Model save
Browse files- README.md +61 -0
- all_results.json +9 -0
- generation_config.json +7 -0
- train_results.json +9 -0
- trainer_state.json +1182 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- dpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: tinyllama-1.1b-chat-v1.0-ui-coding-dpo-2
|
| 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 |
+
# tinyllama-1.1b-chat-v1.0-ui-coding-dpo-2
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-Chat-v1.0](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0) on an unknown dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-07
|
| 39 |
+
- train_batch_size: 8
|
| 40 |
+
- eval_batch_size: 8
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 8
|
| 44 |
+
- gradient_accumulation_steps: 4
|
| 45 |
+
- total_train_batch_size: 256
|
| 46 |
+
- total_eval_batch_size: 64
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: cosine
|
| 49 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 50 |
+
- num_epochs: 2
|
| 51 |
+
|
| 52 |
+
### Training results
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
### Framework versions
|
| 57 |
+
|
| 58 |
+
- Transformers 4.44.1
|
| 59 |
+
- Pytorch 2.1.2+cu121
|
| 60 |
+
- Datasets 2.21.0
|
| 61 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.9960291197882198,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.26563876598520053,
|
| 5 |
+
"train_runtime": 5267.2293,
|
| 6 |
+
"train_samples": 96652,
|
| 7 |
+
"train_samples_per_second": 36.699,
|
| 8 |
+
"train_steps_per_second": 0.143
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 1,
|
| 3 |
+
"eos_token_id": 2,
|
| 4 |
+
"max_length": 2048,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.44.1"
|
| 7 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.9960291197882198,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.26563876598520053,
|
| 5 |
+
"train_runtime": 5267.2293,
|
| 6 |
+
"train_samples": 96652,
|
| 7 |
+
"train_samples_per_second": 36.699,
|
| 8 |
+
"train_steps_per_second": 0.143
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1182 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.9960291197882198,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 754,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0026472534745201853,
|
| 13 |
+
"grad_norm": 4.30932258831093,
|
| 14 |
+
"learning_rate": 6.578947368421052e-09,
|
| 15 |
+
"logits/chosen": -2.923454761505127,
|
| 16 |
+
"logits/rejected": -3.022336483001709,
|
| 17 |
+
"logps/chosen": -491.803955078125,
|
| 18 |
+
"logps/rejected": -509.828369140625,
|
| 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": 0.026472534745201854,
|
| 28 |
+
"grad_norm": 4.3524814860184176,
|
| 29 |
+
"learning_rate": 6.578947368421052e-08,
|
| 30 |
+
"logits/chosen": -2.849837303161621,
|
| 31 |
+
"logits/rejected": -2.918842315673828,
|
| 32 |
+
"logps/chosen": -482.8021240234375,
|
| 33 |
+
"logps/rejected": -468.6262512207031,
|
| 34 |
+
"loss": 0.6931,
|
| 35 |
+
"rewards/accuracies": 0.4444444477558136,
|
| 36 |
+
"rewards/chosen": 1.7793978258850984e-05,
|
| 37 |
+
"rewards/margins": 0.00012061676534358412,
|
| 38 |
+
"rewards/rejected": -0.00010282275616191328,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.05294506949040371,
|
| 43 |
+
"grad_norm": 4.058367941679181,
|
| 44 |
+
"learning_rate": 1.3157894736842104e-07,
|
| 45 |
+
"logits/chosen": -2.8850979804992676,
|
| 46 |
+
"logits/rejected": -2.9686431884765625,
|
| 47 |
+
"logps/chosen": -492.1334533691406,
|
| 48 |
+
"logps/rejected": -473.5765075683594,
|
| 49 |
+
"loss": 0.6929,
|
| 50 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 51 |
+
"rewards/chosen": 0.00013926837709732354,
|
| 52 |
+
"rewards/margins": 0.0005674505373463035,
|
| 53 |
+
"rewards/rejected": -0.0004281821602489799,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.07941760423560557,
|
| 58 |
+
"grad_norm": 4.100822256989,
|
| 59 |
+
"learning_rate": 1.9736842105263157e-07,
|
| 60 |
+
"logits/chosen": -2.868659496307373,
|
| 61 |
+
"logits/rejected": -2.9573891162872314,
|
| 62 |
+
"logps/chosen": -475.2897033691406,
|
| 63 |
+
"logps/rejected": -473.85064697265625,
|
| 64 |
+
"loss": 0.6906,
|
| 65 |
+
"rewards/accuracies": 0.715624988079071,
|
| 66 |
+
"rewards/chosen": 0.003203944070264697,
|
| 67 |
+
"rewards/margins": 0.006764715071767569,
|
| 68 |
+
"rewards/rejected": -0.0035607716999948025,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.10589013898080742,
|
| 73 |
+
"grad_norm": 4.224597004906648,
|
| 74 |
+
"learning_rate": 2.631578947368421e-07,
|
| 75 |
+
"logits/chosen": -2.880610704421997,
|
| 76 |
+
"logits/rejected": -2.9531335830688477,
|
| 77 |
+
"logps/chosen": -477.7626953125,
|
| 78 |
+
"logps/rejected": -471.15203857421875,
|
| 79 |
+
"loss": 0.6838,
|
| 80 |
+
"rewards/accuracies": 0.7906249761581421,
|
| 81 |
+
"rewards/chosen": 0.0102998660877347,
|
| 82 |
+
"rewards/margins": 0.02004994824528694,
|
| 83 |
+
"rewards/rejected": -0.009750082157552242,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.13236267372600927,
|
| 88 |
+
"grad_norm": 4.481790796785453,
|
| 89 |
+
"learning_rate": 3.2894736842105264e-07,
|
| 90 |
+
"logits/chosen": -2.910224199295044,
|
| 91 |
+
"logits/rejected": -2.966555118560791,
|
| 92 |
+
"logps/chosen": -496.03704833984375,
|
| 93 |
+
"logps/rejected": -483.69970703125,
|
| 94 |
+
"loss": 0.6705,
|
| 95 |
+
"rewards/accuracies": 0.8031250238418579,
|
| 96 |
+
"rewards/chosen": 0.02056797966361046,
|
| 97 |
+
"rewards/margins": 0.04747764393687248,
|
| 98 |
+
"rewards/rejected": -0.026909660547971725,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.15883520847121113,
|
| 103 |
+
"grad_norm": 4.727420695183077,
|
| 104 |
+
"learning_rate": 3.9473684210526315e-07,
|
| 105 |
+
"logits/chosen": -2.908859968185425,
|
| 106 |
+
"logits/rejected": -2.9846339225769043,
|
| 107 |
+
"logps/chosen": -487.971923828125,
|
| 108 |
+
"logps/rejected": -487.698486328125,
|
| 109 |
+
"loss": 0.6321,
|
| 110 |
+
"rewards/accuracies": 0.875,
|
| 111 |
+
"rewards/chosen": 0.047008998692035675,
|
| 112 |
+
"rewards/margins": 0.12849071621894836,
|
| 113 |
+
"rewards/rejected": -0.08148171752691269,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.18530774321641297,
|
| 118 |
+
"grad_norm": 5.3651750883138645,
|
| 119 |
+
"learning_rate": 4.6052631578947365e-07,
|
| 120 |
+
"logits/chosen": -2.9276702404022217,
|
| 121 |
+
"logits/rejected": -2.958789348602295,
|
| 122 |
+
"logps/chosen": -488.4877014160156,
|
| 123 |
+
"logps/rejected": -489.53546142578125,
|
| 124 |
+
"loss": 0.5937,
|
| 125 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 126 |
+
"rewards/chosen": 0.025530118495225906,
|
| 127 |
+
"rewards/margins": 0.22059020400047302,
|
| 128 |
+
"rewards/rejected": -0.19506008923053741,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.21178027796161483,
|
| 133 |
+
"grad_norm": 3.6311630185736297,
|
| 134 |
+
"learning_rate": 4.999570604073014e-07,
|
| 135 |
+
"logits/chosen": -2.923982620239258,
|
| 136 |
+
"logits/rejected": -2.9833145141601562,
|
| 137 |
+
"logps/chosen": -503.56134033203125,
|
| 138 |
+
"logps/rejected": -552.4190673828125,
|
| 139 |
+
"loss": 0.4934,
|
| 140 |
+
"rewards/accuracies": 0.840624988079071,
|
| 141 |
+
"rewards/chosen": -0.13993698358535767,
|
| 142 |
+
"rewards/margins": 0.5944468975067139,
|
| 143 |
+
"rewards/rejected": -0.7343839406967163,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.23825281270681667,
|
| 148 |
+
"grad_norm": 4.662865437270245,
|
| 149 |
+
"learning_rate": 4.994741593733563e-07,
|
| 150 |
+
"logits/chosen": -2.9195408821105957,
|
| 151 |
+
"logits/rejected": -2.958824872970581,
|
| 152 |
+
"logps/chosen": -534.3409423828125,
|
| 153 |
+
"logps/rejected": -607.5826416015625,
|
| 154 |
+
"loss": 0.4331,
|
| 155 |
+
"rewards/accuracies": 0.871874988079071,
|
| 156 |
+
"rewards/chosen": -0.41773176193237305,
|
| 157 |
+
"rewards/margins": 0.9263063669204712,
|
| 158 |
+
"rewards/rejected": -1.3440382480621338,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.26472534745201853,
|
| 163 |
+
"grad_norm": 4.763826147961603,
|
| 164 |
+
"learning_rate": 4.984557228946769e-07,
|
| 165 |
+
"logits/chosen": -2.815560817718506,
|
| 166 |
+
"logits/rejected": -2.8625988960266113,
|
| 167 |
+
"logps/chosen": -611.6197509765625,
|
| 168 |
+
"logps/rejected": -758.5867309570312,
|
| 169 |
+
"loss": 0.3704,
|
| 170 |
+
"rewards/accuracies": 0.8343750238418579,
|
| 171 |
+
"rewards/chosen": -1.2842621803283691,
|
| 172 |
+
"rewards/margins": 1.4864912033081055,
|
| 173 |
+
"rewards/rejected": -2.7707533836364746,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.29119788219722037,
|
| 178 |
+
"grad_norm": 4.297222720122574,
|
| 179 |
+
"learning_rate": 4.969039372050355e-07,
|
| 180 |
+
"logits/chosen": -2.787930488586426,
|
| 181 |
+
"logits/rejected": -2.8286757469177246,
|
| 182 |
+
"logps/chosen": -622.7388305664062,
|
| 183 |
+
"logps/rejected": -840.2802734375,
|
| 184 |
+
"loss": 0.3234,
|
| 185 |
+
"rewards/accuracies": 0.8531249761581421,
|
| 186 |
+
"rewards/chosen": -1.4110974073410034,
|
| 187 |
+
"rewards/margins": 2.2375450134277344,
|
| 188 |
+
"rewards/rejected": -3.648642063140869,
|
| 189 |
+
"step": 110
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.31767041694242226,
|
| 193 |
+
"grad_norm": 3.7535819941587065,
|
| 194 |
+
"learning_rate": 4.948221334560093e-07,
|
| 195 |
+
"logits/chosen": -2.831071615219116,
|
| 196 |
+
"logits/rejected": -2.843599796295166,
|
| 197 |
+
"logps/chosen": -592.7720947265625,
|
| 198 |
+
"logps/rejected": -807.1361083984375,
|
| 199 |
+
"loss": 0.3176,
|
| 200 |
+
"rewards/accuracies": 0.8531249761581421,
|
| 201 |
+
"rewards/chosen": -1.023418664932251,
|
| 202 |
+
"rewards/margins": 2.2339682579040527,
|
| 203 |
+
"rewards/rejected": -3.2573866844177246,
|
| 204 |
+
"step": 120
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.3441429516876241,
|
| 208 |
+
"grad_norm": 4.4946406733821185,
|
| 209 |
+
"learning_rate": 4.922147805661402e-07,
|
| 210 |
+
"logits/chosen": -2.83577823638916,
|
| 211 |
+
"logits/rejected": -2.875858783721924,
|
| 212 |
+
"logps/chosen": -629.0977783203125,
|
| 213 |
+
"logps/rejected": -913.8206787109375,
|
| 214 |
+
"loss": 0.2875,
|
| 215 |
+
"rewards/accuracies": 0.862500011920929,
|
| 216 |
+
"rewards/chosen": -1.3898423910140991,
|
| 217 |
+
"rewards/margins": 2.828009605407715,
|
| 218 |
+
"rewards/rejected": -4.2178521156311035,
|
| 219 |
+
"step": 130
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.37061548643282594,
|
| 223 |
+
"grad_norm": 4.541566634311875,
|
| 224 |
+
"learning_rate": 4.890874756276999e-07,
|
| 225 |
+
"logits/chosen": -2.7425622940063477,
|
| 226 |
+
"logits/rejected": -2.8039002418518066,
|
| 227 |
+
"logps/chosen": -634.9403686523438,
|
| 228 |
+
"logps/rejected": -911.8308715820312,
|
| 229 |
+
"loss": 0.2831,
|
| 230 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 231 |
+
"rewards/chosen": -1.5872424840927124,
|
| 232 |
+
"rewards/margins": 2.6655924320220947,
|
| 233 |
+
"rewards/rejected": -4.252835273742676,
|
| 234 |
+
"step": 140
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.3970880211780278,
|
| 238 |
+
"grad_norm": 3.3580041177607978,
|
| 239 |
+
"learning_rate": 4.854469318916532e-07,
|
| 240 |
+
"logits/chosen": -2.8008108139038086,
|
| 241 |
+
"logits/rejected": -2.8558998107910156,
|
| 242 |
+
"logps/chosen": -646.1107177734375,
|
| 243 |
+
"logps/rejected": -980.4015502929688,
|
| 244 |
+
"loss": 0.2668,
|
| 245 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 246 |
+
"rewards/chosen": -1.5013834238052368,
|
| 247 |
+
"rewards/margins": 3.364828109741211,
|
| 248 |
+
"rewards/rejected": -4.866211414337158,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.42356055592322966,
|
| 253 |
+
"grad_norm": 4.675914141471813,
|
| 254 |
+
"learning_rate": 4.8130096435661e-07,
|
| 255 |
+
"logits/chosen": -2.80204176902771,
|
| 256 |
+
"logits/rejected": -2.8504693508148193,
|
| 257 |
+
"logps/chosen": -631.5078125,
|
| 258 |
+
"logps/rejected": -925.1160888671875,
|
| 259 |
+
"loss": 0.2519,
|
| 260 |
+
"rewards/accuracies": 0.893750011920929,
|
| 261 |
+
"rewards/chosen": -1.4621139764785767,
|
| 262 |
+
"rewards/margins": 2.9786953926086426,
|
| 263 |
+
"rewards/rejected": -4.440809726715088,
|
| 264 |
+
"step": 160
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.4500330906684315,
|
| 268 |
+
"grad_norm": 6.483714450702139,
|
| 269 |
+
"learning_rate": 4.766584729927049e-07,
|
| 270 |
+
"logits/chosen": -2.768064498901367,
|
| 271 |
+
"logits/rejected": -2.8086233139038086,
|
| 272 |
+
"logps/chosen": -676.531982421875,
|
| 273 |
+
"logps/rejected": -1029.0057373046875,
|
| 274 |
+
"loss": 0.2799,
|
| 275 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 276 |
+
"rewards/chosen": -1.8374897241592407,
|
| 277 |
+
"rewards/margins": 3.5798168182373047,
|
| 278 |
+
"rewards/rejected": -5.417306900024414,
|
| 279 |
+
"step": 170
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.47650562541363334,
|
| 283 |
+
"grad_norm": 5.396515481877322,
|
| 284 |
+
"learning_rate": 4.7152942363641345e-07,
|
| 285 |
+
"logits/chosen": -2.7385268211364746,
|
| 286 |
+
"logits/rejected": -2.760967493057251,
|
| 287 |
+
"logps/chosen": -642.5712890625,
|
| 288 |
+
"logps/rejected": -997.8201904296875,
|
| 289 |
+
"loss": 0.2558,
|
| 290 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 291 |
+
"rewards/chosen": -1.627642273902893,
|
| 292 |
+
"rewards/margins": 3.5839648246765137,
|
| 293 |
+
"rewards/rejected": -5.211607456207275,
|
| 294 |
+
"step": 180
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.5029781601588352,
|
| 298 |
+
"grad_norm": 5.261191188089043,
|
| 299 |
+
"learning_rate": 4.6592482659732045e-07,
|
| 300 |
+
"logits/chosen": -2.7847423553466797,
|
| 301 |
+
"logits/rejected": -2.8355870246887207,
|
| 302 |
+
"logps/chosen": -674.9547729492188,
|
| 303 |
+
"logps/rejected": -972.1364135742188,
|
| 304 |
+
"loss": 0.2536,
|
| 305 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 306 |
+
"rewards/chosen": -1.8109451532363892,
|
| 307 |
+
"rewards/margins": 3.033442258834839,
|
| 308 |
+
"rewards/rejected": -4.844388008117676,
|
| 309 |
+
"step": 190
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.5294506949040371,
|
| 313 |
+
"grad_norm": 4.9133890145951495,
|
| 314 |
+
"learning_rate": 4.5985671302276166e-07,
|
| 315 |
+
"logits/chosen": -2.7557764053344727,
|
| 316 |
+
"logits/rejected": -2.8086917400360107,
|
| 317 |
+
"logps/chosen": -681.2697143554688,
|
| 318 |
+
"logps/rejected": -989.4937744140625,
|
| 319 |
+
"loss": 0.2602,
|
| 320 |
+
"rewards/accuracies": 0.890625,
|
| 321 |
+
"rewards/chosen": -1.8896667957305908,
|
| 322 |
+
"rewards/margins": 3.2132556438446045,
|
| 323 |
+
"rewards/rejected": -5.102922439575195,
|
| 324 |
+
"step": 200
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.5559232296492389,
|
| 328 |
+
"grad_norm": 4.892431641268176,
|
| 329 |
+
"learning_rate": 4.533381090710776e-07,
|
| 330 |
+
"logits/chosen": -2.7014384269714355,
|
| 331 |
+
"logits/rejected": -2.7446448802948,
|
| 332 |
+
"logps/chosen": -670.3836669921875,
|
| 333 |
+
"logps/rejected": -988.2939453125,
|
| 334 |
+
"loss": 0.2323,
|
| 335 |
+
"rewards/accuracies": 0.909375011920929,
|
| 336 |
+
"rewards/chosen": -1.8962217569351196,
|
| 337 |
+
"rewards/margins": 3.261091947555542,
|
| 338 |
+
"rewards/rejected": -5.157313823699951,
|
| 339 |
+
"step": 210
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.5823957643944407,
|
| 343 |
+
"grad_norm": 4.5229078992046166,
|
| 344 |
+
"learning_rate": 4.463830079489196e-07,
|
| 345 |
+
"logits/chosen": -2.7127437591552734,
|
| 346 |
+
"logits/rejected": -2.735060930252075,
|
| 347 |
+
"logps/chosen": -697.1866455078125,
|
| 348 |
+
"logps/rejected": -1054.2486572265625,
|
| 349 |
+
"loss": 0.2292,
|
| 350 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 351 |
+
"rewards/chosen": -2.145174026489258,
|
| 352 |
+
"rewards/margins": 3.7120985984802246,
|
| 353 |
+
"rewards/rejected": -5.857272148132324,
|
| 354 |
+
"step": 220
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.6088682991396426,
|
| 358 |
+
"grad_norm": 5.024605464390731,
|
| 359 |
+
"learning_rate": 4.390063398726356e-07,
|
| 360 |
+
"logits/chosen": -2.6947622299194336,
|
| 361 |
+
"logits/rejected": -2.6960158348083496,
|
| 362 |
+
"logps/chosen": -719.8890380859375,
|
| 363 |
+
"logps/rejected": -1086.1702880859375,
|
| 364 |
+
"loss": 0.2353,
|
| 365 |
+
"rewards/accuracies": 0.921875,
|
| 366 |
+
"rewards/chosen": -2.32169771194458,
|
| 367 |
+
"rewards/margins": 3.803462266921997,
|
| 368 |
+
"rewards/rejected": -6.12515926361084,
|
| 369 |
+
"step": 230
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.6353408338848445,
|
| 373 |
+
"grad_norm": 4.497134908678504,
|
| 374 |
+
"learning_rate": 4.3122394001821657e-07,
|
| 375 |
+
"logits/chosen": -2.699171781539917,
|
| 376 |
+
"logits/rejected": -2.7273011207580566,
|
| 377 |
+
"logps/chosen": -715.7581176757812,
|
| 378 |
+
"logps/rejected": -1008.9625244140625,
|
| 379 |
+
"loss": 0.2481,
|
| 380 |
+
"rewards/accuracies": 0.875,
|
| 381 |
+
"rewards/chosen": -2.174252986907959,
|
| 382 |
+
"rewards/margins": 3.0458567142486572,
|
| 383 |
+
"rewards/rejected": -5.220109462738037,
|
| 384 |
+
"step": 240
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.6618133686300464,
|
| 388 |
+
"grad_norm": 4.8761682045650625,
|
| 389 |
+
"learning_rate": 4.2305251452860566e-07,
|
| 390 |
+
"logits/chosen": -2.7068121433258057,
|
| 391 |
+
"logits/rejected": -2.7410686016082764,
|
| 392 |
+
"logps/chosen": -667.6207885742188,
|
| 393 |
+
"logps/rejected": -1025.73828125,
|
| 394 |
+
"loss": 0.2279,
|
| 395 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 396 |
+
"rewards/chosen": -1.91938054561615,
|
| 397 |
+
"rewards/margins": 3.5805881023406982,
|
| 398 |
+
"rewards/rejected": -5.4999680519104,
|
| 399 |
+
"step": 250
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.6882859033752482,
|
| 403 |
+
"grad_norm": 6.933116753661563,
|
| 404 |
+
"learning_rate": 4.1450960465134024e-07,
|
| 405 |
+
"logits/chosen": -2.7003915309906006,
|
| 406 |
+
"logits/rejected": -2.7421202659606934,
|
| 407 |
+
"logps/chosen": -725.2587890625,
|
| 408 |
+
"logps/rejected": -1111.21533203125,
|
| 409 |
+
"loss": 0.2322,
|
| 410 |
+
"rewards/accuracies": 0.90625,
|
| 411 |
+
"rewards/chosen": -2.4127094745635986,
|
| 412 |
+
"rewards/margins": 3.868704319000244,
|
| 413 |
+
"rewards/rejected": -6.28141450881958,
|
| 414 |
+
"step": 260
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.71475843812045,
|
| 418 |
+
"grad_norm": 5.180639131738146,
|
| 419 |
+
"learning_rate": 4.0561354908350977e-07,
|
| 420 |
+
"logits/chosen": -2.7455577850341797,
|
| 421 |
+
"logits/rejected": -2.768336057662964,
|
| 422 |
+
"logps/chosen": -686.0287475585938,
|
| 423 |
+
"logps/rejected": -1041.7474365234375,
|
| 424 |
+
"loss": 0.2406,
|
| 425 |
+
"rewards/accuracies": 0.925000011920929,
|
| 426 |
+
"rewards/chosen": -2.0675699710845947,
|
| 427 |
+
"rewards/margins": 3.5837669372558594,
|
| 428 |
+
"rewards/rejected": -5.651337146759033,
|
| 429 |
+
"step": 270
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.7412309728656519,
|
| 433 |
+
"grad_norm": 6.760828116437296,
|
| 434 |
+
"learning_rate": 3.963834446048644e-07,
|
| 435 |
+
"logits/chosen": -2.7979235649108887,
|
| 436 |
+
"logits/rejected": -2.795571804046631,
|
| 437 |
+
"logps/chosen": -757.2296142578125,
|
| 438 |
+
"logps/rejected": -1096.9935302734375,
|
| 439 |
+
"loss": 0.235,
|
| 440 |
+
"rewards/accuracies": 0.878125011920929,
|
| 441 |
+
"rewards/chosen": -2.414430618286133,
|
| 442 |
+
"rewards/margins": 3.682656764984131,
|
| 443 |
+
"rewards/rejected": -6.097087383270264,
|
| 444 |
+
"step": 280
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.7677035076108537,
|
| 448 |
+
"grad_norm": 5.924062579551431,
|
| 449 |
+
"learning_rate": 3.868391050835793e-07,
|
| 450 |
+
"logits/chosen": -2.707353353500366,
|
| 451 |
+
"logits/rejected": -2.708177089691162,
|
| 452 |
+
"logps/chosen": -735.2392578125,
|
| 453 |
+
"logps/rejected": -1096.3634033203125,
|
| 454 |
+
"loss": 0.2165,
|
| 455 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 456 |
+
"rewards/chosen": -2.4990901947021484,
|
| 457 |
+
"rewards/margins": 3.737868547439575,
|
| 458 |
+
"rewards/rejected": -6.2369585037231445,
|
| 459 |
+
"step": 290
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.7941760423560555,
|
| 463 |
+
"grad_norm": 6.316712123280126,
|
| 464 |
+
"learning_rate": 3.770010189426761e-07,
|
| 465 |
+
"logits/chosen": -2.696866512298584,
|
| 466 |
+
"logits/rejected": -2.720412015914917,
|
| 467 |
+
"logps/chosen": -788.0701904296875,
|
| 468 |
+
"logps/rejected": -1174.900390625,
|
| 469 |
+
"loss": 0.2278,
|
| 470 |
+
"rewards/accuracies": 0.893750011920929,
|
| 471 |
+
"rewards/chosen": -2.9900288581848145,
|
| 472 |
+
"rewards/margins": 3.7810866832733154,
|
| 473 |
+
"rewards/rejected": -6.771115779876709,
|
| 474 |
+
"step": 300
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.8206485771012575,
|
| 478 |
+
"grad_norm": 5.395016037418297,
|
| 479 |
+
"learning_rate": 3.66890305178407e-07,
|
| 480 |
+
"logits/chosen": -2.716305732727051,
|
| 481 |
+
"logits/rejected": -2.716601848602295,
|
| 482 |
+
"logps/chosen": -742.38232421875,
|
| 483 |
+
"logps/rejected": -1122.4625244140625,
|
| 484 |
+
"loss": 0.2287,
|
| 485 |
+
"rewards/accuracies": 0.90625,
|
| 486 |
+
"rewards/chosen": -2.601231098175049,
|
| 487 |
+
"rewards/margins": 3.891951084136963,
|
| 488 |
+
"rewards/rejected": -6.4931817054748535,
|
| 489 |
+
"step": 310
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.8471211118464593,
|
| 493 |
+
"grad_norm": 6.246149943149367,
|
| 494 |
+
"learning_rate": 3.565286680250138e-07,
|
| 495 |
+
"logits/chosen": -2.748223066329956,
|
| 496 |
+
"logits/rejected": -2.7793920040130615,
|
| 497 |
+
"logps/chosen": -724.46484375,
|
| 498 |
+
"logps/rejected": -1080.5015869140625,
|
| 499 |
+
"loss": 0.2089,
|
| 500 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 501 |
+
"rewards/chosen": -2.3395774364471436,
|
| 502 |
+
"rewards/margins": 3.757481813430786,
|
| 503 |
+
"rewards/rejected": -6.097059726715088,
|
| 504 |
+
"step": 320
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.8735936465916612,
|
| 508 |
+
"grad_norm": 6.876926943284203,
|
| 509 |
+
"learning_rate": 3.4593835036318225e-07,
|
| 510 |
+
"logits/chosen": -2.691286563873291,
|
| 511 |
+
"logits/rejected": -2.7156758308410645,
|
| 512 |
+
"logps/chosen": -759.7913208007812,
|
| 513 |
+
"logps/rejected": -1132.6741943359375,
|
| 514 |
+
"loss": 0.2212,
|
| 515 |
+
"rewards/accuracies": 0.921875,
|
| 516 |
+
"rewards/chosen": -2.764772415161133,
|
| 517 |
+
"rewards/margins": 3.8325703144073486,
|
| 518 |
+
"rewards/rejected": -6.597343444824219,
|
| 519 |
+
"step": 330
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.900066181336863,
|
| 523 |
+
"grad_norm": 7.266749393582284,
|
| 524 |
+
"learning_rate": 3.35142085972207e-07,
|
| 525 |
+
"logits/chosen": -2.6851227283477783,
|
| 526 |
+
"logits/rejected": -2.722938060760498,
|
| 527 |
+
"logps/chosen": -754.1365356445312,
|
| 528 |
+
"logps/rejected": -1132.1871337890625,
|
| 529 |
+
"loss": 0.218,
|
| 530 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 531 |
+
"rewards/chosen": -2.7608418464660645,
|
| 532 |
+
"rewards/margins": 3.8383007049560547,
|
| 533 |
+
"rewards/rejected": -6.599142551422119,
|
| 534 |
+
"step": 340
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.9265387160820648,
|
| 538 |
+
"grad_norm": 5.8828162546054426,
|
| 539 |
+
"learning_rate": 3.2416305072836555e-07,
|
| 540 |
+
"logits/chosen": -2.703392267227173,
|
| 541 |
+
"logits/rejected": -2.717411756515503,
|
| 542 |
+
"logps/chosen": -759.862060546875,
|
| 543 |
+
"logps/rejected": -1137.0098876953125,
|
| 544 |
+
"loss": 0.2033,
|
| 545 |
+
"rewards/accuracies": 0.909375011920929,
|
| 546 |
+
"rewards/chosen": -2.6198081970214844,
|
| 547 |
+
"rewards/margins": 3.8972651958465576,
|
| 548 |
+
"rewards/rejected": -6.517073154449463,
|
| 549 |
+
"step": 350
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.9530112508272667,
|
| 553 |
+
"grad_norm": 6.02673405904732,
|
| 554 |
+
"learning_rate": 3.1302481285426197e-07,
|
| 555 |
+
"logits/chosen": -2.7140469551086426,
|
| 556 |
+
"logits/rejected": -2.7299208641052246,
|
| 557 |
+
"logps/chosen": -727.9823608398438,
|
| 558 |
+
"logps/rejected": -1152.2041015625,
|
| 559 |
+
"loss": 0.2146,
|
| 560 |
+
"rewards/accuracies": 0.921875,
|
| 561 |
+
"rewards/chosen": -2.4466357231140137,
|
| 562 |
+
"rewards/margins": 4.253617763519287,
|
| 563 |
+
"rewards/rejected": -6.700253486633301,
|
| 564 |
+
"step": 360
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.9794837855724685,
|
| 568 |
+
"grad_norm": 5.891168724337083,
|
| 569 |
+
"learning_rate": 3.017512823259373e-07,
|
| 570 |
+
"logits/chosen": -2.7166543006896973,
|
| 571 |
+
"logits/rejected": -2.7297706604003906,
|
| 572 |
+
"logps/chosen": -730.9952392578125,
|
| 573 |
+
"logps/rejected": -1117.7025146484375,
|
| 574 |
+
"loss": 0.2011,
|
| 575 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 576 |
+
"rewards/chosen": -2.5781686305999756,
|
| 577 |
+
"rewards/margins": 4.00323486328125,
|
| 578 |
+
"rewards/rejected": -6.5814032554626465,
|
| 579 |
+
"step": 370
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 1.0059563203176705,
|
| 583 |
+
"grad_norm": 5.4096108526897115,
|
| 584 |
+
"learning_rate": 2.9036665954635264e-07,
|
| 585 |
+
"logits/chosen": -2.7353827953338623,
|
| 586 |
+
"logits/rejected": -2.765237331390381,
|
| 587 |
+
"logps/chosen": -769.5081176757812,
|
| 588 |
+
"logps/rejected": -1209.970947265625,
|
| 589 |
+
"loss": 0.2135,
|
| 590 |
+
"rewards/accuracies": 0.940625011920929,
|
| 591 |
+
"rewards/chosen": -2.7778820991516113,
|
| 592 |
+
"rewards/margins": 4.453129291534424,
|
| 593 |
+
"rewards/rejected": -7.231011390686035,
|
| 594 |
+
"step": 380
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.0324288550628722,
|
| 598 |
+
"grad_norm": 6.969996057859779,
|
| 599 |
+
"learning_rate": 2.7889538339542523e-07,
|
| 600 |
+
"logits/chosen": -2.715958833694458,
|
| 601 |
+
"logits/rejected": -2.741973638534546,
|
| 602 |
+
"logps/chosen": -760.3748779296875,
|
| 603 |
+
"logps/rejected": -1181.979248046875,
|
| 604 |
+
"loss": 0.1997,
|
| 605 |
+
"rewards/accuracies": 0.940625011920929,
|
| 606 |
+
"rewards/chosen": -2.8078932762145996,
|
| 607 |
+
"rewards/margins": 4.318324089050293,
|
| 608 |
+
"rewards/rejected": -7.126217842102051,
|
| 609 |
+
"step": 390
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.0589013898080741,
|
| 613 |
+
"grad_norm": 6.27267698414733,
|
| 614 |
+
"learning_rate": 2.6736207876813643e-07,
|
| 615 |
+
"logits/chosen": -2.728332042694092,
|
| 616 |
+
"logits/rejected": -2.7608368396759033,
|
| 617 |
+
"logps/chosen": -768.7643432617188,
|
| 618 |
+
"logps/rejected": -1208.543701171875,
|
| 619 |
+
"loss": 0.1963,
|
| 620 |
+
"rewards/accuracies": 0.909375011920929,
|
| 621 |
+
"rewards/chosen": -2.7519562244415283,
|
| 622 |
+
"rewards/margins": 4.398838996887207,
|
| 623 |
+
"rewards/rejected": -7.150795936584473,
|
| 624 |
+
"step": 400
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 1.0853739245532759,
|
| 628 |
+
"grad_norm": 6.331116230324871,
|
| 629 |
+
"learning_rate": 2.5579150371332953e-07,
|
| 630 |
+
"logits/chosen": -2.7396435737609863,
|
| 631 |
+
"logits/rejected": -2.7421391010284424,
|
| 632 |
+
"logps/chosen": -809.5894775390625,
|
| 633 |
+
"logps/rejected": -1248.327880859375,
|
| 634 |
+
"loss": 0.1821,
|
| 635 |
+
"rewards/accuracies": 0.940625011920929,
|
| 636 |
+
"rewards/chosen": -3.0432112216949463,
|
| 637 |
+
"rewards/margins": 4.567535400390625,
|
| 638 |
+
"rewards/rejected": -7.61074686050415,
|
| 639 |
+
"step": 410
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 1.1118464592984778,
|
| 643 |
+
"grad_norm": 6.507894575900424,
|
| 644 |
+
"learning_rate": 2.4420849628667045e-07,
|
| 645 |
+
"logits/chosen": -2.7221198081970215,
|
| 646 |
+
"logits/rejected": -2.7356066703796387,
|
| 647 |
+
"logps/chosen": -795.4509887695312,
|
| 648 |
+
"logps/rejected": -1229.848876953125,
|
| 649 |
+
"loss": 0.205,
|
| 650 |
+
"rewards/accuracies": 0.940625011920929,
|
| 651 |
+
"rewards/chosen": -3.027529716491699,
|
| 652 |
+
"rewards/margins": 4.336198806762695,
|
| 653 |
+
"rewards/rejected": -7.3637285232543945,
|
| 654 |
+
"step": 420
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.1383189940436798,
|
| 658 |
+
"grad_norm": 6.866126936883296,
|
| 659 |
+
"learning_rate": 2.3263792123186352e-07,
|
| 660 |
+
"logits/chosen": -2.7347493171691895,
|
| 661 |
+
"logits/rejected": -2.7523412704467773,
|
| 662 |
+
"logps/chosen": -788.9684448242188,
|
| 663 |
+
"logps/rejected": -1187.0565185546875,
|
| 664 |
+
"loss": 0.2132,
|
| 665 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 666 |
+
"rewards/chosen": -2.9675726890563965,
|
| 667 |
+
"rewards/margins": 4.024303913116455,
|
| 668 |
+
"rewards/rejected": -6.99187707901001,
|
| 669 |
+
"step": 430
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 1.1647915287888815,
|
| 673 |
+
"grad_norm": 5.664097996574406,
|
| 674 |
+
"learning_rate": 2.211046166045748e-07,
|
| 675 |
+
"logits/chosen": -2.699467420578003,
|
| 676 |
+
"logits/rejected": -2.715172529220581,
|
| 677 |
+
"logps/chosen": -787.1725463867188,
|
| 678 |
+
"logps/rejected": -1219.5947265625,
|
| 679 |
+
"loss": 0.1992,
|
| 680 |
+
"rewards/accuracies": 0.909375011920929,
|
| 681 |
+
"rewards/chosen": -2.9378509521484375,
|
| 682 |
+
"rewards/margins": 4.4215216636657715,
|
| 683 |
+
"rewards/rejected": -7.359372138977051,
|
| 684 |
+
"step": 440
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 1.1912640635340834,
|
| 688 |
+
"grad_norm": 5.8438221376315,
|
| 689 |
+
"learning_rate": 2.096333404536474e-07,
|
| 690 |
+
"logits/chosen": -2.6697657108306885,
|
| 691 |
+
"logits/rejected": -2.6788392066955566,
|
| 692 |
+
"logps/chosen": -816.9085693359375,
|
| 693 |
+
"logps/rejected": -1204.7374267578125,
|
| 694 |
+
"loss": 0.1939,
|
| 695 |
+
"rewards/accuracies": 0.925000011920929,
|
| 696 |
+
"rewards/chosen": -3.2424380779266357,
|
| 697 |
+
"rewards/margins": 4.097726821899414,
|
| 698 |
+
"rewards/rejected": -7.340165138244629,
|
| 699 |
+
"step": 450
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.2177365982792852,
|
| 703 |
+
"grad_norm": 6.534950899522826,
|
| 704 |
+
"learning_rate": 1.982487176740627e-07,
|
| 705 |
+
"logits/chosen": -2.6479620933532715,
|
| 706 |
+
"logits/rejected": -2.686981201171875,
|
| 707 |
+
"logps/chosen": -760.5189208984375,
|
| 708 |
+
"logps/rejected": -1172.6590576171875,
|
| 709 |
+
"loss": 0.2034,
|
| 710 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 711 |
+
"rewards/chosen": -2.8477563858032227,
|
| 712 |
+
"rewards/margins": 4.02510929107666,
|
| 713 |
+
"rewards/rejected": -6.872865200042725,
|
| 714 |
+
"step": 460
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 1.244209133024487,
|
| 718 |
+
"grad_norm": 5.919455336024619,
|
| 719 |
+
"learning_rate": 1.8697518714573804e-07,
|
| 720 |
+
"logits/chosen": -2.7183327674865723,
|
| 721 |
+
"logits/rejected": -2.7170839309692383,
|
| 722 |
+
"logps/chosen": -790.9767456054688,
|
| 723 |
+
"logps/rejected": -1145.7747802734375,
|
| 724 |
+
"loss": 0.1854,
|
| 725 |
+
"rewards/accuracies": 0.887499988079071,
|
| 726 |
+
"rewards/chosen": -2.942429304122925,
|
| 727 |
+
"rewards/margins": 3.9219982624053955,
|
| 728 |
+
"rewards/rejected": -6.864427089691162,
|
| 729 |
+
"step": 470
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 1.270681667769689,
|
| 733 |
+
"grad_norm": 8.510303191565324,
|
| 734 |
+
"learning_rate": 1.758369492716345e-07,
|
| 735 |
+
"logits/chosen": -2.7107906341552734,
|
| 736 |
+
"logits/rejected": -2.7333004474639893,
|
| 737 |
+
"logps/chosen": -831.3541870117188,
|
| 738 |
+
"logps/rejected": -1250.9085693359375,
|
| 739 |
+
"loss": 0.204,
|
| 740 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 741 |
+
"rewards/chosen": -3.3519797325134277,
|
| 742 |
+
"rewards/margins": 4.182629585266113,
|
| 743 |
+
"rewards/rejected": -7.534609317779541,
|
| 744 |
+
"step": 480
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 1.2971542025148908,
|
| 748 |
+
"grad_norm": 6.158432969646956,
|
| 749 |
+
"learning_rate": 1.648579140277931e-07,
|
| 750 |
+
"logits/chosen": -2.7101292610168457,
|
| 751 |
+
"logits/rejected": -2.715567111968994,
|
| 752 |
+
"logps/chosen": -815.2239990234375,
|
| 753 |
+
"logps/rejected": -1204.00830078125,
|
| 754 |
+
"loss": 0.2016,
|
| 755 |
+
"rewards/accuracies": 0.9375,
|
| 756 |
+
"rewards/chosen": -3.184443712234497,
|
| 757 |
+
"rewards/margins": 4.096536159515381,
|
| 758 |
+
"rewards/rejected": -7.280980110168457,
|
| 759 |
+
"step": 490
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 1.3236267372600927,
|
| 763 |
+
"grad_norm": 6.688172890404941,
|
| 764 |
+
"learning_rate": 1.5406164963681773e-07,
|
| 765 |
+
"logits/chosen": -2.6899495124816895,
|
| 766 |
+
"logits/rejected": -2.697331666946411,
|
| 767 |
+
"logps/chosen": -788.6990356445312,
|
| 768 |
+
"logps/rejected": -1184.682861328125,
|
| 769 |
+
"loss": 0.2028,
|
| 770 |
+
"rewards/accuracies": 0.918749988079071,
|
| 771 |
+
"rewards/chosen": -2.907625436782837,
|
| 772 |
+
"rewards/margins": 4.130288124084473,
|
| 773 |
+
"rewards/rejected": -7.037914276123047,
|
| 774 |
+
"step": 500
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 1.3500992720052944,
|
| 778 |
+
"grad_norm": 6.177665969794554,
|
| 779 |
+
"learning_rate": 1.4347133197498618e-07,
|
| 780 |
+
"logits/chosen": -2.720623016357422,
|
| 781 |
+
"logits/rejected": -2.713228702545166,
|
| 782 |
+
"logps/chosen": -813.38427734375,
|
| 783 |
+
"logps/rejected": -1158.912353515625,
|
| 784 |
+
"loss": 0.1947,
|
| 785 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 786 |
+
"rewards/chosen": -3.203073024749756,
|
| 787 |
+
"rewards/margins": 3.7714266777038574,
|
| 788 |
+
"rewards/rejected": -6.974499702453613,
|
| 789 |
+
"step": 510
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 1.3765718067504964,
|
| 793 |
+
"grad_norm": 7.6140239843164785,
|
| 794 |
+
"learning_rate": 1.3310969482159296e-07,
|
| 795 |
+
"logits/chosen": -2.7323203086853027,
|
| 796 |
+
"logits/rejected": -2.720503330230713,
|
| 797 |
+
"logps/chosen": -822.5255737304688,
|
| 798 |
+
"logps/rejected": -1203.3837890625,
|
| 799 |
+
"loss": 0.206,
|
| 800 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 801 |
+
"rewards/chosen": -3.2711524963378906,
|
| 802 |
+
"rewards/margins": 4.040999889373779,
|
| 803 |
+
"rewards/rejected": -7.312152862548828,
|
| 804 |
+
"step": 520
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 1.4030443414956983,
|
| 808 |
+
"grad_norm": 7.395646356776199,
|
| 809 |
+
"learning_rate": 1.2299898105732384e-07,
|
| 810 |
+
"logits/chosen": -2.664844036102295,
|
| 811 |
+
"logits/rejected": -2.6827051639556885,
|
| 812 |
+
"logps/chosen": -797.8216552734375,
|
| 813 |
+
"logps/rejected": -1175.84326171875,
|
| 814 |
+
"loss": 0.1975,
|
| 815 |
+
"rewards/accuracies": 0.90625,
|
| 816 |
+
"rewards/chosen": -3.197530746459961,
|
| 817 |
+
"rewards/margins": 3.875856876373291,
|
| 818 |
+
"rewards/rejected": -7.07338809967041,
|
| 819 |
+
"step": 530
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 1.4295168762409,
|
| 823 |
+
"grad_norm": 5.37023800702416,
|
| 824 |
+
"learning_rate": 1.1316089491642075e-07,
|
| 825 |
+
"logits/chosen": -2.6801838874816895,
|
| 826 |
+
"logits/rejected": -2.7150120735168457,
|
| 827 |
+
"logps/chosen": -782.3211669921875,
|
| 828 |
+
"logps/rejected": -1218.0201416015625,
|
| 829 |
+
"loss": 0.177,
|
| 830 |
+
"rewards/accuracies": 0.953125,
|
| 831 |
+
"rewards/chosen": -3.0446789264678955,
|
| 832 |
+
"rewards/margins": 4.468942642211914,
|
| 833 |
+
"rewards/rejected": -7.5136213302612305,
|
| 834 |
+
"step": 540
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 1.4559894109861018,
|
| 838 |
+
"grad_norm": 6.848159521074984,
|
| 839 |
+
"learning_rate": 1.0361655539513564e-07,
|
| 840 |
+
"logits/chosen": -2.7019715309143066,
|
| 841 |
+
"logits/rejected": -2.717634677886963,
|
| 842 |
+
"logps/chosen": -816.9815063476562,
|
| 843 |
+
"logps/rejected": -1239.1295166015625,
|
| 844 |
+
"loss": 0.1866,
|
| 845 |
+
"rewards/accuracies": 0.9375,
|
| 846 |
+
"rewards/chosen": -3.2783799171447754,
|
| 847 |
+
"rewards/margins": 4.237971305847168,
|
| 848 |
+
"rewards/rejected": -7.516351222991943,
|
| 849 |
+
"step": 550
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 1.4824619457313037,
|
| 853 |
+
"grad_norm": 6.1903290336235965,
|
| 854 |
+
"learning_rate": 9.438645091649028e-08,
|
| 855 |
+
"logits/chosen": -2.703291416168213,
|
| 856 |
+
"logits/rejected": -2.717991590499878,
|
| 857 |
+
"logps/chosen": -835.4637451171875,
|
| 858 |
+
"logps/rejected": -1262.481201171875,
|
| 859 |
+
"loss": 0.1941,
|
| 860 |
+
"rewards/accuracies": 0.909375011920929,
|
| 861 |
+
"rewards/chosen": -3.5229759216308594,
|
| 862 |
+
"rewards/margins": 4.230559349060059,
|
| 863 |
+
"rewards/rejected": -7.753535270690918,
|
| 864 |
+
"step": 560
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 1.5089344804765057,
|
| 868 |
+
"grad_norm": 7.319396362423722,
|
| 869 |
+
"learning_rate": 8.549039534865979e-08,
|
| 870 |
+
"logits/chosen": -2.6856772899627686,
|
| 871 |
+
"logits/rejected": -2.7186999320983887,
|
| 872 |
+
"logps/chosen": -839.0086059570312,
|
| 873 |
+
"logps/rejected": -1260.779541015625,
|
| 874 |
+
"loss": 0.1939,
|
| 875 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 876 |
+
"rewards/chosen": -3.4230124950408936,
|
| 877 |
+
"rewards/margins": 4.303016185760498,
|
| 878 |
+
"rewards/rejected": -7.7260284423828125,
|
| 879 |
+
"step": 570
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 1.5354070152217076,
|
| 883 |
+
"grad_norm": 6.830716936510353,
|
| 884 |
+
"learning_rate": 7.694748547139429e-08,
|
| 885 |
+
"logits/chosen": -2.690796136856079,
|
| 886 |
+
"logits/rejected": -2.6854541301727295,
|
| 887 |
+
"logps/chosen": -858.9547119140625,
|
| 888 |
+
"logps/rejected": -1268.125732421875,
|
| 889 |
+
"loss": 0.1822,
|
| 890 |
+
"rewards/accuracies": 0.925000011920929,
|
| 891 |
+
"rewards/chosen": -3.606029510498047,
|
| 892 |
+
"rewards/margins": 4.35842752456665,
|
| 893 |
+
"rewards/rejected": -7.964457035064697,
|
| 894 |
+
"step": 580
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 1.5618795499669094,
|
| 898 |
+
"grad_norm": 7.344475965596424,
|
| 899 |
+
"learning_rate": 6.877605998178343e-08,
|
| 900 |
+
"logits/chosen": -2.6852376461029053,
|
| 901 |
+
"logits/rejected": -2.6851906776428223,
|
| 902 |
+
"logps/chosen": -825.2852783203125,
|
| 903 |
+
"logps/rejected": -1217.6024169921875,
|
| 904 |
+
"loss": 0.1937,
|
| 905 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 906 |
+
"rewards/chosen": -3.3112761974334717,
|
| 907 |
+
"rewards/margins": 4.0954132080078125,
|
| 908 |
+
"rewards/rejected": -7.4066901206970215,
|
| 909 |
+
"step": 590
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 1.588352084712111,
|
| 913 |
+
"grad_norm": 6.774677658543177,
|
| 914 |
+
"learning_rate": 6.099366012736437e-08,
|
| 915 |
+
"logits/chosen": -2.6621475219726562,
|
| 916 |
+
"logits/rejected": -2.6802525520324707,
|
| 917 |
+
"logps/chosen": -823.2789306640625,
|
| 918 |
+
"logps/rejected": -1230.999755859375,
|
| 919 |
+
"loss": 0.1879,
|
| 920 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 921 |
+
"rewards/chosen": -3.3858489990234375,
|
| 922 |
+
"rewards/margins": 4.153179168701172,
|
| 923 |
+
"rewards/rejected": -7.539028167724609,
|
| 924 |
+
"step": 600
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 1.614824619457313,
|
| 928 |
+
"grad_norm": 6.63238227242087,
|
| 929 |
+
"learning_rate": 5.3616992051080415e-08,
|
| 930 |
+
"logits/chosen": -2.677306890487671,
|
| 931 |
+
"logits/rejected": -2.712442636489868,
|
| 932 |
+
"logps/chosen": -788.0901489257812,
|
| 933 |
+
"logps/rejected": -1236.086181640625,
|
| 934 |
+
"loss": 0.1818,
|
| 935 |
+
"rewards/accuracies": 0.925000011920929,
|
| 936 |
+
"rewards/chosen": -3.1388726234436035,
|
| 937 |
+
"rewards/margins": 4.399405002593994,
|
| 938 |
+
"rewards/rejected": -7.538276672363281,
|
| 939 |
+
"step": 610
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 1.641297154202515,
|
| 943 |
+
"grad_norm": 7.377420865927689,
|
| 944 |
+
"learning_rate": 4.666189092892245e-08,
|
| 945 |
+
"logits/chosen": -2.681398630142212,
|
| 946 |
+
"logits/rejected": -2.7074177265167236,
|
| 947 |
+
"logps/chosen": -808.8873291015625,
|
| 948 |
+
"logps/rejected": -1222.2945556640625,
|
| 949 |
+
"loss": 0.192,
|
| 950 |
+
"rewards/accuracies": 0.90625,
|
| 951 |
+
"rewards/chosen": -3.247420072555542,
|
| 952 |
+
"rewards/margins": 4.086552619934082,
|
| 953 |
+
"rewards/rejected": -7.3339738845825195,
|
| 954 |
+
"step": 620
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 1.6677696889477167,
|
| 958 |
+
"grad_norm": 8.778568268489295,
|
| 959 |
+
"learning_rate": 4.0143286977238345e-08,
|
| 960 |
+
"logits/chosen": -2.732050895690918,
|
| 961 |
+
"logits/rejected": -2.743983507156372,
|
| 962 |
+
"logps/chosen": -833.8406372070312,
|
| 963 |
+
"logps/rejected": -1255.4339599609375,
|
| 964 |
+
"loss": 0.2052,
|
| 965 |
+
"rewards/accuracies": 0.909375011920929,
|
| 966 |
+
"rewards/chosen": -3.412902355194092,
|
| 967 |
+
"rewards/margins": 4.319065570831299,
|
| 968 |
+
"rewards/rejected": -7.731966972351074,
|
| 969 |
+
"step": 630
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 1.6942422236929184,
|
| 973 |
+
"grad_norm": 9.757233734434815,
|
| 974 |
+
"learning_rate": 3.407517340267957e-08,
|
| 975 |
+
"logits/chosen": -2.704099655151367,
|
| 976 |
+
"logits/rejected": -2.747129440307617,
|
| 977 |
+
"logps/chosen": -825.1597900390625,
|
| 978 |
+
"logps/rejected": -1279.721923828125,
|
| 979 |
+
"loss": 0.1969,
|
| 980 |
+
"rewards/accuracies": 0.921875,
|
| 981 |
+
"rewards/chosen": -3.3202297687530518,
|
| 982 |
+
"rewards/margins": 4.6456804275512695,
|
| 983 |
+
"rewards/rejected": -7.965909481048584,
|
| 984 |
+
"step": 640
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 1.7207147584381204,
|
| 988 |
+
"grad_norm": 5.790336419749713,
|
| 989 |
+
"learning_rate": 2.847057636358663e-08,
|
| 990 |
+
"logits/chosen": -2.666538715362549,
|
| 991 |
+
"logits/rejected": -2.6833436489105225,
|
| 992 |
+
"logps/chosen": -791.0831298828125,
|
| 993 |
+
"logps/rejected": -1239.421875,
|
| 994 |
+
"loss": 0.2025,
|
| 995 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 996 |
+
"rewards/chosen": -3.060492753982544,
|
| 997 |
+
"rewards/margins": 4.618899345397949,
|
| 998 |
+
"rewards/rejected": -7.679392337799072,
|
| 999 |
+
"step": 650
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 1.7471872931833223,
|
| 1003 |
+
"grad_norm": 7.090658258583613,
|
| 1004 |
+
"learning_rate": 2.3341527007295107e-08,
|
| 1005 |
+
"logits/chosen": -2.724691867828369,
|
| 1006 |
+
"logits/rejected": -2.7381081581115723,
|
| 1007 |
+
"logps/chosen": -795.6064453125,
|
| 1008 |
+
"logps/rejected": -1221.3892822265625,
|
| 1009 |
+
"loss": 0.1934,
|
| 1010 |
+
"rewards/accuracies": 0.918749988079071,
|
| 1011 |
+
"rewards/chosen": -3.000559091567993,
|
| 1012 |
+
"rewards/margins": 4.30582332611084,
|
| 1013 |
+
"rewards/rejected": -7.306382656097412,
|
| 1014 |
+
"step": 660
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 1.7736598279285243,
|
| 1018 |
+
"grad_norm": 7.043130947329383,
|
| 1019 |
+
"learning_rate": 1.8699035643389927e-08,
|
| 1020 |
+
"logits/chosen": -2.6909706592559814,
|
| 1021 |
+
"logits/rejected": -2.706798791885376,
|
| 1022 |
+
"logps/chosen": -813.7172241210938,
|
| 1023 |
+
"logps/rejected": -1210.0374755859375,
|
| 1024 |
+
"loss": 0.1935,
|
| 1025 |
+
"rewards/accuracies": 0.9375,
|
| 1026 |
+
"rewards/chosen": -3.150881290435791,
|
| 1027 |
+
"rewards/margins": 4.167717933654785,
|
| 1028 |
+
"rewards/rejected": -7.318598747253418,
|
| 1029 |
+
"step": 670
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 1.800132362673726,
|
| 1033 |
+
"grad_norm": 8.93025020392657,
|
| 1034 |
+
"learning_rate": 1.4553068108346778e-08,
|
| 1035 |
+
"logits/chosen": -2.7232556343078613,
|
| 1036 |
+
"logits/rejected": -2.758570671081543,
|
| 1037 |
+
"logps/chosen": -816.0042724609375,
|
| 1038 |
+
"logps/rejected": -1228.6907958984375,
|
| 1039 |
+
"loss": 0.1969,
|
| 1040 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 1041 |
+
"rewards/chosen": -3.222668409347534,
|
| 1042 |
+
"rewards/margins": 4.130661487579346,
|
| 1043 |
+
"rewards/rejected": -7.353329658508301,
|
| 1044 |
+
"step": 680
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 1.8266048974189277,
|
| 1048 |
+
"grad_norm": 8.105462657724475,
|
| 1049 |
+
"learning_rate": 1.0912524372300031e-08,
|
| 1050 |
+
"logits/chosen": -2.6958508491516113,
|
| 1051 |
+
"logits/rejected": -2.7145791053771973,
|
| 1052 |
+
"logps/chosen": -800.2642211914062,
|
| 1053 |
+
"logps/rejected": -1204.597412109375,
|
| 1054 |
+
"loss": 0.1979,
|
| 1055 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 1056 |
+
"rewards/chosen": -3.0551960468292236,
|
| 1057 |
+
"rewards/margins": 4.35771369934082,
|
| 1058 |
+
"rewards/rejected": -7.412909507751465,
|
| 1059 |
+
"step": 690
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 1.8530774321641297,
|
| 1063 |
+
"grad_norm": 6.305134676239398,
|
| 1064 |
+
"learning_rate": 7.785219433859846e-09,
|
| 1065 |
+
"logits/chosen": -2.666156768798828,
|
| 1066 |
+
"logits/rejected": -2.6787142753601074,
|
| 1067 |
+
"logps/chosen": -795.4716186523438,
|
| 1068 |
+
"logps/rejected": -1216.070068359375,
|
| 1069 |
+
"loss": 0.1889,
|
| 1070 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 1071 |
+
"rewards/chosen": -3.1992480754852295,
|
| 1072 |
+
"rewards/margins": 4.258774757385254,
|
| 1073 |
+
"rewards/rejected": -7.4580230712890625,
|
| 1074 |
+
"step": 700
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 1.8795499669093316,
|
| 1078 |
+
"grad_norm": 8.35714287832233,
|
| 1079 |
+
"learning_rate": 5.177866543990689e-09,
|
| 1080 |
+
"logits/chosen": -2.6713449954986572,
|
| 1081 |
+
"logits/rejected": -2.7099664211273193,
|
| 1082 |
+
"logps/chosen": -809.7697143554688,
|
| 1083 |
+
"logps/rejected": -1260.0615234375,
|
| 1084 |
+
"loss": 0.1768,
|
| 1085 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1086 |
+
"rewards/chosen": -3.31097412109375,
|
| 1087 |
+
"rewards/margins": 4.475188255310059,
|
| 1088 |
+
"rewards/rejected": -7.786163330078125,
|
| 1089 |
+
"step": 710
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 1.9060225016545336,
|
| 1093 |
+
"grad_norm": 6.357842566297534,
|
| 1094 |
+
"learning_rate": 3.0960627949644105e-09,
|
| 1095 |
+
"logits/chosen": -2.661090850830078,
|
| 1096 |
+
"logits/rejected": -2.6854565143585205,
|
| 1097 |
+
"logps/chosen": -799.8544921875,
|
| 1098 |
+
"logps/rejected": -1197.60595703125,
|
| 1099 |
+
"loss": 0.195,
|
| 1100 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1101 |
+
"rewards/chosen": -3.179225206375122,
|
| 1102 |
+
"rewards/margins": 4.040897846221924,
|
| 1103 |
+
"rewards/rejected": -7.220122337341309,
|
| 1104 |
+
"step": 720
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 1.9324950363997353,
|
| 1108 |
+
"grad_norm": 6.897936781273812,
|
| 1109 |
+
"learning_rate": 1.5442771053230663e-09,
|
| 1110 |
+
"logits/chosen": -2.654930353164673,
|
| 1111 |
+
"logits/rejected": -2.6900641918182373,
|
| 1112 |
+
"logps/chosen": -794.2493286132812,
|
| 1113 |
+
"logps/rejected": -1218.8597412109375,
|
| 1114 |
+
"loss": 0.1752,
|
| 1115 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1116 |
+
"rewards/chosen": -3.182406187057495,
|
| 1117 |
+
"rewards/margins": 4.386531352996826,
|
| 1118 |
+
"rewards/rejected": -7.5689377784729,
|
| 1119 |
+
"step": 730
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 1.958967571144937,
|
| 1123 |
+
"grad_norm": 6.958740857504488,
|
| 1124 |
+
"learning_rate": 5.25840626643681e-10,
|
| 1125 |
+
"logits/chosen": -2.731703519821167,
|
| 1126 |
+
"logits/rejected": -2.728083848953247,
|
| 1127 |
+
"logps/chosen": -804.791259765625,
|
| 1128 |
+
"logps/rejected": -1165.920654296875,
|
| 1129 |
+
"loss": 0.1871,
|
| 1130 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 1131 |
+
"rewards/chosen": -3.141540765762329,
|
| 1132 |
+
"rewards/margins": 3.8939075469970703,
|
| 1133 |
+
"rewards/rejected": -7.035449028015137,
|
| 1134 |
+
"step": 740
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 1.985440105890139,
|
| 1138 |
+
"grad_norm": 7.712958898241185,
|
| 1139 |
+
"learning_rate": 4.293959269863201e-11,
|
| 1140 |
+
"logits/chosen": -2.6934590339660645,
|
| 1141 |
+
"logits/rejected": -2.702648639678955,
|
| 1142 |
+
"logps/chosen": -818.2516479492188,
|
| 1143 |
+
"logps/rejected": -1221.8408203125,
|
| 1144 |
+
"loss": 0.2042,
|
| 1145 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1146 |
+
"rewards/chosen": -3.207921266555786,
|
| 1147 |
+
"rewards/margins": 4.332727909088135,
|
| 1148 |
+
"rewards/rejected": -7.5406494140625,
|
| 1149 |
+
"step": 750
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 1.9960291197882198,
|
| 1153 |
+
"step": 754,
|
| 1154 |
+
"total_flos": 0.0,
|
| 1155 |
+
"train_loss": 0.26563876598520053,
|
| 1156 |
+
"train_runtime": 5267.2293,
|
| 1157 |
+
"train_samples_per_second": 36.699,
|
| 1158 |
+
"train_steps_per_second": 0.143
|
| 1159 |
+
}
|
| 1160 |
+
],
|
| 1161 |
+
"logging_steps": 10,
|
| 1162 |
+
"max_steps": 754,
|
| 1163 |
+
"num_input_tokens_seen": 0,
|
| 1164 |
+
"num_train_epochs": 2,
|
| 1165 |
+
"save_steps": 100,
|
| 1166 |
+
"stateful_callbacks": {
|
| 1167 |
+
"TrainerControl": {
|
| 1168 |
+
"args": {
|
| 1169 |
+
"should_epoch_stop": false,
|
| 1170 |
+
"should_evaluate": false,
|
| 1171 |
+
"should_log": false,
|
| 1172 |
+
"should_save": true,
|
| 1173 |
+
"should_training_stop": true
|
| 1174 |
+
},
|
| 1175 |
+
"attributes": {}
|
| 1176 |
+
}
|
| 1177 |
+
},
|
| 1178 |
+
"total_flos": 0.0,
|
| 1179 |
+
"train_batch_size": 8,
|
| 1180 |
+
"trial_name": null,
|
| 1181 |
+
"trial_params": null
|
| 1182 |
+
}
|