
+
"grad_norm": 267.49129367815516,
|
| 406 |
+
"learning_rate": 6.346514092574479e-07,
|
| 407 |
+
"logits/chosen": -2.6791954040527344,
|
| 408 |
+
"logits/rejected": -2.6473562717437744,
|
| 409 |
+
"logps/chosen": -81.90107727050781,
|
| 410 |
+
"logps/rejected": -76.54811096191406,
|
| 411 |
+
"loss": 0.8387,
|
| 412 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 413 |
+
"rewards/chosen": 2.1782379150390625,
|
| 414 |
+
"rewards/margins": 0.5636407136917114,
|
| 415 |
+
"rewards/rejected": 1.6145970821380615,
|
| 416 |
+
"step": 240
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.8710801393728222,
|
| 420 |
+
"grad_norm": 359.51061878716314,
|
| 421 |
+
"learning_rate": 6.26153247727851e-07,
|
| 422 |
+
"logits/chosen": -2.676499843597412,
|
| 423 |
+
"logits/rejected": -2.6430983543395996,
|
| 424 |
+
"logps/chosen": -90.66349792480469,
|
| 425 |
+
"logps/rejected": -86.98454284667969,
|
| 426 |
+
"loss": 0.6777,
|
| 427 |
+
"rewards/accuracies": 0.39375001192092896,
|
| 428 |
+
"rewards/chosen": 2.1536006927490234,
|
| 429 |
+
"rewards/margins": 0.8208580017089844,
|
| 430 |
+
"rewards/rejected": 1.3327425718307495,
|
| 431 |
+
"step": 250
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.9059233449477352,
|
| 435 |
+
"grad_norm": 219.1541113628816,
|
| 436 |
+
"learning_rate": 6.172001941403836e-07,
|
| 437 |
+
"logits/chosen": -2.585671901702881,
|
| 438 |
+
"logits/rejected": -2.599630832672119,
|
| 439 |
+
"logps/chosen": -56.668235778808594,
|
| 440 |
+
"logps/rejected": -64.54667663574219,
|
| 441 |
+
"loss": 0.8203,
|
| 442 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 443 |
+
"rewards/chosen": 1.0715070962905884,
|
| 444 |
+
"rewards/margins": 0.584784209728241,
|
| 445 |
+
"rewards/rejected": 0.4867229461669922,
|
| 446 |
+
"step": 260
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.9407665505226481,
|
| 450 |
+
"grad_norm": 447.78049155610177,
|
| 451 |
+
"learning_rate": 6.078069963678453e-07,
|
| 452 |
+
"logits/chosen": -2.667015314102173,
|
| 453 |
+
"logits/rejected": -2.6670918464660645,
|
| 454 |
+
"logps/chosen": -67.22764587402344,
|
| 455 |
+
"logps/rejected": -82.75209045410156,
|
| 456 |
+
"loss": 0.7889,
|
| 457 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 458 |
+
"rewards/chosen": 0.7436366081237793,
|
| 459 |
+
"rewards/margins": 0.842780590057373,
|
| 460 |
+
"rewards/rejected": -0.09914406388998032,
|
| 461 |
+
"step": 270
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.975609756097561,
|
| 465 |
+
"grad_norm": 255.49301673895127,
|
| 466 |
+
"learning_rate": 5.979891273083455e-07,
|
| 467 |
+
"logits/chosen": -2.560213804244995,
|
| 468 |
+
"logits/rejected": -2.5412333011627197,
|
| 469 |
+
"logps/chosen": -65.57135009765625,
|
| 470 |
+
"logps/rejected": -70.57018280029297,
|
| 471 |
+
"loss": 0.6995,
|
| 472 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 473 |
+
"rewards/chosen": 1.0183632373809814,
|
| 474 |
+
"rewards/margins": 0.8994625806808472,
|
| 475 |
+
"rewards/rejected": 0.11890069395303726,
|
| 476 |
+
"step": 280
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.0104529616724738,
|
| 480 |
+
"grad_norm": 62.440035652192904,
|
| 481 |
+
"learning_rate": 5.8776275939765e-07,
|
| 482 |
+
"logits/chosen": -2.583773612976074,
|
| 483 |
+
"logits/rejected": -2.558880567550659,
|
| 484 |
+
"logps/chosen": -67.36951446533203,
|
| 485 |
+
"logps/rejected": -66.36976623535156,
|
| 486 |
+
"loss": 0.6156,
|
| 487 |
+
"rewards/accuracies": 0.375,
|
| 488 |
+
"rewards/chosen": 2.1141774654388428,
|
| 489 |
+
"rewards/margins": 2.7693214416503906,
|
| 490 |
+
"rewards/rejected": -0.6551438570022583,
|
| 491 |
+
"step": 290
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.0452961672473868,
|
| 495 |
+
"grad_norm": 22.68193419500007,
|
| 496 |
+
"learning_rate": 5.771447379692167e-07,
|
| 497 |
+
"logits/chosen": -2.6168670654296875,
|
| 498 |
+
"logits/rejected": -2.603750228881836,
|
| 499 |
+
"logps/chosen": -58.1750373840332,
|
| 500 |
+
"logps/rejected": -80.46726989746094,
|
| 501 |
+
"loss": 0.4002,
|
| 502 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 503 |
+
"rewards/chosen": 3.7682430744171143,
|
| 504 |
+
"rewards/margins": 8.851958274841309,
|
| 505 |
+
"rewards/rejected": -5.083714962005615,
|
| 506 |
+
"step": 300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.0452961672473868,
|
| 510 |
+
"eval_logits/chosen": -2.6408894062042236,
|
| 511 |
+
"eval_logits/rejected": -2.6256282329559326,
|
| 512 |
+
"eval_logps/chosen": -73.48590087890625,
|
| 513 |
+
"eval_logps/rejected": -82.12760162353516,
|
| 514 |
+
"eval_loss": 0.8162885904312134,
|
| 515 |
+
"eval_rewards/accuracies": 0.3472222089767456,
|
| 516 |
+
"eval_rewards/chosen": 0.45239681005477905,
|
| 517 |
+
"eval_rewards/margins": 0.9020630717277527,
|
| 518 |
+
"eval_rewards/rejected": -0.44966623187065125,
|
| 519 |
+
"eval_runtime": 113.8253,
|
| 520 |
+
"eval_samples_per_second": 17.571,
|
| 521 |
+
"eval_steps_per_second": 0.553,
|
| 522 |
+
"step": 300
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.0801393728222997,
|
| 526 |
+
"grad_norm": 5.506688225969992,
|
| 527 |
+
"learning_rate": 5.66152553505804e-07,
|
| 528 |
+
"logits/chosen": -2.588723659515381,
|
| 529 |
+
"logits/rejected": -2.5905752182006836,
|
| 530 |
+
"logps/chosen": -58.51898193359375,
|
| 531 |
+
"logps/rejected": -90.38175964355469,
|
| 532 |
+
"loss": 0.4013,
|
| 533 |
+
"rewards/accuracies": 0.4375,
|
| 534 |
+
"rewards/chosen": 4.1029510498046875,
|
| 535 |
+
"rewards/margins": 10.913629531860352,
|
| 536 |
+
"rewards/rejected": -6.810678958892822,
|
| 537 |
+
"step": 310
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.1149825783972125,
|
| 541 |
+
"grad_norm": 84.44098198698735,
|
| 542 |
+
"learning_rate": 5.548043128283609e-07,
|
| 543 |
+
"logits/chosen": -2.6132149696350098,
|
| 544 |
+
"logits/rejected": -2.600966691970825,
|
| 545 |
+
"logps/chosen": -64.97319030761719,
|
| 546 |
+
"logps/rejected": -90.65545654296875,
|
| 547 |
+
"loss": 0.3888,
|
| 548 |
+
"rewards/accuracies": 0.46875,
|
| 549 |
+
"rewards/chosen": 4.044018745422363,
|
| 550 |
+
"rewards/margins": 9.642068862915039,
|
| 551 |
+
"rewards/rejected": -5.598048686981201,
|
| 552 |
+
"step": 320
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.1498257839721253,
|
| 556 |
+
"grad_norm": 54.66825088919817,
|
| 557 |
+
"learning_rate": 5.43118709269656e-07,
|
| 558 |
+
"logits/chosen": -2.5924811363220215,
|
| 559 |
+
"logits/rejected": -2.5696775913238525,
|
| 560 |
+
"logps/chosen": -71.87019348144531,
|
| 561 |
+
"logps/rejected": -88.80474853515625,
|
| 562 |
+
"loss": 0.3669,
|
| 563 |
+
"rewards/accuracies": 0.53125,
|
| 564 |
+
"rewards/chosen": 5.513075828552246,
|
| 565 |
+
"rewards/margins": 10.093175888061523,
|
| 566 |
+
"rewards/rejected": -4.5801005363464355,
|
| 567 |
+
"step": 330
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.1846689895470384,
|
| 571 |
+
"grad_norm": 16.225429361874546,
|
| 572 |
+
"learning_rate": 5.311149918817793e-07,
|
| 573 |
+
"logits/chosen": -2.58941650390625,
|
| 574 |
+
"logits/rejected": -2.597695827484131,
|
| 575 |
+
"logps/chosen": -71.19819641113281,
|
| 576 |
+
"logps/rejected": -106.7779769897461,
|
| 577 |
+
"loss": 0.3766,
|
| 578 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 579 |
+
"rewards/chosen": 5.647311687469482,
|
| 580 |
+
"rewards/margins": 10.82011890411377,
|
| 581 |
+
"rewards/rejected": -5.172807216644287,
|
| 582 |
+
"step": 340
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.2195121951219512,
|
| 586 |
+
"grad_norm": 78.05238402288815,
|
| 587 |
+
"learning_rate": 5.188129337282367e-07,
|
| 588 |
+
"logits/chosen": -2.643777370452881,
|
| 589 |
+
"logits/rejected": -2.613312005996704,
|
| 590 |
+
"logps/chosen": -56.92789840698242,
|
| 591 |
+
"logps/rejected": -74.79656982421875,
|
| 592 |
+
"loss": 0.3934,
|
| 593 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 594 |
+
"rewards/chosen": 5.251105308532715,
|
| 595 |
+
"rewards/margins": 9.353957176208496,
|
| 596 |
+
"rewards/rejected": -4.102850914001465,
|
| 597 |
+
"step": 350
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.254355400696864,
|
| 601 |
+
"grad_norm": 58.627931535437625,
|
| 602 |
+
"learning_rate": 5.062327993128697e-07,
|
| 603 |
+
"logits/chosen": -2.6389856338500977,
|
| 604 |
+
"logits/rejected": -2.612027406692505,
|
| 605 |
+
"logps/chosen": -61.96246337890625,
|
| 606 |
+
"logps/rejected": -73.02973175048828,
|
| 607 |
+
"loss": 0.3632,
|
| 608 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 609 |
+
"rewards/chosen": 4.449435710906982,
|
| 610 |
+
"rewards/margins": 7.763936519622803,
|
| 611 |
+
"rewards/rejected": -3.3145008087158203,
|
| 612 |
+
"step": 360
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.289198606271777,
|
| 616 |
+
"grad_norm": 103.01146780644565,
|
| 617 |
+
"learning_rate": 4.933953111992535e-07,
|
| 618 |
+
"logits/chosen": -2.6231448650360107,
|
| 619 |
+
"logits/rejected": -2.6383228302001953,
|
| 620 |
+
"logps/chosen": -60.911376953125,
|
| 621 |
+
"logps/rejected": -85.25511932373047,
|
| 622 |
+
"loss": 0.3915,
|
| 623 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 624 |
+
"rewards/chosen": 4.569601535797119,
|
| 625 |
+
"rewards/margins": 9.484132766723633,
|
| 626 |
+
"rewards/rejected": -4.9145307540893555,
|
| 627 |
+
"step": 370
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 1.32404181184669,
|
| 631 |
+
"grad_norm": 98.7337260301499,
|
| 632 |
+
"learning_rate": 4.803216158755572e-07,
|
| 633 |
+
"logits/chosen": -2.643336772918701,
|
| 634 |
+
"logits/rejected": -2.638636589050293,
|
| 635 |
+
"logps/chosen": -74.83843994140625,
|
| 636 |
+
"logps/rejected": -102.59355163574219,
|
| 637 |
+
"loss": 0.3849,
|
| 638 |
+
"rewards/accuracies": 0.543749988079071,
|
| 639 |
+
"rewards/chosen": 5.416236877441406,
|
| 640 |
+
"rewards/margins": 12.673513412475586,
|
| 641 |
+
"rewards/rejected": -7.2572784423828125,
|
| 642 |
+
"step": 380
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 1.3588850174216027,
|
| 646 |
+
"grad_norm": 88.65160189656861,
|
| 647 |
+
"learning_rate": 4.6703324892109645e-07,
|
| 648 |
+
"logits/chosen": -2.7196712493896484,
|
| 649 |
+
"logits/rejected": -2.7034525871276855,
|
| 650 |
+
"logps/chosen": -58.760398864746094,
|
| 651 |
+
"logps/rejected": -83.69285583496094,
|
| 652 |
+
"loss": 0.3793,
|
| 653 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 654 |
+
"rewards/chosen": 3.705737352371216,
|
| 655 |
+
"rewards/margins": 9.30022144317627,
|
| 656 |
+
"rewards/rejected": -5.594483375549316,
|
| 657 |
+
"step": 390
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.3937282229965158,
|
| 661 |
+
"grad_norm": 24.91794224364189,
|
| 662 |
+
"learning_rate": 4.535520995319585e-07,
|
| 663 |
+
"logits/chosen": -2.7114458084106445,
|
| 664 |
+
"logits/rejected": -2.6843841075897217,
|
| 665 |
+
"logps/chosen": -75.79566955566406,
|
| 666 |
+
"logps/rejected": -109.73543548583984,
|
| 667 |
+
"loss": 0.3982,
|
| 668 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 669 |
+
"rewards/chosen": 4.6581196784973145,
|
| 670 |
+
"rewards/margins": 9.757375717163086,
|
| 671 |
+
"rewards/rejected": -5.099255084991455,
|
| 672 |
+
"step": 400
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.3937282229965158,
|
| 676 |
+
"eval_logits/chosen": -2.726470470428467,
|
| 677 |
+
"eval_logits/rejected": -2.7106189727783203,
|
| 678 |
+
"eval_logps/chosen": -71.78788757324219,
|
| 679 |
+
"eval_logps/rejected": -80.97720336914062,
|
| 680 |
+
"eval_loss": 0.887217104434967,
|
| 681 |
+
"eval_rewards/accuracies": 0.3313491940498352,
|
| 682 |
+
"eval_rewards/chosen": 1.2165056467056274,
|
| 683 |
+
"eval_rewards/margins": 1.1484907865524292,
|
| 684 |
+
"eval_rewards/rejected": 0.06801486760377884,
|
| 685 |
+
"eval_runtime": 113.9417,
|
| 686 |
+
"eval_samples_per_second": 17.553,
|
| 687 |
+
"eval_steps_per_second": 0.553,
|
| 688 |
+
"step": 400
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 1.4285714285714286,
|
| 692 |
+
"grad_norm": 2.035285199174074,
|
| 693 |
+
"learning_rate": 4.3990037446413313e-07,
|
| 694 |
+
"logits/chosen": -2.70420503616333,
|
| 695 |
+
"logits/rejected": -2.6959073543548584,
|
| 696 |
+
"logps/chosen": -72.02146911621094,
|
| 697 |
+
"logps/rejected": -90.1993637084961,
|
| 698 |
+
"loss": 0.5091,
|
| 699 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 700 |
+
"rewards/chosen": 4.766329288482666,
|
| 701 |
+
"rewards/margins": 9.186702728271484,
|
| 702 |
+
"rewards/rejected": -4.420374393463135,
|
| 703 |
+
"step": 410
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.4634146341463414,
|
| 707 |
+
"grad_norm": 106.25796220752933,
|
| 708 |
+
"learning_rate": 4.2610056145354496e-07,
|
| 709 |
+
"logits/chosen": -2.7422072887420654,
|
| 710 |
+
"logits/rejected": -2.740999937057495,
|
| 711 |
+
"logps/chosen": -67.51850891113281,
|
| 712 |
+
"logps/rejected": -96.81071472167969,
|
| 713 |
+
"loss": 0.4118,
|
| 714 |
+
"rewards/accuracies": 0.4375,
|
| 715 |
+
"rewards/chosen": 3.858595371246338,
|
| 716 |
+
"rewards/margins": 9.231226921081543,
|
| 717 |
+
"rewards/rejected": -5.372632026672363,
|
| 718 |
+
"step": 420
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.4982578397212545,
|
| 722 |
+
"grad_norm": 7.263239275981174,
|
| 723 |
+
"learning_rate": 4.1217539217324226e-07,
|
| 724 |
+
"logits/chosen": -2.714473247528076,
|
| 725 |
+
"logits/rejected": -2.699317455291748,
|
| 726 |
+
"logps/chosen": -59.63612747192383,
|
| 727 |
+
"logps/rejected": -85.27086639404297,
|
| 728 |
+
"loss": 0.4066,
|
| 729 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 730 |
+
"rewards/chosen": 4.061794281005859,
|
| 731 |
+
"rewards/margins": 9.813599586486816,
|
| 732 |
+
"rewards/rejected": -5.751805782318115,
|
| 733 |
+
"step": 430
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.533101045296167,
|
| 737 |
+
"grad_norm": 132.98971170304543,
|
| 738 |
+
"learning_rate": 3.9814780478876267e-07,
|
| 739 |
+
"logits/chosen": -2.725816249847412,
|
| 740 |
+
"logits/rejected": -2.73162841796875,
|
| 741 |
+
"logps/chosen": -59.733802795410156,
|
| 742 |
+
"logps/rejected": -84.18232727050781,
|
| 743 |
+
"loss": 0.4048,
|
| 744 |
+
"rewards/accuracies": 0.41874998807907104,
|
| 745 |
+
"rewards/chosen": 4.610183238983154,
|
| 746 |
+
"rewards/margins": 9.417313575744629,
|
| 747 |
+
"rewards/rejected": -4.807129859924316,
|
| 748 |
+
"step": 440
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.5679442508710801,
|
| 752 |
+
"grad_norm": 28.002507272860324,
|
| 753 |
+
"learning_rate": 3.8404090617335413e-07,
|
| 754 |
+
"logits/chosen": -2.7826929092407227,
|
| 755 |
+
"logits/rejected": -2.749206304550171,
|
| 756 |
+
"logps/chosen": -80.09606170654297,
|
| 757 |
+
"logps/rejected": -101.27384948730469,
|
| 758 |
+
"loss": 0.4032,
|
| 759 |
+
"rewards/accuracies": 0.512499988079071,
|
| 760 |
+
"rewards/chosen": 4.835663795471191,
|
| 761 |
+
"rewards/margins": 11.725584030151367,
|
| 762 |
+
"rewards/rejected": -6.88992166519165,
|
| 763 |
+
"step": 450
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.6027874564459932,
|
| 767 |
+
"grad_norm": 37.52261766419059,
|
| 768 |
+
"learning_rate": 3.698779338452938e-07,
|
| 769 |
+
"logits/chosen": -2.769883632659912,
|
| 770 |
+
"logits/rejected": -2.7476718425750732,
|
| 771 |
+
"logps/chosen": -66.14527130126953,
|
| 772 |
+
"logps/rejected": -93.9597396850586,
|
| 773 |
+
"loss": 0.4069,
|
| 774 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 775 |
+
"rewards/chosen": 4.515407085418701,
|
| 776 |
+
"rewards/margins": 12.145170211791992,
|
| 777 |
+
"rewards/rejected": -7.629762172698975,
|
| 778 |
+
"step": 460
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.6376306620209058,
|
| 782 |
+
"grad_norm": 6.272455332024739,
|
| 783 |
+
"learning_rate": 3.556822176900017e-07,
|
| 784 |
+
"logits/chosen": -2.739861011505127,
|
| 785 |
+
"logits/rejected": -2.731259822845459,
|
| 786 |
+
"logps/chosen": -54.19404983520508,
|
| 787 |
+
"logps/rejected": -83.7978286743164,
|
| 788 |
+
"loss": 0.4057,
|
| 789 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 790 |
+
"rewards/chosen": 3.198758840560913,
|
| 791 |
+
"rewards/margins": 8.655891418457031,
|
| 792 |
+
"rewards/rejected": -5.4571332931518555,
|
| 793 |
+
"step": 470
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.6724738675958188,
|
| 797 |
+
"grad_norm": 46.64342674711966,
|
| 798 |
+
"learning_rate": 3.414771415300036e-07,
|
| 799 |
+
"logits/chosen": -2.737431049346924,
|
| 800 |
+
"logits/rejected": -2.7219510078430176,
|
| 801 |
+
"logps/chosen": -47.62141799926758,
|
| 802 |
+
"logps/rejected": -61.3114128112793,
|
| 803 |
+
"loss": 0.4139,
|
| 804 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 805 |
+
"rewards/chosen": 2.8980441093444824,
|
| 806 |
+
"rewards/margins": 7.096662998199463,
|
| 807 |
+
"rewards/rejected": -4.198617935180664,
|
| 808 |
+
"step": 480
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 1.7073170731707317,
|
| 812 |
+
"grad_norm": 28.98635525863854,
|
| 813 |
+
"learning_rate": 3.2728610460604674e-07,
|
| 814 |
+
"logits/chosen": -2.712428569793701,
|
| 815 |
+
"logits/rejected": -2.704003095626831,
|
| 816 |
+
"logps/chosen": -66.01129913330078,
|
| 817 |
+
"logps/rejected": -81.80970764160156,
|
| 818 |
+
"loss": 0.4894,
|
| 819 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 820 |
+
"rewards/chosen": 3.2447237968444824,
|
| 821 |
+
"rewards/margins": 8.53111743927002,
|
| 822 |
+
"rewards/rejected": -5.286392688751221,
|
| 823 |
+
"step": 490
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.7421602787456445,
|
| 827 |
+
"grad_norm": 80.97214190913306,
|
| 828 |
+
"learning_rate": 3.131324830328163e-07,
|
| 829 |
+
"logits/chosen": -2.6181178092956543,
|
| 830 |
+
"logits/rejected": -2.6126463413238525,
|
| 831 |
+
"logps/chosen": -67.113037109375,
|
| 832 |
+
"logps/rejected": -96.601318359375,
|
| 833 |
+
"loss": 0.389,
|
| 834 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 835 |
+
"rewards/chosen": 4.128693580627441,
|
| 836 |
+
"rewards/margins": 9.92380142211914,
|
| 837 |
+
"rewards/rejected": -5.795108795166016,
|
| 838 |
+
"step": 500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.7421602787456445,
|
| 842 |
+
"eval_logits/chosen": -2.7346041202545166,
|
| 843 |
+
"eval_logits/rejected": -2.7188162803649902,
|
| 844 |
+
"eval_logps/chosen": -73.7844009399414,
|
| 845 |
+
"eval_logps/rejected": -83.26041412353516,
|
| 846 |
+
"eval_loss": 0.910707414150238,
|
| 847 |
+
"eval_rewards/accuracies": 0.335317462682724,
|
| 848 |
+
"eval_rewards/chosen": 0.318076491355896,
|
| 849 |
+
"eval_rewards/margins": 1.277505874633789,
|
| 850 |
+
"eval_rewards/rejected": -0.9594294428825378,
|
| 851 |
+
"eval_runtime": 113.9969,
|
| 852 |
+
"eval_samples_per_second": 17.544,
|
| 853 |
+
"eval_steps_per_second": 0.553,
|
| 854 |
+
"step": 500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.7770034843205575,
|
| 858 |
+
"grad_norm": 51.127719184203606,
|
| 859 |
+
"learning_rate": 2.9903959129274836e-07,
|
| 860 |
+
"logits/chosen": -2.6881165504455566,
|
| 861 |
+
"logits/rejected": -2.669621467590332,
|
| 862 |
+
"logps/chosen": -61.860618591308594,
|
| 863 |
+
"logps/rejected": -81.27119445800781,
|
| 864 |
+
"loss": 0.3864,
|
| 865 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 866 |
+
"rewards/chosen": 4.003110885620117,
|
| 867 |
+
"rewards/margins": 9.262310028076172,
|
| 868 |
+
"rewards/rejected": -5.259199619293213,
|
| 869 |
+
"step": 510
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.8118466898954704,
|
| 873 |
+
"grad_norm": 41.46570799336418,
|
| 874 |
+
"learning_rate": 2.850306438313643e-07,
|
| 875 |
+
"logits/chosen": -2.680405855178833,
|
| 876 |
+
"logits/rejected": -2.675224781036377,
|
| 877 |
+
"logps/chosen": -64.68285369873047,
|
| 878 |
+
"logps/rejected": -88.90663146972656,
|
| 879 |
+
"loss": 0.4,
|
| 880 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 881 |
+
"rewards/chosen": 4.314764499664307,
|
| 882 |
+
"rewards/margins": 9.576990127563477,
|
| 883 |
+
"rewards/rejected": -5.262225151062012,
|
| 884 |
+
"step": 520
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.8466898954703832,
|
| 888 |
+
"grad_norm": 33.602492864431376,
|
| 889 |
+
"learning_rate": 2.711287168173922e-07,
|
| 890 |
+
"logits/chosen": -2.647599697113037,
|
| 891 |
+
"logits/rejected": -2.642518997192383,
|
| 892 |
+
"logps/chosen": -61.43975830078125,
|
| 893 |
+
"logps/rejected": -83.95713806152344,
|
| 894 |
+
"loss": 0.4019,
|
| 895 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 896 |
+
"rewards/chosen": 4.453371047973633,
|
| 897 |
+
"rewards/margins": 10.12047290802002,
|
| 898 |
+
"rewards/rejected": -5.667101860046387,
|
| 899 |
+
"step": 530
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.8815331010452963,
|
| 903 |
+
"grad_norm": 24.453834226726745,
|
| 904 |
+
"learning_rate": 2.573567101306622e-07,
|
| 905 |
+
"logits/chosen": -2.65397310256958,
|
| 906 |
+
"logits/rejected": -2.66703462600708,
|
| 907 |
+
"logps/chosen": -52.19524002075195,
|
| 908 |
+
"logps/rejected": -87.27598571777344,
|
| 909 |
+
"loss": 0.4031,
|
| 910 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 911 |
+
"rewards/chosen": 5.190995693206787,
|
| 912 |
+
"rewards/margins": 11.461599349975586,
|
| 913 |
+
"rewards/rejected": -6.270604133605957,
|
| 914 |
+
"step": 540
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.916376306620209,
|
| 918 |
+
"grad_norm": 2.684221407433218,
|
| 919 |
+
"learning_rate": 2.4373730964039504e-07,
|
| 920 |
+
"logits/chosen": -2.627750873565674,
|
| 921 |
+
"logits/rejected": -2.610384464263916,
|
| 922 |
+
"logps/chosen": -77.6567611694336,
|
| 923 |
+
"logps/rejected": -100.83102416992188,
|
| 924 |
+
"loss": 0.3791,
|
| 925 |
+
"rewards/accuracies": 0.5,
|
| 926 |
+
"rewards/chosen": 5.930109977722168,
|
| 927 |
+
"rewards/margins": 12.205314636230469,
|
| 928 |
+
"rewards/rejected": -6.275204658508301,
|
| 929 |
+
"step": 550
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 1.951219512195122,
|
| 933 |
+
"grad_norm": 79.1320535853655,
|
| 934 |
+
"learning_rate": 2.3029294983601597e-07,
|
| 935 |
+
"logits/chosen": -2.643815517425537,
|
| 936 |
+
"logits/rejected": -2.647261142730713,
|
| 937 |
+
"logps/chosen": -54.161102294921875,
|
| 938 |
+
"logps/rejected": -79.52574157714844,
|
| 939 |
+
"loss": 0.419,
|
| 940 |
+
"rewards/accuracies": 0.4375,
|
| 941 |
+
"rewards/chosen": 4.301507949829102,
|
| 942 |
+
"rewards/margins": 8.454828262329102,
|
| 943 |
+
"rewards/rejected": -4.153320789337158,
|
| 944 |
+
"step": 560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.986062717770035,
|
| 948 |
+
"grad_norm": 29.398825057181867,
|
| 949 |
+
"learning_rate": 2.1704577687205507e-07,
|
| 950 |
+
"logits/chosen": -2.730459690093994,
|
| 951 |
+
"logits/rejected": -2.7015600204467773,
|
| 952 |
+
"logps/chosen": -56.092933654785156,
|
| 953 |
+
"logps/rejected": -70.2969970703125,
|
| 954 |
+
"loss": 0.4011,
|
| 955 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 956 |
+
"rewards/chosen": 4.104133605957031,
|
| 957 |
+
"rewards/margins": 7.584607124328613,
|
| 958 |
+
"rewards/rejected": -3.4804725646972656,
|
| 959 |
+
"step": 570
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 2.0209059233449476,
|
| 963 |
+
"grad_norm": 69.01444268439244,
|
| 964 |
+
"learning_rate": 2.040176120880048e-07,
|
| 965 |
+
"logits/chosen": -2.742163896560669,
|
| 966 |
+
"logits/rejected": -2.739891529083252,
|
| 967 |
+
"logps/chosen": -66.87298583984375,
|
| 968 |
+
"logps/rejected": -92.02427673339844,
|
| 969 |
+
"loss": 0.3602,
|
| 970 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 971 |
+
"rewards/chosen": 5.334141731262207,
|
| 972 |
+
"rewards/margins": 10.981359481811523,
|
| 973 |
+
"rewards/rejected": -5.647217750549316,
|
| 974 |
+
"step": 580
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 2.0557491289198606,
|
| 978 |
+
"grad_norm": 0.14017497005010535,
|
| 979 |
+
"learning_rate": 1.9122991606322655e-07,
|
| 980 |
+
"logits/chosen": -2.6563334465026855,
|
| 981 |
+
"logits/rejected": -2.6057848930358887,
|
| 982 |
+
"logps/chosen": -80.1333236694336,
|
| 983 |
+
"logps/rejected": -92.28086853027344,
|
| 984 |
+
"loss": 0.3413,
|
| 985 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 986 |
+
"rewards/chosen": 6.459235191345215,
|
| 987 |
+
"rewards/margins": 12.748309135437012,
|
| 988 |
+
"rewards/rejected": -6.289073944091797,
|
| 989 |
+
"step": 590
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 2.0905923344947737,
|
| 993 |
+
"grad_norm": 0.4665024187650401,
|
| 994 |
+
"learning_rate": 1.7870375326612014e-07,
|
| 995 |
+
"logits/chosen": -2.6802449226379395,
|
| 996 |
+
"logits/rejected": -2.6527862548828125,
|
| 997 |
+
"logps/chosen": -52.02629470825195,
|
| 998 |
+
"logps/rejected": -71.58445739746094,
|
| 999 |
+
"loss": 0.3707,
|
| 1000 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 1001 |
+
"rewards/chosen": 4.42466926574707,
|
| 1002 |
+
"rewards/margins": 9.960657119750977,
|
| 1003 |
+
"rewards/rejected": -5.535989284515381,
|
| 1004 |
+
"step": 600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 2.0905923344947737,
|
| 1008 |
+
"eval_logits/chosen": -2.706523895263672,
|
| 1009 |
+
"eval_logits/rejected": -2.6904332637786865,
|
| 1010 |
+
"eval_logps/chosen": -72.95612335205078,
|
| 1011 |
+
"eval_logps/rejected": -82.87360382080078,
|
| 1012 |
+
"eval_loss": 0.8992136716842651,
|
| 1013 |
+
"eval_rewards/accuracies": 0.3472222089767456,
|
| 1014 |
+
"eval_rewards/chosen": 0.6907990574836731,
|
| 1015 |
+
"eval_rewards/margins": 1.4761611223220825,
|
| 1016 |
+
"eval_rewards/rejected": -0.7853620052337646,
|
| 1017 |
+
"eval_runtime": 114.0517,
|
| 1018 |
+
"eval_samples_per_second": 17.536,
|
| 1019 |
+
"eval_steps_per_second": 0.552,
|
| 1020 |
+
"step": 600
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 2.1254355400696863,
|
| 1024 |
+
"grad_norm": 2.1413870154178674,
|
| 1025 |
+
"learning_rate": 1.6645975735578165e-07,
|
| 1026 |
+
"logits/chosen": -2.7286293506622314,
|
| 1027 |
+
"logits/rejected": -2.6968209743499756,
|
| 1028 |
+
"logps/chosen": -73.44158172607422,
|
| 1029 |
+
"logps/rejected": -84.2601547241211,
|
| 1030 |
+
"loss": 0.3631,
|
| 1031 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 1032 |
+
"rewards/chosen": 4.519391059875488,
|
| 1033 |
+
"rewards/margins": 10.450955390930176,
|
| 1034 |
+
"rewards/rejected": -5.931563377380371,
|
| 1035 |
+
"step": 610
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 2.1602787456445993,
|
| 1039 |
+
"grad_norm": 7.439328512600098,
|
| 1040 |
+
"learning_rate": 1.5451809719331295e-07,
|
| 1041 |
+
"logits/chosen": -2.6922688484191895,
|
| 1042 |
+
"logits/rejected": -2.64046049118042,
|
| 1043 |
+
"logps/chosen": -66.28971099853516,
|
| 1044 |
+
"logps/rejected": -97.03089904785156,
|
| 1045 |
+
"loss": 0.3667,
|
| 1046 |
+
"rewards/accuracies": 0.518750011920929,
|
| 1047 |
+
"rewards/chosen": 5.458328723907471,
|
| 1048 |
+
"rewards/margins": 14.20256233215332,
|
| 1049 |
+
"rewards/rejected": -8.744232177734375,
|
| 1050 |
+
"step": 620
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 2.1951219512195124,
|
| 1054 |
+
"grad_norm": 1.6583151417624387,
|
| 1055 |
+
"learning_rate": 1.4289844361876528e-07,
|
| 1056 |
+
"logits/chosen": -2.6175191402435303,
|
| 1057 |
+
"logits/rejected": -2.6281955242156982,
|
| 1058 |
+
"logps/chosen": -62.218841552734375,
|
| 1059 |
+
"logps/rejected": -105.92720031738281,
|
| 1060 |
+
"loss": 0.3438,
|
| 1061 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1062 |
+
"rewards/chosen": 4.9540114402771,
|
| 1063 |
+
"rewards/margins": 13.651411056518555,
|
| 1064 |
+
"rewards/rejected": -8.69740104675293,
|
| 1065 |
+
"step": 630
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 2.229965156794425,
|
| 1069 |
+
"grad_norm": 0.8793732938037636,
|
| 1070 |
+
"learning_rate": 1.3161993704844647e-07,
|
| 1071 |
+
"logits/chosen": -2.642472505569458,
|
| 1072 |
+
"logits/rejected": -2.6352057456970215,
|
| 1073 |
+
"logps/chosen": -59.35064697265625,
|
| 1074 |
+
"logps/rejected": -101.8673324584961,
|
| 1075 |
+
"loss": 0.3678,
|
| 1076 |
+
"rewards/accuracies": 0.46875,
|
| 1077 |
+
"rewards/chosen": 4.605126857757568,
|
| 1078 |
+
"rewards/margins": 13.10174560546875,
|
| 1079 |
+
"rewards/rejected": -8.49661922454834,
|
| 1080 |
+
"step": 640
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 2.264808362369338,
|
| 1084 |
+
"grad_norm": 10.705002096620795,
|
| 1085 |
+
"learning_rate": 1.2070115594596576e-07,
|
| 1086 |
+
"logits/chosen": -2.6452431678771973,
|
| 1087 |
+
"logits/rejected": -2.619326114654541,
|
| 1088 |
+
"logps/chosen": -69.06687927246094,
|
| 1089 |
+
"logps/rejected": -92.87254333496094,
|
| 1090 |
+
"loss": 0.3908,
|
| 1091 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 1092 |
+
"rewards/chosen": 4.869100093841553,
|
| 1093 |
+
"rewards/margins": 11.985551834106445,
|
| 1094 |
+
"rewards/rejected": -7.116452217102051,
|
| 1095 |
+
"step": 650
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 2.2996515679442506,
|
| 1099 |
+
"grad_norm": 1.1342333232981456,
|
| 1100 |
+
"learning_rate": 1.1016008621895228e-07,
|
| 1101 |
+
"logits/chosen": -2.6762545108795166,
|
| 1102 |
+
"logits/rejected": -2.6614186763763428,
|
| 1103 |
+
"logps/chosen": -68.25106811523438,
|
| 1104 |
+
"logps/rejected": -91.7679214477539,
|
| 1105 |
+
"loss": 0.3573,
|
| 1106 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 1107 |
+
"rewards/chosen": 4.843748569488525,
|
| 1108 |
+
"rewards/margins": 12.167219161987305,
|
| 1109 |
+
"rewards/rejected": -7.323469638824463,
|
| 1110 |
+
"step": 660
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 2.3344947735191637,
|
| 1114 |
+
"grad_norm": 0.6331495120745031,
|
| 1115 |
+
"learning_rate": 1.000140915918589e-07,
|
| 1116 |
+
"logits/chosen": -2.6580915451049805,
|
| 1117 |
+
"logits/rejected": -2.6534488201141357,
|
| 1118 |
+
"logps/chosen": -65.1015396118164,
|
| 1119 |
+
"logps/rejected": -91.0811538696289,
|
| 1120 |
+
"loss": 0.3679,
|
| 1121 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 1122 |
+
"rewards/chosen": 3.4471499919891357,
|
| 1123 |
+
"rewards/margins": 11.26557731628418,
|
| 1124 |
+
"rewards/rejected": -7.818428039550781,
|
| 1125 |
+
"step": 670
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 2.3693379790940767,
|
| 1129 |
+
"grad_norm": 2.2726954763261324,
|
| 1130 |
+
"learning_rate": 9.027988500365347e-08,
|
| 1131 |
+
"logits/chosen": -2.693312406539917,
|
| 1132 |
+
"logits/rejected": -2.680427312850952,
|
| 1133 |
+
"logps/chosen": -64.97016906738281,
|
| 1134 |
+
"logps/rejected": -90.56646728515625,
|
| 1135 |
+
"loss": 0.3555,
|
| 1136 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 1137 |
+
"rewards/chosen": 3.795478343963623,
|
| 1138 |
+
"rewards/margins": 12.02349853515625,
|
| 1139 |
+
"rewards/rejected": -8.228020668029785,
|
| 1140 |
+
"step": 680
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 2.40418118466899,
|
| 1144 |
+
"grad_norm": 0.22531368042473615,
|
| 1145 |
+
"learning_rate": 8.097350107751374e-08,
|
| 1146 |
+
"logits/chosen": -2.6593518257141113,
|
| 1147 |
+
"logits/rejected": -2.629815101623535,
|
| 1148 |
+
"logps/chosen": -72.8058853149414,
|
| 1149 |
+
"logps/rejected": -104.3930435180664,
|
| 1150 |
+
"loss": 0.3723,
|
| 1151 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 1152 |
+
"rewards/chosen": 4.151662349700928,
|
| 1153 |
+
"rewards/margins": 11.557449340820312,
|
| 1154 |
+
"rewards/rejected": -7.405786991119385,
|
| 1155 |
+
"step": 690
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 2.4390243902439024,
|
| 1159 |
+
"grad_norm": 2.7004181516244823,
|
| 1160 |
+
"learning_rate": 7.211026970787468e-08,
|
| 1161 |
+
"logits/chosen": -2.6526448726654053,
|
| 1162 |
+
"logits/rejected": -2.6184093952178955,
|
| 1163 |
+
"logps/chosen": -78.18067932128906,
|
| 1164 |
+
"logps/rejected": -97.9962158203125,
|
| 1165 |
+
"loss": 0.3672,
|
| 1166 |
+
"rewards/accuracies": 0.53125,
|
| 1167 |
+
"rewards/chosen": 4.823997497558594,
|
| 1168 |
+
"rewards/margins": 12.184492111206055,
|
| 1169 |
+
"rewards/rejected": -7.360495090484619,
|
| 1170 |
+
"step": 700
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 2.4390243902439024,
|
| 1174 |
+
"eval_logits/chosen": -2.6822805404663086,
|
| 1175 |
+
"eval_logits/rejected": -2.666159152984619,
|
| 1176 |
+
"eval_logps/chosen": -75.6268539428711,
|
| 1177 |
+
"eval_logps/rejected": -86.10513305664062,
|
| 1178 |
+
"eval_loss": 0.9353695511817932,
|
| 1179 |
+
"eval_rewards/accuracies": 0.3492063581943512,
|
| 1180 |
+
"eval_rewards/chosen": -0.5110280513763428,
|
| 1181 |
+
"eval_rewards/margins": 1.7285233736038208,
|
| 1182 |
+
"eval_rewards/rejected": -2.239551305770874,
|
| 1183 |
+
"eval_runtime": 113.8285,
|
| 1184 |
+
"eval_samples_per_second": 17.57,
|
| 1185 |
+
"eval_steps_per_second": 0.553,
|
| 1186 |
+
"step": 700
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 2.4738675958188154,
|
| 1190 |
+
"grad_norm": 0.0879833853193303,
|
| 1191 |
+
"learning_rate": 6.370479080833579e-08,
|
| 1192 |
+
"logits/chosen": -2.700568675994873,
|
| 1193 |
+
"logits/rejected": -2.6803741455078125,
|
| 1194 |
+
"logps/chosen": -84.60392761230469,
|
| 1195 |
+
"logps/rejected": -123.07450103759766,
|
| 1196 |
+
"loss": 0.3643,
|
| 1197 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1198 |
+
"rewards/chosen": 4.764341831207275,
|
| 1199 |
+
"rewards/margins": 15.03064250946045,
|
| 1200 |
+
"rewards/rejected": -10.266302108764648,
|
| 1201 |
+
"step": 710
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 2.508710801393728,
|
| 1205 |
+
"grad_norm": 1.1049586982136799,
|
| 1206 |
+
"learning_rate": 5.5770910262027175e-08,
|
| 1207 |
+
"logits/chosen": -2.6708908081054688,
|
| 1208 |
+
"logits/rejected": -2.638007164001465,
|
| 1209 |
+
"logps/chosen": -60.929107666015625,
|
| 1210 |
+
"logps/rejected": -78.13368225097656,
|
| 1211 |
+
"loss": 0.4049,
|
| 1212 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 1213 |
+
"rewards/chosen": 4.554830551147461,
|
| 1214 |
+
"rewards/margins": 11.321012496948242,
|
| 1215 |
+
"rewards/rejected": -6.766181945800781,
|
| 1216 |
+
"step": 720
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 2.543554006968641,
|
| 1220 |
+
"grad_norm": 0.17464151203641626,
|
| 1221 |
+
"learning_rate": 4.832169711404716e-08,
|
| 1222 |
+
"logits/chosen": -2.702913284301758,
|
| 1223 |
+
"logits/rejected": -2.682276725769043,
|
| 1224 |
+
"logps/chosen": -61.146484375,
|
| 1225 |
+
"logps/rejected": -76.15191650390625,
|
| 1226 |
+
"loss": 0.384,
|
| 1227 |
+
"rewards/accuracies": 0.34375,
|
| 1228 |
+
"rewards/chosen": 4.145053863525391,
|
| 1229 |
+
"rewards/margins": 10.881394386291504,
|
| 1230 |
+
"rewards/rejected": -6.7363386154174805,
|
| 1231 |
+
"step": 730
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 2.578397212543554,
|
| 1235 |
+
"grad_norm": 1.4963244248499354,
|
| 1236 |
+
"learning_rate": 4.1369422043543185e-08,
|
| 1237 |
+
"logits/chosen": -2.6026740074157715,
|
| 1238 |
+
"logits/rejected": -2.5997722148895264,
|
| 1239 |
+
"logps/chosen": -58.41170120239258,
|
| 1240 |
+
"logps/rejected": -99.71952056884766,
|
| 1241 |
+
"loss": 0.3645,
|
| 1242 |
+
"rewards/accuracies": 0.4375,
|
| 1243 |
+
"rewards/chosen": 3.420952320098877,
|
| 1244 |
+
"rewards/margins": 12.743513107299805,
|
| 1245 |
+
"rewards/rejected": -9.322561264038086,
|
| 1246 |
+
"step": 740
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 2.6132404181184667,
|
| 1250 |
+
"grad_norm": 0.17470713028005413,
|
| 1251 |
+
"learning_rate": 3.492553715089692e-08,
|
| 1252 |
+
"logits/chosen": -2.6855924129486084,
|
| 1253 |
+
"logits/rejected": -2.6744086742401123,
|
| 1254 |
+
"logps/chosen": -59.6028938293457,
|
| 1255 |
+
"logps/rejected": -80.73404693603516,
|
| 1256 |
+
"loss": 0.3419,
|
| 1257 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 1258 |
+
"rewards/chosen": 3.875894069671631,
|
| 1259 |
+
"rewards/margins": 9.938748359680176,
|
| 1260 |
+
"rewards/rejected": -6.062855243682861,
|
| 1261 |
+
"step": 750
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 2.64808362369338,
|
| 1265 |
+
"grad_norm": 0.7912656232401193,
|
| 1266 |
+
"learning_rate": 2.9000657093309096e-08,
|
| 1267 |
+
"logits/chosen": -2.6053149700164795,
|
| 1268 |
+
"logits/rejected": -2.594897985458374,
|
| 1269 |
+
"logps/chosen": -45.781890869140625,
|
| 1270 |
+
"logps/rejected": -70.77694702148438,
|
| 1271 |
+
"loss": 0.3866,
|
| 1272 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 1273 |
+
"rewards/chosen": 3.3890724182128906,
|
| 1274 |
+
"rewards/margins": 9.678762435913086,
|
| 1275 |
+
"rewards/rejected": -6.289690017700195,
|
| 1276 |
+
"step": 760
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 2.682926829268293,
|
| 1280 |
+
"grad_norm": 0.03850043403510083,
|
| 1281 |
+
"learning_rate": 2.3604541599858524e-08,
|
| 1282 |
+
"logits/chosen": -2.7165355682373047,
|
| 1283 |
+
"logits/rejected": -2.698086738586426,
|
| 1284 |
+
"logps/chosen": -67.92710876464844,
|
| 1285 |
+
"logps/rejected": -97.56449890136719,
|
| 1286 |
+
"loss": 0.3756,
|
| 1287 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 1288 |
+
"rewards/chosen": 3.9307703971862793,
|
| 1289 |
+
"rewards/margins": 12.25821590423584,
|
| 1290 |
+
"rewards/rejected": -8.327445030212402,
|
| 1291 |
+
"step": 770
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 2.7177700348432055,
|
| 1295 |
+
"grad_norm": 0.23519403290223553,
|
| 1296 |
+
"learning_rate": 1.8746079394836706e-08,
|
| 1297 |
+
"logits/chosen": -2.656900405883789,
|
| 1298 |
+
"logits/rejected": -2.6522555351257324,
|
| 1299 |
+
"logps/chosen": -68.84635925292969,
|
| 1300 |
+
"logps/rejected": -102.07672119140625,
|
| 1301 |
+
"loss": 0.3447,
|
| 1302 |
+
"rewards/accuracies": 0.512499988079071,
|
| 1303 |
+
"rewards/chosen": 5.431530475616455,
|
| 1304 |
+
"rewards/margins": 14.358503341674805,
|
| 1305 |
+
"rewards/rejected": -8.926973342895508,
|
| 1306 |
+
"step": 780
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 2.7526132404181185,
|
| 1310 |
+
"grad_norm": 0.605439153331925,
|
| 1311 |
+
"learning_rate": 1.4433273555842e-08,
|
| 1312 |
+
"logits/chosen": -2.6824231147766113,
|
| 1313 |
+
"logits/rejected": -2.6712381839752197,
|
| 1314 |
+
"logps/chosen": -57.7734260559082,
|
| 1315 |
+
"logps/rejected": -90.67354583740234,
|
| 1316 |
+
"loss": 0.3745,
|
| 1317 |
+
"rewards/accuracies": 0.46875,
|
| 1318 |
+
"rewards/chosen": 4.336173057556152,
|
| 1319 |
+
"rewards/margins": 11.832128524780273,
|
| 1320 |
+
"rewards/rejected": -7.495955467224121,
|
| 1321 |
+
"step": 790
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 2.7874564459930316,
|
| 1325 |
+
"grad_norm": 8.838304462962064,
|
| 1326 |
+
"learning_rate": 1.0673228330749007e-08,
|
| 1327 |
+
"logits/chosen": -2.649266242980957,
|
| 1328 |
+
"logits/rejected": -2.6187663078308105,
|
| 1329 |
+
"logps/chosen": -66.30155944824219,
|
| 1330 |
+
"logps/rejected": -96.27810668945312,
|
| 1331 |
+
"loss": 0.3596,
|
| 1332 |
+
"rewards/accuracies": 0.46875,
|
| 1333 |
+
"rewards/chosen": 6.211874961853027,
|
| 1334 |
+
"rewards/margins": 15.416938781738281,
|
| 1335 |
+
"rewards/rejected": -9.205063819885254,
|
| 1336 |
+
"step": 800
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.7874564459930316,
|
| 1340 |
+
"eval_logits/chosen": -2.6886420249938965,
|
| 1341 |
+
"eval_logits/rejected": -2.672652006149292,
|
| 1342 |
+
"eval_logps/chosen": -75.24068450927734,
|
| 1343 |
+
"eval_logps/rejected": -85.62494659423828,
|
| 1344 |
+
"eval_loss": 0.9344265460968018,
|
| 1345 |
+
"eval_rewards/accuracies": 0.3452380895614624,
|
| 1346 |
+
"eval_rewards/chosen": -0.33725666999816895,
|
| 1347 |
+
"eval_rewards/margins": 1.686206340789795,
|
| 1348 |
+
"eval_rewards/rejected": -2.023463249206543,
|
| 1349 |
+
"eval_runtime": 114.2576,
|
| 1350 |
+
"eval_samples_per_second": 17.504,
|
| 1351 |
+
"eval_steps_per_second": 0.551,
|
| 1352 |
+
"step": 800
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 2.822299651567944,
|
| 1356 |
+
"grad_norm": 0.5583388154267291,
|
| 1357 |
+
"learning_rate": 7.472137435272619e-09,
|
| 1358 |
+
"logits/chosen": -2.6408095359802246,
|
| 1359 |
+
"logits/rejected": -2.6051411628723145,
|
| 1360 |
+
"logps/chosen": -79.69121551513672,
|
| 1361 |
+
"logps/rejected": -97.32070922851562,
|
| 1362 |
+
"loss": 0.3607,
|
| 1363 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 1364 |
+
"rewards/chosen": 4.549583435058594,
|
| 1365 |
+
"rewards/margins": 12.599637985229492,
|
| 1366 |
+
"rewards/rejected": -8.050054550170898,
|
| 1367 |
+
"step": 810
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 2.857142857142857,
|
| 1371 |
+
"grad_norm": 2.894525046100427,
|
| 1372 |
+
"learning_rate": 4.835273850400123e-09,
|
| 1373 |
+
"logits/chosen": -2.661485195159912,
|
| 1374 |
+
"logits/rejected": -2.6633734703063965,
|
| 1375 |
+
"logps/chosen": -79.87845611572266,
|
| 1376 |
+
"logps/rejected": -116.92414855957031,
|
| 1377 |
+
"loss": 0.3792,
|
| 1378 |
+
"rewards/accuracies": 0.512499988079071,
|
| 1379 |
+
"rewards/chosen": 5.126074314117432,
|
| 1380 |
+
"rewards/margins": 13.770315170288086,
|
| 1381 |
+
"rewards/rejected": -8.644243240356445,
|
| 1382 |
+
"step": 820
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 2.89198606271777,
|
| 1386 |
+
"grad_norm": 0.19653863059909565,
|
| 1387 |
+
"learning_rate": 2.766981136500024e-09,
|
| 1388 |
+
"logits/chosen": -2.5719292163848877,
|
| 1389 |
+
"logits/rejected": -2.577578067779541,
|
| 1390 |
+
"logps/chosen": -54.612770080566406,
|
| 1391 |
+
"logps/rejected": -91.96238708496094,
|
| 1392 |
+
"loss": 0.3571,
|
| 1393 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 1394 |
+
"rewards/chosen": 3.002854585647583,
|
| 1395 |
+
"rewards/margins": 11.87660026550293,
|
| 1396 |
+
"rewards/rejected": -8.873745918273926,
|
| 1397 |
+
"step": 830
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 2.926829268292683,
|
| 1401 |
+
"grad_norm": 0.16246150387256475,
|
| 1402 |
+
"learning_rate": 1.2706662784136513e-09,
|
| 1403 |
+
"logits/chosen": -2.6564536094665527,
|
| 1404 |
+
"logits/rejected": -2.619932174682617,
|
| 1405 |
+
"logps/chosen": -71.76846313476562,
|
| 1406 |
+
"logps/rejected": -101.76387023925781,
|
| 1407 |
+
"loss": 0.3726,
|
| 1408 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 1409 |
+
"rewards/chosen": 4.6055145263671875,
|
| 1410 |
+
"rewards/margins": 12.795547485351562,
|
| 1411 |
+
"rewards/rejected": -8.190032958984375,
|
| 1412 |
+
"step": 840
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 2.961672473867596,
|
| 1416 |
+
"grad_norm": 0.90921168828385,
|
| 1417 |
+
"learning_rate": 3.4879407331657175e-10,
|
| 1418 |
+
"logits/chosen": -2.68048357963562,
|
| 1419 |
+
"logits/rejected": -2.6529481410980225,
|
| 1420 |
+
"logps/chosen": -80.40882873535156,
|
| 1421 |
+
"logps/rejected": -104.63221740722656,
|
| 1422 |
+
"loss": 0.3615,
|
| 1423 |
+
"rewards/accuracies": 0.518750011920929,
|
| 1424 |
+
"rewards/chosen": 5.170698165893555,
|
| 1425 |
+
"rewards/margins": 13.889617919921875,
|
| 1426 |
+
"rewards/rejected": -8.718920707702637,
|
| 1427 |
+
"step": 850
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.996515679442509,
|
| 1431 |
+
"grad_norm": 1.296285561761296,
|
| 1432 |
+
"learning_rate": 2.8830705936344624e-12,
|
| 1433 |
+
"logits/chosen": -2.618508815765381,
|
| 1434 |
+
"logits/rejected": -2.619737386703491,
|
| 1435 |
+
"logps/chosen": -62.67692947387695,
|
| 1436 |
+
"logps/rejected": -99.06525421142578,
|
| 1437 |
+
"loss": 0.3391,
|
| 1438 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 1439 |
+
"rewards/chosen": 5.471669673919678,
|
| 1440 |
+
"rewards/margins": 15.48302936553955,
|
| 1441 |
+
"rewards/rejected": -10.011358261108398,
|
| 1442 |
+
"step": 860
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 3.0,
|
| 1446 |
+
"step": 861,
|
| 1447 |
+
"total_flos": 0.0,
|
| 1448 |
+
"train_loss": 0.49774221859604084,
|
| 1449 |
+
"train_runtime": 9884.8529,
|
| 1450 |
+
"train_samples_per_second": 5.566,
|
| 1451 |
+
"train_steps_per_second": 0.087
|
| 1452 |
+
}
|
| 1453 |
+
],
|
| 1454 |
+
"logging_steps": 10,
|
| 1455 |
+
"max_steps": 861,
|
| 1456 |
+
"num_input_tokens_seen": 0,
|
| 1457 |
+
"num_train_epochs": 3,
|
| 1458 |
+
"save_steps": 100,
|
| 1459 |
+
"stateful_callbacks": {
|
| 1460 |
+
"TrainerControl": {
|
| 1461 |
+
"args": {
|
| 1462 |
+
"should_epoch_stop": false,
|
| 1463 |
+
"should_evaluate": false,
|
| 1464 |
+
"should_log": false,
|
| 1465 |
+
"should_save": true,
|
| 1466 |
+
"should_training_stop": true
|
| 1467 |
+
},
|
| 1468 |
+
"attributes": {}
|
| 1469 |
+
}
|
| 1470 |
+
},
|
| 1471 |
+
"total_flos": 0.0,
|
| 1472 |
+
"train_batch_size": 8,
|
| 1473 |
+
"trial_name": null,
|
| 1474 |
+
"trial_params": null
|
| 1475 |
+
}
|