Model save
Browse files- README.md +74 -0
- all_results.json +9 -0
- generation_config.json +7 -0
- train_results.json +9 -0
- trainer_state.json +2518 -0
README.md
ADDED
|
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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-math-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-math-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 the None dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.3224
|
| 22 |
+
- Rewards/chosen: -3.3080
|
| 23 |
+
- Rewards/rejected: -7.0491
|
| 24 |
+
- Rewards/accuracies: 0.9062
|
| 25 |
+
- Rewards/margins: 3.7411
|
| 26 |
+
- Logps/rejected: -1048.2871
|
| 27 |
+
- Logps/chosen: -689.9403
|
| 28 |
+
- Logits/rejected: -1.4398
|
| 29 |
+
- Logits/chosen: -1.4575
|
| 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: 5e-07
|
| 49 |
+
- train_batch_size: 8
|
| 50 |
+
- eval_batch_size: 8
|
| 51 |
+
- seed: 42
|
| 52 |
+
- distributed_type: multi-GPU
|
| 53 |
+
- num_devices: 8
|
| 54 |
+
- gradient_accumulation_steps: 4
|
| 55 |
+
- total_train_batch_size: 256
|
| 56 |
+
- total_eval_batch_size: 64
|
| 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: 2
|
| 61 |
+
|
| 62 |
+
### Training results
|
| 63 |
+
|
| 64 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 65 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 66 |
+
| 0.2293 | 1.2210 | 1000 | 0.3224 | -3.3080 | -7.0491 | 0.9062 | 3.7411 | -1048.2871 | -689.9403 | -1.4398 | -1.4575 |
|
| 67 |
+
|
| 68 |
+
|
| 69 |
+
### Framework versions
|
| 70 |
+
|
| 71 |
+
- Transformers 4.44.1
|
| 72 |
+
- Pytorch 2.1.2+cu121
|
| 73 |
+
- Datasets 2.21.0
|
| 74 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.2817063624168927,
|
| 5 |
+
"train_runtime": 11455.6006,
|
| 6 |
+
"train_samples": 209650,
|
| 7 |
+
"train_samples_per_second": 36.602,
|
| 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": 2.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.2817063624168927,
|
| 5 |
+
"train_runtime": 11455.6006,
|
| 6 |
+
"train_samples": 209650,
|
| 7 |
+
"train_samples_per_second": 36.602,
|
| 8 |
+
"train_steps_per_second": 0.143
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2518 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 1638,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.001221001221001221,
|
| 13 |
+
"grad_norm": 3.104700246084015,
|
| 14 |
+
"learning_rate": 3.048780487804878e-09,
|
| 15 |
+
"logits/chosen": -2.611332893371582,
|
| 16 |
+
"logits/rejected": -2.6034297943115234,
|
| 17 |
+
"logps/chosen": -424.76251220703125,
|
| 18 |
+
"logps/rejected": -401.40936279296875,
|
| 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": 0.01221001221001221,
|
| 28 |
+
"grad_norm": 3.069950647256397,
|
| 29 |
+
"learning_rate": 3.048780487804878e-08,
|
| 30 |
+
"logits/chosen": -2.483457326889038,
|
| 31 |
+
"logits/rejected": -2.5166726112365723,
|
| 32 |
+
"logps/chosen": -394.9632873535156,
|
| 33 |
+
"logps/rejected": -370.91741943359375,
|
| 34 |
+
"loss": 0.6932,
|
| 35 |
+
"rewards/accuracies": 0.4444444477558136,
|
| 36 |
+
"rewards/chosen": -0.00035367117379792035,
|
| 37 |
+
"rewards/margins": -0.00030399090610444546,
|
| 38 |
+
"rewards/rejected": -4.968021312379278e-05,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.02442002442002442,
|
| 43 |
+
"grad_norm": 3.064473375199238,
|
| 44 |
+
"learning_rate": 6.097560975609756e-08,
|
| 45 |
+
"logits/chosen": -2.4844443798065186,
|
| 46 |
+
"logits/rejected": -2.5088613033294678,
|
| 47 |
+
"logps/chosen": -393.47589111328125,
|
| 48 |
+
"logps/rejected": -384.7496337890625,
|
| 49 |
+
"loss": 0.6931,
|
| 50 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 51 |
+
"rewards/chosen": -0.0005507160676643252,
|
| 52 |
+
"rewards/margins": -0.0003282243851572275,
|
| 53 |
+
"rewards/rejected": -0.00022249165340326726,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.03663003663003663,
|
| 58 |
+
"grad_norm": 2.881896681708338,
|
| 59 |
+
"learning_rate": 9.146341463414634e-08,
|
| 60 |
+
"logits/chosen": -2.478980779647827,
|
| 61 |
+
"logits/rejected": -2.4746451377868652,
|
| 62 |
+
"logps/chosen": -400.7593994140625,
|
| 63 |
+
"logps/rejected": -374.46099853515625,
|
| 64 |
+
"loss": 0.693,
|
| 65 |
+
"rewards/accuracies": 0.53125,
|
| 66 |
+
"rewards/chosen": 0.0005641734460368752,
|
| 67 |
+
"rewards/margins": 0.00045302818762138486,
|
| 68 |
+
"rewards/rejected": 0.00011114527296740562,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.04884004884004884,
|
| 73 |
+
"grad_norm": 3.0295854115026777,
|
| 74 |
+
"learning_rate": 1.219512195121951e-07,
|
| 75 |
+
"logits/chosen": -2.462890386581421,
|
| 76 |
+
"logits/rejected": -2.483200788497925,
|
| 77 |
+
"logps/chosen": -397.7459411621094,
|
| 78 |
+
"logps/rejected": -385.4634704589844,
|
| 79 |
+
"loss": 0.6922,
|
| 80 |
+
"rewards/accuracies": 0.5625,
|
| 81 |
+
"rewards/chosen": 0.0019774003885686398,
|
| 82 |
+
"rewards/margins": 0.0016252705827355385,
|
| 83 |
+
"rewards/rejected": 0.00035212995135225356,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.06105006105006105,
|
| 88 |
+
"grad_norm": 2.9757197429390847,
|
| 89 |
+
"learning_rate": 1.524390243902439e-07,
|
| 90 |
+
"logits/chosen": -2.4540820121765137,
|
| 91 |
+
"logits/rejected": -2.4638664722442627,
|
| 92 |
+
"logps/chosen": -391.7041931152344,
|
| 93 |
+
"logps/rejected": -370.6739196777344,
|
| 94 |
+
"loss": 0.6906,
|
| 95 |
+
"rewards/accuracies": 0.6656249761581421,
|
| 96 |
+
"rewards/chosen": 0.005458775907754898,
|
| 97 |
+
"rewards/margins": 0.004881708417087793,
|
| 98 |
+
"rewards/rejected": 0.0005770674324594438,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.07326007326007326,
|
| 103 |
+
"grad_norm": 3.0702933396650827,
|
| 104 |
+
"learning_rate": 1.8292682926829268e-07,
|
| 105 |
+
"logits/chosen": -2.453782796859741,
|
| 106 |
+
"logits/rejected": -2.4780030250549316,
|
| 107 |
+
"logps/chosen": -375.0745544433594,
|
| 108 |
+
"logps/rejected": -376.1570739746094,
|
| 109 |
+
"loss": 0.6872,
|
| 110 |
+
"rewards/accuracies": 0.7281249761581421,
|
| 111 |
+
"rewards/chosen": 0.011652452871203423,
|
| 112 |
+
"rewards/margins": 0.010647130198776722,
|
| 113 |
+
"rewards/rejected": 0.0010053229052573442,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.08547008547008547,
|
| 118 |
+
"grad_norm": 2.9490982076861703,
|
| 119 |
+
"learning_rate": 2.1341463414634144e-07,
|
| 120 |
+
"logits/chosen": -2.5242810249328613,
|
| 121 |
+
"logits/rejected": -2.513012170791626,
|
| 122 |
+
"logps/chosen": -402.56036376953125,
|
| 123 |
+
"logps/rejected": -372.23004150390625,
|
| 124 |
+
"loss": 0.6833,
|
| 125 |
+
"rewards/accuracies": 0.7906249761581421,
|
| 126 |
+
"rewards/chosen": 0.01956186816096306,
|
| 127 |
+
"rewards/margins": 0.018751125782728195,
|
| 128 |
+
"rewards/rejected": 0.0008107417961582541,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.09768009768009768,
|
| 133 |
+
"grad_norm": 2.995960567198045,
|
| 134 |
+
"learning_rate": 2.439024390243902e-07,
|
| 135 |
+
"logits/chosen": -2.4664769172668457,
|
| 136 |
+
"logits/rejected": -2.512596607208252,
|
| 137 |
+
"logps/chosen": -388.77276611328125,
|
| 138 |
+
"logps/rejected": -386.55120849609375,
|
| 139 |
+
"loss": 0.6754,
|
| 140 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 141 |
+
"rewards/chosen": 0.0328640341758728,
|
| 142 |
+
"rewards/margins": 0.035563308745622635,
|
| 143 |
+
"rewards/rejected": -0.0026992757339030504,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.10989010989010989,
|
| 148 |
+
"grad_norm": 3.0842908329196925,
|
| 149 |
+
"learning_rate": 2.7439024390243906e-07,
|
| 150 |
+
"logits/chosen": -2.583833932876587,
|
| 151 |
+
"logits/rejected": -2.6459927558898926,
|
| 152 |
+
"logps/chosen": -411.39117431640625,
|
| 153 |
+
"logps/rejected": -416.40771484375,
|
| 154 |
+
"loss": 0.6616,
|
| 155 |
+
"rewards/accuracies": 0.809374988079071,
|
| 156 |
+
"rewards/chosen": 0.04710187762975693,
|
| 157 |
+
"rewards/margins": 0.06961944699287415,
|
| 158 |
+
"rewards/rejected": -0.022517573088407516,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.1221001221001221,
|
| 163 |
+
"grad_norm": 2.588661111129042,
|
| 164 |
+
"learning_rate": 3.048780487804878e-07,
|
| 165 |
+
"logits/chosen": -2.5411438941955566,
|
| 166 |
+
"logits/rejected": -2.5637898445129395,
|
| 167 |
+
"logps/chosen": -382.1836853027344,
|
| 168 |
+
"logps/rejected": -380.8400573730469,
|
| 169 |
+
"loss": 0.6506,
|
| 170 |
+
"rewards/accuracies": 0.824999988079071,
|
| 171 |
+
"rewards/chosen": 0.053000353276729584,
|
| 172 |
+
"rewards/margins": 0.09493989497423172,
|
| 173 |
+
"rewards/rejected": -0.04193953052163124,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.1343101343101343,
|
| 178 |
+
"grad_norm": 2.921339715123654,
|
| 179 |
+
"learning_rate": 3.353658536585366e-07,
|
| 180 |
+
"logits/chosen": -2.5204200744628906,
|
| 181 |
+
"logits/rejected": -2.5622482299804688,
|
| 182 |
+
"logps/chosen": -393.4702453613281,
|
| 183 |
+
"logps/rejected": -395.1978454589844,
|
| 184 |
+
"loss": 0.6293,
|
| 185 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 186 |
+
"rewards/chosen": 0.03344795107841492,
|
| 187 |
+
"rewards/margins": 0.16028472781181335,
|
| 188 |
+
"rewards/rejected": -0.12683679163455963,
|
| 189 |
+
"step": 110
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.14652014652014653,
|
| 193 |
+
"grad_norm": 2.963460425484982,
|
| 194 |
+
"learning_rate": 3.6585365853658536e-07,
|
| 195 |
+
"logits/chosen": -2.508514881134033,
|
| 196 |
+
"logits/rejected": -2.5441009998321533,
|
| 197 |
+
"logps/chosen": -383.43133544921875,
|
| 198 |
+
"logps/rejected": -407.09210205078125,
|
| 199 |
+
"loss": 0.5887,
|
| 200 |
+
"rewards/accuracies": 0.809374988079071,
|
| 201 |
+
"rewards/chosen": -0.013098609633743763,
|
| 202 |
+
"rewards/margins": 0.2991253733634949,
|
| 203 |
+
"rewards/rejected": -0.31222397089004517,
|
| 204 |
+
"step": 120
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.15873015873015872,
|
| 208 |
+
"grad_norm": 3.0254473498968553,
|
| 209 |
+
"learning_rate": 3.9634146341463414e-07,
|
| 210 |
+
"logits/chosen": -2.542419910430908,
|
| 211 |
+
"logits/rejected": -2.5631086826324463,
|
| 212 |
+
"logps/chosen": -407.2526550292969,
|
| 213 |
+
"logps/rejected": -442.8262634277344,
|
| 214 |
+
"loss": 0.5587,
|
| 215 |
+
"rewards/accuracies": 0.800000011920929,
|
| 216 |
+
"rewards/chosen": -0.11188334226608276,
|
| 217 |
+
"rewards/margins": 0.4068102240562439,
|
| 218 |
+
"rewards/rejected": -0.5186935663223267,
|
| 219 |
+
"step": 130
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.17094017094017094,
|
| 223 |
+
"grad_norm": 3.612015672751943,
|
| 224 |
+
"learning_rate": 4.268292682926829e-07,
|
| 225 |
+
"logits/chosen": -2.5460963249206543,
|
| 226 |
+
"logits/rejected": -2.595196485519409,
|
| 227 |
+
"logps/chosen": -435.4662170410156,
|
| 228 |
+
"logps/rejected": -484.08270263671875,
|
| 229 |
+
"loss": 0.5266,
|
| 230 |
+
"rewards/accuracies": 0.796875,
|
| 231 |
+
"rewards/chosen": -0.2931057810783386,
|
| 232 |
+
"rewards/margins": 0.5774662494659424,
|
| 233 |
+
"rewards/rejected": -0.8705719709396362,
|
| 234 |
+
"step": 140
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.18315018315018314,
|
| 238 |
+
"grad_norm": 4.357042344454044,
|
| 239 |
+
"learning_rate": 4.573170731707317e-07,
|
| 240 |
+
"logits/chosen": -2.4378104209899902,
|
| 241 |
+
"logits/rejected": -2.4749557971954346,
|
| 242 |
+
"logps/chosen": -473.6615295410156,
|
| 243 |
+
"logps/rejected": -519.7752685546875,
|
| 244 |
+
"loss": 0.5035,
|
| 245 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 246 |
+
"rewards/chosen": -0.6644935607910156,
|
| 247 |
+
"rewards/margins": 0.7163723707199097,
|
| 248 |
+
"rewards/rejected": -1.3808658123016357,
|
| 249 |
+
"step": 150
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.19536019536019536,
|
| 253 |
+
"grad_norm": 4.657915772314801,
|
| 254 |
+
"learning_rate": 4.878048780487804e-07,
|
| 255 |
+
"logits/chosen": -2.3850626945495605,
|
| 256 |
+
"logits/rejected": -2.404574394226074,
|
| 257 |
+
"logps/chosen": -475.3724060058594,
|
| 258 |
+
"logps/rejected": -554.40576171875,
|
| 259 |
+
"loss": 0.4585,
|
| 260 |
+
"rewards/accuracies": 0.793749988079071,
|
| 261 |
+
"rewards/chosen": -0.9302708506584167,
|
| 262 |
+
"rewards/margins": 0.9146777987480164,
|
| 263 |
+
"rewards/rejected": -1.8449485301971436,
|
| 264 |
+
"step": 160
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.20757020757020758,
|
| 268 |
+
"grad_norm": 5.145238611541166,
|
| 269 |
+
"learning_rate": 4.999795585653115e-07,
|
| 270 |
+
"logits/chosen": -2.4656128883361816,
|
| 271 |
+
"logits/rejected": -2.5041518211364746,
|
| 272 |
+
"logps/chosen": -500.9999084472656,
|
| 273 |
+
"logps/rejected": -621.8275146484375,
|
| 274 |
+
"loss": 0.4203,
|
| 275 |
+
"rewards/accuracies": 0.846875011920929,
|
| 276 |
+
"rewards/chosen": -1.0742014646530151,
|
| 277 |
+
"rewards/margins": 1.363214135169983,
|
| 278 |
+
"rewards/rejected": -2.437415599822998,
|
| 279 |
+
"step": 170
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.21978021978021978,
|
| 283 |
+
"grad_norm": 5.498621709371294,
|
| 284 |
+
"learning_rate": 4.998546507921325e-07,
|
| 285 |
+
"logits/chosen": -2.3531060218811035,
|
| 286 |
+
"logits/rejected": -2.3719255924224854,
|
| 287 |
+
"logps/chosen": -516.4880981445312,
|
| 288 |
+
"logps/rejected": -656.3202514648438,
|
| 289 |
+
"loss": 0.3875,
|
| 290 |
+
"rewards/accuracies": 0.8125,
|
| 291 |
+
"rewards/chosen": -1.3029719591140747,
|
| 292 |
+
"rewards/margins": 1.5469024181365967,
|
| 293 |
+
"rewards/rejected": -2.8498740196228027,
|
| 294 |
+
"step": 180
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.231990231990232,
|
| 298 |
+
"grad_norm": 4.777943322815817,
|
| 299 |
+
"learning_rate": 4.996162482680895e-07,
|
| 300 |
+
"logits/chosen": -2.383643388748169,
|
| 301 |
+
"logits/rejected": -2.390869617462158,
|
| 302 |
+
"logps/chosen": -557.1913452148438,
|
| 303 |
+
"logps/rejected": -686.9656372070312,
|
| 304 |
+
"loss": 0.3856,
|
| 305 |
+
"rewards/accuracies": 0.828125,
|
| 306 |
+
"rewards/chosen": -1.459194540977478,
|
| 307 |
+
"rewards/margins": 1.6121686697006226,
|
| 308 |
+
"rewards/rejected": -3.0713634490966797,
|
| 309 |
+
"step": 190
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.2442002442002442,
|
| 313 |
+
"grad_norm": 5.801515238018521,
|
| 314 |
+
"learning_rate": 4.992644592858842e-07,
|
| 315 |
+
"logits/chosen": -2.3923563957214355,
|
| 316 |
+
"logits/rejected": -2.444206476211548,
|
| 317 |
+
"logps/chosen": -532.6453247070312,
|
| 318 |
+
"logps/rejected": -670.8153076171875,
|
| 319 |
+
"loss": 0.3645,
|
| 320 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 321 |
+
"rewards/chosen": -1.294365406036377,
|
| 322 |
+
"rewards/margins": 1.502074122428894,
|
| 323 |
+
"rewards/rejected": -2.7964394092559814,
|
| 324 |
+
"step": 200
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2564102564102564,
|
| 328 |
+
"grad_norm": 7.316959768144989,
|
| 329 |
+
"learning_rate": 4.987994436432335e-07,
|
| 330 |
+
"logits/chosen": -2.342782497406006,
|
| 331 |
+
"logits/rejected": -2.362506151199341,
|
| 332 |
+
"logps/chosen": -557.2277221679688,
|
| 333 |
+
"logps/rejected": -723.668701171875,
|
| 334 |
+
"loss": 0.3664,
|
| 335 |
+
"rewards/accuracies": 0.846875011920929,
|
| 336 |
+
"rewards/chosen": -1.531425952911377,
|
| 337 |
+
"rewards/margins": 1.8236634731292725,
|
| 338 |
+
"rewards/rejected": -3.3550896644592285,
|
| 339 |
+
"step": 210
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.2686202686202686,
|
| 343 |
+
"grad_norm": 5.84414999103754,
|
| 344 |
+
"learning_rate": 4.982214125702845e-07,
|
| 345 |
+
"logits/chosen": -2.363788366317749,
|
| 346 |
+
"logits/rejected": -2.4076294898986816,
|
| 347 |
+
"logps/chosen": -515.8175659179688,
|
| 348 |
+
"logps/rejected": -681.577880859375,
|
| 349 |
+
"loss": 0.3513,
|
| 350 |
+
"rewards/accuracies": 0.856249988079071,
|
| 351 |
+
"rewards/chosen": -1.2184154987335205,
|
| 352 |
+
"rewards/margins": 1.8515632152557373,
|
| 353 |
+
"rewards/rejected": -3.069979190826416,
|
| 354 |
+
"step": 220
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.28083028083028083,
|
| 358 |
+
"grad_norm": 5.57281060925085,
|
| 359 |
+
"learning_rate": 4.975306286336627e-07,
|
| 360 |
+
"logits/chosen": -2.390061855316162,
|
| 361 |
+
"logits/rejected": -2.4118847846984863,
|
| 362 |
+
"logps/chosen": -558.8645629882812,
|
| 363 |
+
"logps/rejected": -748.7266845703125,
|
| 364 |
+
"loss": 0.3603,
|
| 365 |
+
"rewards/accuracies": 0.824999988079071,
|
| 366 |
+
"rewards/chosen": -1.5887969732284546,
|
| 367 |
+
"rewards/margins": 1.980838418006897,
|
| 368 |
+
"rewards/rejected": -3.5696358680725098,
|
| 369 |
+
"step": 230
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.29304029304029305,
|
| 373 |
+
"grad_norm": 5.714762684377855,
|
| 374 |
+
"learning_rate": 4.967274056172044e-07,
|
| 375 |
+
"logits/chosen": -2.3940510749816895,
|
| 376 |
+
"logits/rejected": -2.4176127910614014,
|
| 377 |
+
"logps/chosen": -543.5166015625,
|
| 378 |
+
"logps/rejected": -715.0543212890625,
|
| 379 |
+
"loss": 0.3498,
|
| 380 |
+
"rewards/accuracies": 0.84375,
|
| 381 |
+
"rewards/chosen": -1.3804460763931274,
|
| 382 |
+
"rewards/margins": 1.9572741985321045,
|
| 383 |
+
"rewards/rejected": -3.3377201557159424,
|
| 384 |
+
"step": 240
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.3052503052503053,
|
| 388 |
+
"grad_norm": 5.674577347175447,
|
| 389 |
+
"learning_rate": 4.958121083794216e-07,
|
| 390 |
+
"logits/chosen": -2.360697031021118,
|
| 391 |
+
"logits/rejected": -2.3762993812561035,
|
| 392 |
+
"logps/chosen": -573.5714111328125,
|
| 393 |
+
"logps/rejected": -768.0198364257812,
|
| 394 |
+
"loss": 0.3375,
|
| 395 |
+
"rewards/accuracies": 0.8343750238418579,
|
| 396 |
+
"rewards/chosen": -1.5868339538574219,
|
| 397 |
+
"rewards/margins": 2.1389665603637695,
|
| 398 |
+
"rewards/rejected": -3.7258007526397705,
|
| 399 |
+
"step": 250
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.31746031746031744,
|
| 403 |
+
"grad_norm": 6.3452788515852365,
|
| 404 |
+
"learning_rate": 4.947851526877681e-07,
|
| 405 |
+
"logits/chosen": -2.3265128135681152,
|
| 406 |
+
"logits/rejected": -2.354276180267334,
|
| 407 |
+
"logps/chosen": -530.7965087890625,
|
| 408 |
+
"logps/rejected": -724.9639282226562,
|
| 409 |
+
"loss": 0.3366,
|
| 410 |
+
"rewards/accuracies": 0.8125,
|
| 411 |
+
"rewards/chosen": -1.4949758052825928,
|
| 412 |
+
"rewards/margins": 2.002520799636841,
|
| 413 |
+
"rewards/rejected": -3.4974968433380127,
|
| 414 |
+
"step": 260
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.32967032967032966,
|
| 418 |
+
"grad_norm": 6.281289415201168,
|
| 419 |
+
"learning_rate": 4.936470050297798e-07,
|
| 420 |
+
"logits/chosen": -2.301875352859497,
|
| 421 |
+
"logits/rejected": -2.3351759910583496,
|
| 422 |
+
"logps/chosen": -551.602294921875,
|
| 423 |
+
"logps/rejected": -743.7047119140625,
|
| 424 |
+
"loss": 0.3241,
|
| 425 |
+
"rewards/accuracies": 0.8343750238418579,
|
| 426 |
+
"rewards/chosen": -1.5574215650558472,
|
| 427 |
+
"rewards/margins": 2.1133522987365723,
|
| 428 |
+
"rewards/rejected": -3.67077374458313,
|
| 429 |
+
"step": 270
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.3418803418803419,
|
| 433 |
+
"grad_norm": 6.620858897078139,
|
| 434 |
+
"learning_rate": 4.92398182401176e-07,
|
| 435 |
+
"logits/chosen": -2.3097803592681885,
|
| 436 |
+
"logits/rejected": -2.3185763359069824,
|
| 437 |
+
"logps/chosen": -573.0301513671875,
|
| 438 |
+
"logps/rejected": -778.9459228515625,
|
| 439 |
+
"loss": 0.3358,
|
| 440 |
+
"rewards/accuracies": 0.871874988079071,
|
| 441 |
+
"rewards/chosen": -1.798607587814331,
|
| 442 |
+
"rewards/margins": 2.2159924507141113,
|
| 443 |
+
"rewards/rejected": -4.0146002769470215,
|
| 444 |
+
"step": 280
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.3540903540903541,
|
| 448 |
+
"grad_norm": 6.930259744533563,
|
| 449 |
+
"learning_rate": 4.910392520710174e-07,
|
| 450 |
+
"logits/chosen": -2.301408529281616,
|
| 451 |
+
"logits/rejected": -2.3531241416931152,
|
| 452 |
+
"logps/chosen": -536.4937133789062,
|
| 453 |
+
"logps/rejected": -733.7653198242188,
|
| 454 |
+
"loss": 0.3233,
|
| 455 |
+
"rewards/accuracies": 0.846875011920929,
|
| 456 |
+
"rewards/chosen": -1.5255537033081055,
|
| 457 |
+
"rewards/margins": 2.0018668174743652,
|
| 458 |
+
"rewards/rejected": -3.52742075920105,
|
| 459 |
+
"step": 290
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.3663003663003663,
|
| 463 |
+
"grad_norm": 6.863827814496594,
|
| 464 |
+
"learning_rate": 4.895708313240285e-07,
|
| 465 |
+
"logits/chosen": -2.3684794902801514,
|
| 466 |
+
"logits/rejected": -2.3983054161071777,
|
| 467 |
+
"logps/chosen": -560.7042236328125,
|
| 468 |
+
"logps/rejected": -786.914794921875,
|
| 469 |
+
"loss": 0.3143,
|
| 470 |
+
"rewards/accuracies": 0.878125011920929,
|
| 471 |
+
"rewards/chosen": -1.5986013412475586,
|
| 472 |
+
"rewards/margins": 2.3771414756774902,
|
| 473 |
+
"rewards/rejected": -3.9757423400878906,
|
| 474 |
+
"step": 300
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.3785103785103785,
|
| 478 |
+
"grad_norm": 6.618762045262814,
|
| 479 |
+
"learning_rate": 4.879935871802001e-07,
|
| 480 |
+
"logits/chosen": -2.340313196182251,
|
| 481 |
+
"logits/rejected": -2.348755121231079,
|
| 482 |
+
"logps/chosen": -555.6011962890625,
|
| 483 |
+
"logps/rejected": -804.3768310546875,
|
| 484 |
+
"loss": 0.3199,
|
| 485 |
+
"rewards/accuracies": 0.875,
|
| 486 |
+
"rewards/chosen": -1.6778042316436768,
|
| 487 |
+
"rewards/margins": 2.592184543609619,
|
| 488 |
+
"rewards/rejected": -4.269989013671875,
|
| 489 |
+
"step": 310
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.3907203907203907,
|
| 493 |
+
"grad_norm": 6.104172520347164,
|
| 494 |
+
"learning_rate": 4.863082360917998e-07,
|
| 495 |
+
"logits/chosen": -2.300529956817627,
|
| 496 |
+
"logits/rejected": -2.3536949157714844,
|
| 497 |
+
"logps/chosen": -565.5293579101562,
|
| 498 |
+
"logps/rejected": -806.9231567382812,
|
| 499 |
+
"loss": 0.3146,
|
| 500 |
+
"rewards/accuracies": 0.856249988079071,
|
| 501 |
+
"rewards/chosen": -1.7859094142913818,
|
| 502 |
+
"rewards/margins": 2.411153793334961,
|
| 503 |
+
"rewards/rejected": -4.19706392288208,
|
| 504 |
+
"step": 320
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.40293040293040294,
|
| 508 |
+
"grad_norm": 13.362839259142513,
|
| 509 |
+
"learning_rate": 4.845155436179286e-07,
|
| 510 |
+
"logits/chosen": -2.2448084354400635,
|
| 511 |
+
"logits/rejected": -2.265852451324463,
|
| 512 |
+
"logps/chosen": -558.632080078125,
|
| 513 |
+
"logps/rejected": -792.7662963867188,
|
| 514 |
+
"loss": 0.2963,
|
| 515 |
+
"rewards/accuracies": 0.8687499761581421,
|
| 516 |
+
"rewards/chosen": -1.8193552494049072,
|
| 517 |
+
"rewards/margins": 2.476144790649414,
|
| 518 |
+
"rewards/rejected": -4.2955002784729,
|
| 519 |
+
"step": 330
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.41514041514041516,
|
| 523 |
+
"grad_norm": 8.193239382667205,
|
| 524 |
+
"learning_rate": 4.826163240767716e-07,
|
| 525 |
+
"logits/chosen": -2.2685139179229736,
|
| 526 |
+
"logits/rejected": -2.291156053543091,
|
| 527 |
+
"logps/chosen": -578.7394409179688,
|
| 528 |
+
"logps/rejected": -828.9251098632812,
|
| 529 |
+
"loss": 0.2845,
|
| 530 |
+
"rewards/accuracies": 0.909375011920929,
|
| 531 |
+
"rewards/chosen": -1.7241359949111938,
|
| 532 |
+
"rewards/margins": 2.655815601348877,
|
| 533 |
+
"rewards/rejected": -4.379951000213623,
|
| 534 |
+
"step": 340
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.42735042735042733,
|
| 538 |
+
"grad_norm": 11.339490692180462,
|
| 539 |
+
"learning_rate": 4.806114401756988e-07,
|
| 540 |
+
"logits/chosen": -2.278423309326172,
|
| 541 |
+
"logits/rejected": -2.281816244125366,
|
| 542 |
+
"logps/chosen": -609.6286010742188,
|
| 543 |
+
"logps/rejected": -881.9080810546875,
|
| 544 |
+
"loss": 0.3005,
|
| 545 |
+
"rewards/accuracies": 0.887499988079071,
|
| 546 |
+
"rewards/chosen": -2.0114948749542236,
|
| 547 |
+
"rewards/margins": 2.8581748008728027,
|
| 548 |
+
"rewards/rejected": -4.8696699142456055,
|
| 549 |
+
"step": 350
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.43956043956043955,
|
| 553 |
+
"grad_norm": 7.458240778141653,
|
| 554 |
+
"learning_rate": 4.785018026193862e-07,
|
| 555 |
+
"logits/chosen": -2.3067989349365234,
|
| 556 |
+
"logits/rejected": -2.302696943283081,
|
| 557 |
+
"logps/chosen": -579.6629638671875,
|
| 558 |
+
"logps/rejected": -773.7774658203125,
|
| 559 |
+
"loss": 0.3119,
|
| 560 |
+
"rewards/accuracies": 0.8687499761581421,
|
| 561 |
+
"rewards/chosen": -1.6714990139007568,
|
| 562 |
+
"rewards/margins": 2.2192506790161133,
|
| 563 |
+
"rewards/rejected": -3.890749454498291,
|
| 564 |
+
"step": 360
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.4517704517704518,
|
| 568 |
+
"grad_norm": 7.753160256439322,
|
| 569 |
+
"learning_rate": 4.762883696961353e-07,
|
| 570 |
+
"logits/chosen": -2.2718212604522705,
|
| 571 |
+
"logits/rejected": -2.275289535522461,
|
| 572 |
+
"logps/chosen": -606.1307373046875,
|
| 573 |
+
"logps/rejected": -799.2538452148438,
|
| 574 |
+
"loss": 0.2968,
|
| 575 |
+
"rewards/accuracies": 0.8687499761581421,
|
| 576 |
+
"rewards/chosen": -2.069481134414673,
|
| 577 |
+
"rewards/margins": 2.2473983764648438,
|
| 578 |
+
"rewards/rejected": -4.316879749298096,
|
| 579 |
+
"step": 370
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.463980463980464,
|
| 583 |
+
"grad_norm": 7.668480163136406,
|
| 584 |
+
"learning_rate": 4.739721468425763e-07,
|
| 585 |
+
"logits/chosen": -2.196530342102051,
|
| 586 |
+
"logits/rejected": -2.2193641662597656,
|
| 587 |
+
"logps/chosen": -598.3923950195312,
|
| 588 |
+
"logps/rejected": -832.3982543945312,
|
| 589 |
+
"loss": 0.2843,
|
| 590 |
+
"rewards/accuracies": 0.840624988079071,
|
| 591 |
+
"rewards/chosen": -2.077108383178711,
|
| 592 |
+
"rewards/margins": 2.470067024230957,
|
| 593 |
+
"rewards/rejected": -4.547175407409668,
|
| 594 |
+
"step": 380
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 0.47619047619047616,
|
| 598 |
+
"grad_norm": 7.295306779149404,
|
| 599 |
+
"learning_rate": 4.715541861869562e-07,
|
| 600 |
+
"logits/chosen": -2.260920763015747,
|
| 601 |
+
"logits/rejected": -2.2953712940216064,
|
| 602 |
+
"logps/chosen": -582.3689575195312,
|
| 603 |
+
"logps/rejected": -889.9528198242188,
|
| 604 |
+
"loss": 0.2779,
|
| 605 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 606 |
+
"rewards/chosen": -1.7534271478652954,
|
| 607 |
+
"rewards/margins": 3.1180617809295654,
|
| 608 |
+
"rewards/rejected": -4.871488571166992,
|
| 609 |
+
"step": 390
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.4884004884004884,
|
| 613 |
+
"grad_norm": 8.066116525077385,
|
| 614 |
+
"learning_rate": 4.690355860712163e-07,
|
| 615 |
+
"logits/chosen": -2.222865581512451,
|
| 616 |
+
"logits/rejected": -2.278761386871338,
|
| 617 |
+
"logps/chosen": -597.5587158203125,
|
| 618 |
+
"logps/rejected": -863.9420776367188,
|
| 619 |
+
"loss": 0.2852,
|
| 620 |
+
"rewards/accuracies": 0.893750011920929,
|
| 621 |
+
"rewards/chosen": -1.9404022693634033,
|
| 622 |
+
"rewards/margins": 2.7501840591430664,
|
| 623 |
+
"rewards/rejected": -4.690586090087891,
|
| 624 |
+
"step": 400
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 0.5006105006105006,
|
| 628 |
+
"grad_norm": 7.116028285618711,
|
| 629 |
+
"learning_rate": 4.664174905520782e-07,
|
| 630 |
+
"logits/chosen": -2.1432394981384277,
|
| 631 |
+
"logits/rejected": -2.1721110343933105,
|
| 632 |
+
"logps/chosen": -567.879638671875,
|
| 633 |
+
"logps/rejected": -829.4773559570312,
|
| 634 |
+
"loss": 0.2967,
|
| 635 |
+
"rewards/accuracies": 0.846875011920929,
|
| 636 |
+
"rewards/chosen": -1.8987783193588257,
|
| 637 |
+
"rewards/margins": 2.688133716583252,
|
| 638 |
+
"rewards/rejected": -4.586911678314209,
|
| 639 |
+
"step": 410
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5128205128205128,
|
| 643 |
+
"grad_norm": 7.04360607650549,
|
| 644 |
+
"learning_rate": 4.637010888813638e-07,
|
| 645 |
+
"logits/chosen": -2.182570695877075,
|
| 646 |
+
"logits/rejected": -2.1820969581604004,
|
| 647 |
+
"logps/chosen": -577.9962768554688,
|
| 648 |
+
"logps/rejected": -821.7763671875,
|
| 649 |
+
"loss": 0.2965,
|
| 650 |
+
"rewards/accuracies": 0.84375,
|
| 651 |
+
"rewards/chosen": -1.8270905017852783,
|
| 652 |
+
"rewards/margins": 2.6087398529052734,
|
| 653 |
+
"rewards/rejected": -4.435830116271973,
|
| 654 |
+
"step": 420
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.525030525030525,
|
| 658 |
+
"grad_norm": 8.266012865923875,
|
| 659 |
+
"learning_rate": 4.608876149657862e-07,
|
| 660 |
+
"logits/chosen": -2.1726815700531006,
|
| 661 |
+
"logits/rejected": -2.179741621017456,
|
| 662 |
+
"logps/chosen": -615.1981201171875,
|
| 663 |
+
"logps/rejected": -858.2872314453125,
|
| 664 |
+
"loss": 0.2806,
|
| 665 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 666 |
+
"rewards/chosen": -2.058166980743408,
|
| 667 |
+
"rewards/margins": 2.7252743244171143,
|
| 668 |
+
"rewards/rejected": -4.783441066741943,
|
| 669 |
+
"step": 430
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.5372405372405372,
|
| 673 |
+
"grad_norm": 7.654156099725738,
|
| 674 |
+
"learning_rate": 4.5797834680645553e-07,
|
| 675 |
+
"logits/chosen": -2.149817943572998,
|
| 676 |
+
"logits/rejected": -2.1750922203063965,
|
| 677 |
+
"logps/chosen": -626.6456909179688,
|
| 678 |
+
"logps/rejected": -924.1580810546875,
|
| 679 |
+
"loss": 0.2855,
|
| 680 |
+
"rewards/accuracies": 0.887499988079071,
|
| 681 |
+
"rewards/chosen": -2.245823621749878,
|
| 682 |
+
"rewards/margins": 3.026350736618042,
|
| 683 |
+
"rewards/rejected": -5.27217435836792,
|
| 684 |
+
"step": 440
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 0.5494505494505495,
|
| 688 |
+
"grad_norm": 7.2185121675912445,
|
| 689 |
+
"learning_rate": 4.549746059183561e-07,
|
| 690 |
+
"logits/chosen": -2.134995460510254,
|
| 691 |
+
"logits/rejected": -2.1586225032806396,
|
| 692 |
+
"logps/chosen": -588.483642578125,
|
| 693 |
+
"logps/rejected": -902.3905029296875,
|
| 694 |
+
"loss": 0.2651,
|
| 695 |
+
"rewards/accuracies": 0.893750011920929,
|
| 696 |
+
"rewards/chosen": -2.05491304397583,
|
| 697 |
+
"rewards/margins": 3.2267088890075684,
|
| 698 |
+
"rewards/rejected": -5.281621932983398,
|
| 699 |
+
"step": 450
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.5616605616605617,
|
| 703 |
+
"grad_norm": 8.705469465511992,
|
| 704 |
+
"learning_rate": 4.5187775673005744e-07,
|
| 705 |
+
"logits/chosen": -2.0956122875213623,
|
| 706 |
+
"logits/rejected": -2.122889757156372,
|
| 707 |
+
"logps/chosen": -614.2383422851562,
|
| 708 |
+
"logps/rejected": -854.7008666992188,
|
| 709 |
+
"loss": 0.2706,
|
| 710 |
+
"rewards/accuracies": 0.875,
|
| 711 |
+
"rewards/chosen": -2.2875373363494873,
|
| 712 |
+
"rewards/margins": 2.5731403827667236,
|
| 713 |
+
"rewards/rejected": -4.860678195953369,
|
| 714 |
+
"step": 460
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.5738705738705738,
|
| 718 |
+
"grad_norm": 6.674714090077302,
|
| 719 |
+
"learning_rate": 4.4868920596393197e-07,
|
| 720 |
+
"logits/chosen": -2.122455596923828,
|
| 721 |
+
"logits/rejected": -2.1570394039154053,
|
| 722 |
+
"logps/chosen": -628.7379150390625,
|
| 723 |
+
"logps/rejected": -913.1719970703125,
|
| 724 |
+
"loss": 0.2635,
|
| 725 |
+
"rewards/accuracies": 0.8531249761581421,
|
| 726 |
+
"rewards/chosen": -2.345153331756592,
|
| 727 |
+
"rewards/margins": 2.962648630142212,
|
| 728 |
+
"rewards/rejected": -5.307801723480225,
|
| 729 |
+
"step": 470
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.5860805860805861,
|
| 733 |
+
"grad_norm": 10.484528534722092,
|
| 734 |
+
"learning_rate": 4.4541040199716063e-07,
|
| 735 |
+
"logits/chosen": -2.1315033435821533,
|
| 736 |
+
"logits/rejected": -2.133057117462158,
|
| 737 |
+
"logps/chosen": -652.1898803710938,
|
| 738 |
+
"logps/rejected": -918.3726806640625,
|
| 739 |
+
"loss": 0.2841,
|
| 740 |
+
"rewards/accuracies": 0.887499988079071,
|
| 741 |
+
"rewards/chosen": -2.442749261856079,
|
| 742 |
+
"rewards/margins": 2.9514050483703613,
|
| 743 |
+
"rewards/rejected": -5.3941545486450195,
|
| 744 |
+
"step": 480
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.5982905982905983,
|
| 748 |
+
"grad_norm": 6.764572032034108,
|
| 749 |
+
"learning_rate": 4.4204283420381827e-07,
|
| 750 |
+
"logits/chosen": -2.0328996181488037,
|
| 751 |
+
"logits/rejected": -2.0834460258483887,
|
| 752 |
+
"logps/chosen": -605.2128295898438,
|
| 753 |
+
"logps/rejected": -897.3580322265625,
|
| 754 |
+
"loss": 0.2652,
|
| 755 |
+
"rewards/accuracies": 0.871874988079071,
|
| 756 |
+
"rewards/chosen": -2.2144479751586914,
|
| 757 |
+
"rewards/margins": 2.9639675617218018,
|
| 758 |
+
"rewards/rejected": -5.178415775299072,
|
| 759 |
+
"step": 490
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.6105006105006106,
|
| 763 |
+
"grad_norm": 7.778145469907922,
|
| 764 |
+
"learning_rate": 4.3858803227833526e-07,
|
| 765 |
+
"logits/chosen": -1.9765386581420898,
|
| 766 |
+
"logits/rejected": -2.015956401824951,
|
| 767 |
+
"logps/chosen": -608.0709838867188,
|
| 768 |
+
"logps/rejected": -899.2232666015625,
|
| 769 |
+
"loss": 0.2715,
|
| 770 |
+
"rewards/accuracies": 0.90625,
|
| 771 |
+
"rewards/chosen": -2.045722484588623,
|
| 772 |
+
"rewards/margins": 3.0392613410949707,
|
| 773 |
+
"rewards/rejected": -5.084983825683594,
|
| 774 |
+
"step": 500
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.6227106227106227,
|
| 778 |
+
"grad_norm": 7.943002170259063,
|
| 779 |
+
"learning_rate": 4.350475655406445e-07,
|
| 780 |
+
"logits/chosen": -2.0267040729522705,
|
| 781 |
+
"logits/rejected": -2.0642428398132324,
|
| 782 |
+
"logps/chosen": -634.7288208007812,
|
| 783 |
+
"logps/rejected": -928.9196166992188,
|
| 784 |
+
"loss": 0.2677,
|
| 785 |
+
"rewards/accuracies": 0.875,
|
| 786 |
+
"rewards/chosen": -2.3875765800476074,
|
| 787 |
+
"rewards/margins": 3.001603126525879,
|
| 788 |
+
"rewards/rejected": -5.3891801834106445,
|
| 789 |
+
"step": 510
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.6349206349206349,
|
| 793 |
+
"grad_norm": 8.177071581241291,
|
| 794 |
+
"learning_rate": 4.314230422233286e-07,
|
| 795 |
+
"logits/chosen": -1.9996439218521118,
|
| 796 |
+
"logits/rejected": -2.001002788543701,
|
| 797 |
+
"logps/chosen": -606.2533569335938,
|
| 798 |
+
"logps/rejected": -888.8444213867188,
|
| 799 |
+
"loss": 0.2508,
|
| 800 |
+
"rewards/accuracies": 0.875,
|
| 801 |
+
"rewards/chosen": -2.237772226333618,
|
| 802 |
+
"rewards/margins": 2.8773951530456543,
|
| 803 |
+
"rewards/rejected": -5.115168571472168,
|
| 804 |
+
"step": 520
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 0.6471306471306472,
|
| 808 |
+
"grad_norm": 7.600982395285539,
|
| 809 |
+
"learning_rate": 4.2771610874109166e-07,
|
| 810 |
+
"logits/chosen": -1.9806162118911743,
|
| 811 |
+
"logits/rejected": -2.036025285720825,
|
| 812 |
+
"logps/chosen": -653.9401245117188,
|
| 813 |
+
"logps/rejected": -974.8654174804688,
|
| 814 |
+
"loss": 0.2529,
|
| 815 |
+
"rewards/accuracies": 0.90625,
|
| 816 |
+
"rewards/chosen": -2.513798236846924,
|
| 817 |
+
"rewards/margins": 3.2604928016662598,
|
| 818 |
+
"rewards/rejected": -5.774291038513184,
|
| 819 |
+
"step": 530
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.6593406593406593,
|
| 823 |
+
"grad_norm": 9.082170828519892,
|
| 824 |
+
"learning_rate": 4.2392844894288605e-07,
|
| 825 |
+
"logits/chosen": -1.9804267883300781,
|
| 826 |
+
"logits/rejected": -1.9906890392303467,
|
| 827 |
+
"logps/chosen": -633.1802368164062,
|
| 828 |
+
"logps/rejected": -931.8502197265625,
|
| 829 |
+
"loss": 0.2571,
|
| 830 |
+
"rewards/accuracies": 0.921875,
|
| 831 |
+
"rewards/chosen": -2.3836662769317627,
|
| 832 |
+
"rewards/margins": 3.0242981910705566,
|
| 833 |
+
"rewards/rejected": -5.40796422958374,
|
| 834 |
+
"step": 540
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 0.6715506715506715,
|
| 838 |
+
"grad_norm": 7.837053802282412,
|
| 839 |
+
"learning_rate": 4.2006178334703636e-07,
|
| 840 |
+
"logits/chosen": -2.006538152694702,
|
| 841 |
+
"logits/rejected": -2.0499379634857178,
|
| 842 |
+
"logps/chosen": -650.6314697265625,
|
| 843 |
+
"logps/rejected": -932.8741455078125,
|
| 844 |
+
"loss": 0.2636,
|
| 845 |
+
"rewards/accuracies": 0.8656250238418579,
|
| 846 |
+
"rewards/chosen": -2.4409451484680176,
|
| 847 |
+
"rewards/margins": 3.0085086822509766,
|
| 848 |
+
"rewards/rejected": -5.449453830718994,
|
| 849 |
+
"step": 550
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.6837606837606838,
|
| 853 |
+
"grad_norm": 6.941390377752381,
|
| 854 |
+
"learning_rate": 4.161178683597054e-07,
|
| 855 |
+
"logits/chosen": -2.0168557167053223,
|
| 856 |
+
"logits/rejected": -1.9983936548233032,
|
| 857 |
+
"logps/chosen": -596.7725830078125,
|
| 858 |
+
"logps/rejected": -862.6604614257812,
|
| 859 |
+
"loss": 0.2596,
|
| 860 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 861 |
+
"rewards/chosen": -2.015202283859253,
|
| 862 |
+
"rewards/margins": 2.8551807403564453,
|
| 863 |
+
"rewards/rejected": -4.870382785797119,
|
| 864 |
+
"step": 560
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.6959706959706959,
|
| 868 |
+
"grad_norm": 9.116023944385924,
|
| 869 |
+
"learning_rate": 4.1209849547705916e-07,
|
| 870 |
+
"logits/chosen": -1.9668136835098267,
|
| 871 |
+
"logits/rejected": -1.9564120769500732,
|
| 872 |
+
"logps/chosen": -597.05419921875,
|
| 873 |
+
"logps/rejected": -889.5735473632812,
|
| 874 |
+
"loss": 0.2673,
|
| 875 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 876 |
+
"rewards/chosen": -1.908453345298767,
|
| 877 |
+
"rewards/margins": 3.2164673805236816,
|
| 878 |
+
"rewards/rejected": -5.124920845031738,
|
| 879 |
+
"step": 570
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.7081807081807082,
|
| 883 |
+
"grad_norm": 9.73618947224433,
|
| 884 |
+
"learning_rate": 4.080054904714917e-07,
|
| 885 |
+
"logits/chosen": -2.0258185863494873,
|
| 886 |
+
"logits/rejected": -2.048612594604492,
|
| 887 |
+
"logps/chosen": -642.0574951171875,
|
| 888 |
+
"logps/rejected": -941.4654541015625,
|
| 889 |
+
"loss": 0.2523,
|
| 890 |
+
"rewards/accuracies": 0.893750011920929,
|
| 891 |
+
"rewards/chosen": -2.526458263397217,
|
| 892 |
+
"rewards/margins": 3.056541919708252,
|
| 893 |
+
"rewards/rejected": -5.583000183105469,
|
| 894 |
+
"step": 580
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.7203907203907204,
|
| 898 |
+
"grad_norm": 8.420821275679248,
|
| 899 |
+
"learning_rate": 4.038407125622806e-07,
|
| 900 |
+
"logits/chosen": -1.9966434240341187,
|
| 901 |
+
"logits/rejected": -1.9993374347686768,
|
| 902 |
+
"logps/chosen": -637.5452880859375,
|
| 903 |
+
"logps/rejected": -916.9855346679688,
|
| 904 |
+
"loss": 0.2664,
|
| 905 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 906 |
+
"rewards/chosen": -2.3783979415893555,
|
| 907 |
+
"rewards/margins": 3.0379176139831543,
|
| 908 |
+
"rewards/rejected": -5.41631555557251,
|
| 909 |
+
"step": 590
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.7326007326007326,
|
| 913 |
+
"grad_norm": 11.758932811938493,
|
| 914 |
+
"learning_rate": 3.9960605357105e-07,
|
| 915 |
+
"logits/chosen": -2.070080041885376,
|
| 916 |
+
"logits/rejected": -2.069638729095459,
|
| 917 |
+
"logps/chosen": -621.4031372070312,
|
| 918 |
+
"logps/rejected": -907.4441528320312,
|
| 919 |
+
"loss": 0.2693,
|
| 920 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 921 |
+
"rewards/chosen": -2.1775975227355957,
|
| 922 |
+
"rewards/margins": 3.1478095054626465,
|
| 923 |
+
"rewards/rejected": -5.3254075050354,
|
| 924 |
+
"step": 600
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 0.7448107448107448,
|
| 928 |
+
"grad_norm": 7.854542997142867,
|
| 929 |
+
"learning_rate": 3.95303437062423e-07,
|
| 930 |
+
"logits/chosen": -1.9713811874389648,
|
| 931 |
+
"logits/rejected": -1.977330207824707,
|
| 932 |
+
"logps/chosen": -642.8685913085938,
|
| 933 |
+
"logps/rejected": -944.4246826171875,
|
| 934 |
+
"loss": 0.2427,
|
| 935 |
+
"rewards/accuracies": 0.909375011920929,
|
| 936 |
+
"rewards/chosen": -2.425691604614258,
|
| 937 |
+
"rewards/margins": 3.1849162578582764,
|
| 938 |
+
"rewards/rejected": -5.610608100891113,
|
| 939 |
+
"step": 610
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.757020757020757,
|
| 943 |
+
"grad_norm": 7.703815141377651,
|
| 944 |
+
"learning_rate": 3.9093481747025615e-07,
|
| 945 |
+
"logits/chosen": -1.9850308895111084,
|
| 946 |
+
"logits/rejected": -1.9987919330596924,
|
| 947 |
+
"logps/chosen": -668.91943359375,
|
| 948 |
+
"logps/rejected": -992.9368896484375,
|
| 949 |
+
"loss": 0.2573,
|
| 950 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 951 |
+
"rewards/chosen": -2.7951459884643555,
|
| 952 |
+
"rewards/margins": 3.3580493927001953,
|
| 953 |
+
"rewards/rejected": -6.153195381164551,
|
| 954 |
+
"step": 620
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.7692307692307693,
|
| 958 |
+
"grad_norm": 8.680106719787165,
|
| 959 |
+
"learning_rate": 3.86502179209851e-07,
|
| 960 |
+
"logits/chosen": -1.9663703441619873,
|
| 961 |
+
"logits/rejected": -1.9948896169662476,
|
| 962 |
+
"logps/chosen": -610.5440673828125,
|
| 963 |
+
"logps/rejected": -898.1494140625,
|
| 964 |
+
"loss": 0.2614,
|
| 965 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 966 |
+
"rewards/chosen": -2.289752960205078,
|
| 967 |
+
"rewards/margins": 2.931398868560791,
|
| 968 |
+
"rewards/rejected": -5.221152305603027,
|
| 969 |
+
"step": 630
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.7814407814407814,
|
| 973 |
+
"grad_norm": 10.429529355264094,
|
| 974 |
+
"learning_rate": 3.8200753577654765e-07,
|
| 975 |
+
"logits/chosen": -1.944758653640747,
|
| 976 |
+
"logits/rejected": -1.981778860092163,
|
| 977 |
+
"logps/chosen": -658.198486328125,
|
| 978 |
+
"logps/rejected": -976.310546875,
|
| 979 |
+
"loss": 0.252,
|
| 980 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 981 |
+
"rewards/chosen": -2.5240378379821777,
|
| 982 |
+
"rewards/margins": 3.3471717834472656,
|
| 983 |
+
"rewards/rejected": -5.871209621429443,
|
| 984 |
+
"step": 640
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 0.7936507936507936,
|
| 988 |
+
"grad_norm": 9.854306625880072,
|
| 989 |
+
"learning_rate": 3.7745292883110784e-07,
|
| 990 |
+
"logits/chosen": -1.9690608978271484,
|
| 991 |
+
"logits/rejected": -1.9564428329467773,
|
| 992 |
+
"logps/chosen": -661.0819091796875,
|
| 993 |
+
"logps/rejected": -980.1275634765625,
|
| 994 |
+
"loss": 0.2457,
|
| 995 |
+
"rewards/accuracies": 0.921875,
|
| 996 |
+
"rewards/chosen": -2.6281650066375732,
|
| 997 |
+
"rewards/margins": 3.3420677185058594,
|
| 998 |
+
"rewards/rejected": -5.970232963562012,
|
| 999 |
+
"step": 650
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.8058608058608059,
|
| 1003 |
+
"grad_norm": 10.023665069599845,
|
| 1004 |
+
"learning_rate": 3.7284042727230506e-07,
|
| 1005 |
+
"logits/chosen": -1.9493147134780884,
|
| 1006 |
+
"logits/rejected": -1.9569787979125977,
|
| 1007 |
+
"logps/chosen": -632.8013916015625,
|
| 1008 |
+
"logps/rejected": -883.5313720703125,
|
| 1009 |
+
"loss": 0.2535,
|
| 1010 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 1011 |
+
"rewards/chosen": -2.2697696685791016,
|
| 1012 |
+
"rewards/margins": 2.83363676071167,
|
| 1013 |
+
"rewards/rejected": -5.1034064292907715,
|
| 1014 |
+
"step": 660
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 0.818070818070818,
|
| 1018 |
+
"grad_norm": 8.667472255038128,
|
| 1019 |
+
"learning_rate": 3.681721262971413e-07,
|
| 1020 |
+
"logits/chosen": -2.0272722244262695,
|
| 1021 |
+
"logits/rejected": -2.039168119430542,
|
| 1022 |
+
"logps/chosen": -667.6822509765625,
|
| 1023 |
+
"logps/rejected": -953.4684448242188,
|
| 1024 |
+
"loss": 0.2344,
|
| 1025 |
+
"rewards/accuracies": 0.871874988079071,
|
| 1026 |
+
"rewards/chosen": -2.6374218463897705,
|
| 1027 |
+
"rewards/margins": 3.003962516784668,
|
| 1028 |
+
"rewards/rejected": -5.641384601593018,
|
| 1029 |
+
"step": 670
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.8302808302808303,
|
| 1033 |
+
"grad_norm": 11.27138748410362,
|
| 1034 |
+
"learning_rate": 3.634501464491183e-07,
|
| 1035 |
+
"logits/chosen": -2.0476953983306885,
|
| 1036 |
+
"logits/rejected": -2.064707040786743,
|
| 1037 |
+
"logps/chosen": -683.5264282226562,
|
| 1038 |
+
"logps/rejected": -1021.5324096679688,
|
| 1039 |
+
"loss": 0.233,
|
| 1040 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 1041 |
+
"rewards/chosen": -2.805328845977783,
|
| 1042 |
+
"rewards/margins": 3.454031467437744,
|
| 1043 |
+
"rewards/rejected": -6.259359836578369,
|
| 1044 |
+
"step": 680
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 0.8424908424908425,
|
| 1048 |
+
"grad_norm": 10.056725367187733,
|
| 1049 |
+
"learning_rate": 3.5867663265499553e-07,
|
| 1050 |
+
"logits/chosen": -1.9634895324707031,
|
| 1051 |
+
"logits/rejected": -1.978262186050415,
|
| 1052 |
+
"logps/chosen": -688.0383911132812,
|
| 1053 |
+
"logps/rejected": -1032.1566162109375,
|
| 1054 |
+
"loss": 0.2429,
|
| 1055 |
+
"rewards/accuracies": 0.90625,
|
| 1056 |
+
"rewards/chosen": -2.8448052406311035,
|
| 1057 |
+
"rewards/margins": 3.5894615650177,
|
| 1058 |
+
"rewards/rejected": -6.434267520904541,
|
| 1059 |
+
"step": 690
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.8547008547008547,
|
| 1063 |
+
"grad_norm": 10.779176958531542,
|
| 1064 |
+
"learning_rate": 3.5385375325047163e-07,
|
| 1065 |
+
"logits/chosen": -1.9880058765411377,
|
| 1066 |
+
"logits/rejected": -1.948150396347046,
|
| 1067 |
+
"logps/chosen": -678.7589111328125,
|
| 1068 |
+
"logps/rejected": -950.1603393554688,
|
| 1069 |
+
"loss": 0.2498,
|
| 1070 |
+
"rewards/accuracies": 0.875,
|
| 1071 |
+
"rewards/chosen": -2.7791597843170166,
|
| 1072 |
+
"rewards/margins": 2.923888921737671,
|
| 1073 |
+
"rewards/rejected": -5.7030487060546875,
|
| 1074 |
+
"step": 700
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.8669108669108669,
|
| 1078 |
+
"grad_norm": 7.2521850880280665,
|
| 1079 |
+
"learning_rate": 3.4898369899523323e-07,
|
| 1080 |
+
"logits/chosen": -1.9596385955810547,
|
| 1081 |
+
"logits/rejected": -1.9468225240707397,
|
| 1082 |
+
"logps/chosen": -670.3238525390625,
|
| 1083 |
+
"logps/rejected": -978.1044921875,
|
| 1084 |
+
"loss": 0.245,
|
| 1085 |
+
"rewards/accuracies": 0.934374988079071,
|
| 1086 |
+
"rewards/chosen": -2.5701966285705566,
|
| 1087 |
+
"rewards/margins": 3.3597824573516846,
|
| 1088 |
+
"rewards/rejected": -5.92997932434082,
|
| 1089 |
+
"step": 710
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.8791208791208791,
|
| 1093 |
+
"grad_norm": 8.740369578606884,
|
| 1094 |
+
"learning_rate": 3.4406868207781725e-07,
|
| 1095 |
+
"logits/chosen": -1.9431158304214478,
|
| 1096 |
+
"logits/rejected": -1.9668858051300049,
|
| 1097 |
+
"logps/chosen": -652.6699829101562,
|
| 1098 |
+
"logps/rejected": -978.8201293945312,
|
| 1099 |
+
"loss": 0.252,
|
| 1100 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 1101 |
+
"rewards/chosen": -2.550971746444702,
|
| 1102 |
+
"rewards/margins": 3.3800606727600098,
|
| 1103 |
+
"rewards/rejected": -5.931032180786133,
|
| 1104 |
+
"step": 720
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 0.8913308913308914,
|
| 1108 |
+
"grad_norm": 8.917488139539987,
|
| 1109 |
+
"learning_rate": 3.3911093511073984e-07,
|
| 1110 |
+
"logits/chosen": -1.9911832809448242,
|
| 1111 |
+
"logits/rejected": -2.0098257064819336,
|
| 1112 |
+
"logps/chosen": -659.505859375,
|
| 1113 |
+
"logps/rejected": -971.2593994140625,
|
| 1114 |
+
"loss": 0.2467,
|
| 1115 |
+
"rewards/accuracies": 0.909375011920929,
|
| 1116 |
+
"rewards/chosen": -2.533627986907959,
|
| 1117 |
+
"rewards/margins": 3.233253002166748,
|
| 1118 |
+
"rewards/rejected": -5.766880989074707,
|
| 1119 |
+
"step": 730
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.9035409035409036,
|
| 1123 |
+
"grad_norm": 10.10992640831635,
|
| 1124 |
+
"learning_rate": 3.3411271011634697e-07,
|
| 1125 |
+
"logits/chosen": -1.8714303970336914,
|
| 1126 |
+
"logits/rejected": -1.8533025979995728,
|
| 1127 |
+
"logps/chosen": -610.3406372070312,
|
| 1128 |
+
"logps/rejected": -920.5524291992188,
|
| 1129 |
+
"loss": 0.2487,
|
| 1130 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 1131 |
+
"rewards/chosen": -2.253237247467041,
|
| 1132 |
+
"rewards/margins": 3.4128048419952393,
|
| 1133 |
+
"rewards/rejected": -5.666041374206543,
|
| 1134 |
+
"step": 740
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 0.9157509157509157,
|
| 1138 |
+
"grad_norm": 10.16349038435686,
|
| 1139 |
+
"learning_rate": 3.290762775038494e-07,
|
| 1140 |
+
"logits/chosen": -1.8609752655029297,
|
| 1141 |
+
"logits/rejected": -1.8657268285751343,
|
| 1142 |
+
"logps/chosen": -650.5752563476562,
|
| 1143 |
+
"logps/rejected": -990.2158203125,
|
| 1144 |
+
"loss": 0.2209,
|
| 1145 |
+
"rewards/accuracies": 0.918749988079071,
|
| 1146 |
+
"rewards/chosen": -2.499966859817505,
|
| 1147 |
+
"rewards/margins": 3.5357258319854736,
|
| 1148 |
+
"rewards/rejected": -6.0356926918029785,
|
| 1149 |
+
"step": 750
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.927960927960928,
|
| 1153 |
+
"grad_norm": 7.957411909596094,
|
| 1154 |
+
"learning_rate": 3.2400392503800477e-07,
|
| 1155 |
+
"logits/chosen": -1.904550313949585,
|
| 1156 |
+
"logits/rejected": -1.8870503902435303,
|
| 1157 |
+
"logps/chosen": -671.4937744140625,
|
| 1158 |
+
"logps/rejected": -1035.06005859375,
|
| 1159 |
+
"loss": 0.2278,
|
| 1160 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1161 |
+
"rewards/chosen": -2.7466225624084473,
|
| 1162 |
+
"rewards/margins": 3.689298629760742,
|
| 1163 |
+
"rewards/rejected": -6.435920715332031,
|
| 1164 |
+
"step": 760
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.9401709401709402,
|
| 1168 |
+
"grad_norm": 10.052468847497332,
|
| 1169 |
+
"learning_rate": 3.188979567999161e-07,
|
| 1170 |
+
"logits/chosen": -1.846875786781311,
|
| 1171 |
+
"logits/rejected": -1.8232959508895874,
|
| 1172 |
+
"logps/chosen": -650.9820556640625,
|
| 1173 |
+
"logps/rejected": -956.2470703125,
|
| 1174 |
+
"loss": 0.2315,
|
| 1175 |
+
"rewards/accuracies": 0.90625,
|
| 1176 |
+
"rewards/chosen": -2.5622196197509766,
|
| 1177 |
+
"rewards/margins": 3.281222105026245,
|
| 1178 |
+
"rewards/rejected": -5.843441963195801,
|
| 1179 |
+
"step": 770
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.9523809523809523,
|
| 1183 |
+
"grad_norm": 8.845964368891687,
|
| 1184 |
+
"learning_rate": 3.137606921404191e-07,
|
| 1185 |
+
"logits/chosen": -1.8606138229370117,
|
| 1186 |
+
"logits/rejected": -1.8387963771820068,
|
| 1187 |
+
"logps/chosen": -645.6654663085938,
|
| 1188 |
+
"logps/rejected": -939.6808471679688,
|
| 1189 |
+
"loss": 0.2448,
|
| 1190 |
+
"rewards/accuracies": 0.909375011920929,
|
| 1191 |
+
"rewards/chosen": -2.5100626945495605,
|
| 1192 |
+
"rewards/margins": 3.1421546936035156,
|
| 1193 |
+
"rewards/rejected": -5.652216911315918,
|
| 1194 |
+
"step": 780
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 0.9645909645909646,
|
| 1198 |
+
"grad_norm": 8.337030706793822,
|
| 1199 |
+
"learning_rate": 3.0859446462653273e-07,
|
| 1200 |
+
"logits/chosen": -1.8315823078155518,
|
| 1201 |
+
"logits/rejected": -1.8353532552719116,
|
| 1202 |
+
"logps/chosen": -647.4951782226562,
|
| 1203 |
+
"logps/rejected": -976.1868286132812,
|
| 1204 |
+
"loss": 0.2333,
|
| 1205 |
+
"rewards/accuracies": 0.921875,
|
| 1206 |
+
"rewards/chosen": -2.6538143157958984,
|
| 1207 |
+
"rewards/margins": 3.3457908630371094,
|
| 1208 |
+
"rewards/rejected": -5.999605178833008,
|
| 1209 |
+
"step": 790
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 0.9768009768009768,
|
| 1213 |
+
"grad_norm": 12.589579709309238,
|
| 1214 |
+
"learning_rate": 3.034016209814529e-07,
|
| 1215 |
+
"logits/chosen": -1.7978140115737915,
|
| 1216 |
+
"logits/rejected": -1.8211300373077393,
|
| 1217 |
+
"logps/chosen": -673.1063232421875,
|
| 1218 |
+
"logps/rejected": -1035.5140380859375,
|
| 1219 |
+
"loss": 0.2326,
|
| 1220 |
+
"rewards/accuracies": 0.90625,
|
| 1221 |
+
"rewards/chosen": -2.7855618000030518,
|
| 1222 |
+
"rewards/margins": 3.614232301712036,
|
| 1223 |
+
"rewards/rejected": -6.399794101715088,
|
| 1224 |
+
"step": 800
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.989010989010989,
|
| 1228 |
+
"grad_norm": 8.693339105078397,
|
| 1229 |
+
"learning_rate": 2.9818452001856926e-07,
|
| 1230 |
+
"logits/chosen": -1.937139868736267,
|
| 1231 |
+
"logits/rejected": -1.9104955196380615,
|
| 1232 |
+
"logps/chosen": -709.281982421875,
|
| 1233 |
+
"logps/rejected": -1029.5799560546875,
|
| 1234 |
+
"loss": 0.2363,
|
| 1235 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 1236 |
+
"rewards/chosen": -2.88238263130188,
|
| 1237 |
+
"rewards/margins": 3.4413890838623047,
|
| 1238 |
+
"rewards/rejected": -6.3237714767456055,
|
| 1239 |
+
"step": 810
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 1.0012210012210012,
|
| 1243 |
+
"grad_norm": 7.149313221308795,
|
| 1244 |
+
"learning_rate": 2.929455315699908e-07,
|
| 1245 |
+
"logits/chosen": -1.8502495288848877,
|
| 1246 |
+
"logits/rejected": -1.8472564220428467,
|
| 1247 |
+
"logps/chosen": -633.6524658203125,
|
| 1248 |
+
"logps/rejected": -995.9943237304688,
|
| 1249 |
+
"loss": 0.2329,
|
| 1250 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 1251 |
+
"rewards/chosen": -2.5007681846618652,
|
| 1252 |
+
"rewards/margins": 3.637205123901367,
|
| 1253 |
+
"rewards/rejected": -6.137973308563232,
|
| 1254 |
+
"step": 820
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 1.0134310134310134,
|
| 1258 |
+
"grad_norm": 10.404080063241945,
|
| 1259 |
+
"learning_rate": 2.8768703541006574e-07,
|
| 1260 |
+
"logits/chosen": -1.9154850244522095,
|
| 1261 |
+
"logits/rejected": -1.9522449970245361,
|
| 1262 |
+
"logps/chosen": -641.5855712890625,
|
| 1263 |
+
"logps/rejected": -996.3082275390625,
|
| 1264 |
+
"loss": 0.2282,
|
| 1265 |
+
"rewards/accuracies": 0.903124988079071,
|
| 1266 |
+
"rewards/chosen": -2.4940600395202637,
|
| 1267 |
+
"rewards/margins": 3.6092581748962402,
|
| 1268 |
+
"rewards/rejected": -6.103318214416504,
|
| 1269 |
+
"step": 830
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.0256410256410255,
|
| 1273 |
+
"grad_norm": 9.587186254816348,
|
| 1274 |
+
"learning_rate": 2.8241142017438557e-07,
|
| 1275 |
+
"logits/chosen": -1.873044729232788,
|
| 1276 |
+
"logits/rejected": -1.8523098230361938,
|
| 1277 |
+
"logps/chosen": -667.5242309570312,
|
| 1278 |
+
"logps/rejected": -989.6135864257812,
|
| 1279 |
+
"loss": 0.2229,
|
| 1280 |
+
"rewards/accuracies": 0.90625,
|
| 1281 |
+
"rewards/chosen": -2.665759563446045,
|
| 1282 |
+
"rewards/margins": 3.5281364917755127,
|
| 1283 |
+
"rewards/rejected": -6.193896293640137,
|
| 1284 |
+
"step": 840
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.037851037851038,
|
| 1288 |
+
"grad_norm": 9.366883940315772,
|
| 1289 |
+
"learning_rate": 2.771210822747639e-07,
|
| 1290 |
+
"logits/chosen": -1.9449422359466553,
|
| 1291 |
+
"logits/rejected": -1.8745979070663452,
|
| 1292 |
+
"logps/chosen": -703.3695068359375,
|
| 1293 |
+
"logps/rejected": -1020.8590087890625,
|
| 1294 |
+
"loss": 0.2436,
|
| 1295 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1296 |
+
"rewards/chosen": -2.8462772369384766,
|
| 1297 |
+
"rewards/margins": 3.382256031036377,
|
| 1298 |
+
"rewards/rejected": -6.2285332679748535,
|
| 1299 |
+
"step": 850
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 1.05006105006105,
|
| 1303 |
+
"grad_norm": 7.56593773353816,
|
| 1304 |
+
"learning_rate": 2.718184248106828e-07,
|
| 1305 |
+
"logits/chosen": -1.834773063659668,
|
| 1306 |
+
"logits/rejected": -1.8442004919052124,
|
| 1307 |
+
"logps/chosen": -658.0390014648438,
|
| 1308 |
+
"logps/rejected": -1019.7131958007812,
|
| 1309 |
+
"loss": 0.2224,
|
| 1310 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 1311 |
+
"rewards/chosen": -2.64483380317688,
|
| 1312 |
+
"rewards/margins": 3.6114120483398438,
|
| 1313 |
+
"rewards/rejected": -6.256246089935303,
|
| 1314 |
+
"step": 860
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 1.0622710622710623,
|
| 1318 |
+
"grad_norm": 8.841079507264354,
|
| 1319 |
+
"learning_rate": 2.665058564777014e-07,
|
| 1320 |
+
"logits/chosen": -1.8463929891586304,
|
| 1321 |
+
"logits/rejected": -1.839784026145935,
|
| 1322 |
+
"logps/chosen": -652.1268920898438,
|
| 1323 |
+
"logps/rejected": -1015.1273193359375,
|
| 1324 |
+
"loss": 0.2131,
|
| 1325 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 1326 |
+
"rewards/chosen": -2.7053442001342773,
|
| 1327 |
+
"rewards/margins": 3.6853573322296143,
|
| 1328 |
+
"rewards/rejected": -6.3907012939453125,
|
| 1329 |
+
"step": 870
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 1.0744810744810744,
|
| 1333 |
+
"grad_norm": 9.314872473095937,
|
| 1334 |
+
"learning_rate": 2.611857904733227e-07,
|
| 1335 |
+
"logits/chosen": -1.7795578241348267,
|
| 1336 |
+
"logits/rejected": -1.7798683643341064,
|
| 1337 |
+
"logps/chosen": -688.0113525390625,
|
| 1338 |
+
"logps/rejected": -1035.625244140625,
|
| 1339 |
+
"loss": 0.2242,
|
| 1340 |
+
"rewards/accuracies": 0.909375011920929,
|
| 1341 |
+
"rewards/chosen": -2.99015474319458,
|
| 1342 |
+
"rewards/margins": 3.6745171546936035,
|
| 1343 |
+
"rewards/rejected": -6.664670467376709,
|
| 1344 |
+
"step": 880
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 1.0866910866910866,
|
| 1348 |
+
"grad_norm": 8.334325457508486,
|
| 1349 |
+
"learning_rate": 2.5586064340081516e-07,
|
| 1350 |
+
"logits/chosen": -1.8280951976776123,
|
| 1351 |
+
"logits/rejected": -1.8405084609985352,
|
| 1352 |
+
"logps/chosen": -666.4359741210938,
|
| 1353 |
+
"logps/rejected": -1007.44091796875,
|
| 1354 |
+
"loss": 0.23,
|
| 1355 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 1356 |
+
"rewards/chosen": -2.6223227977752686,
|
| 1357 |
+
"rewards/margins": 3.5808663368225098,
|
| 1358 |
+
"rewards/rejected": -6.203189373016357,
|
| 1359 |
+
"step": 890
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 1.098901098901099,
|
| 1363 |
+
"grad_norm": 7.206748373825152,
|
| 1364 |
+
"learning_rate": 2.505328341714873e-07,
|
| 1365 |
+
"logits/chosen": -1.7787061929702759,
|
| 1366 |
+
"logits/rejected": -1.8012263774871826,
|
| 1367 |
+
"logps/chosen": -663.2637329101562,
|
| 1368 |
+
"logps/rejected": -1036.8768310546875,
|
| 1369 |
+
"loss": 0.2129,
|
| 1370 |
+
"rewards/accuracies": 0.9375,
|
| 1371 |
+
"rewards/chosen": -2.8313677310943604,
|
| 1372 |
+
"rewards/margins": 3.7177510261535645,
|
| 1373 |
+
"rewards/rejected": -6.5491180419921875,
|
| 1374 |
+
"step": 900
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.1111111111111112,
|
| 1378 |
+
"grad_norm": 10.177149465840422,
|
| 1379 |
+
"learning_rate": 2.4520478290591416e-07,
|
| 1380 |
+
"logits/chosen": -1.7504581212997437,
|
| 1381 |
+
"logits/rejected": -1.7739044427871704,
|
| 1382 |
+
"logps/chosen": -702.8826904296875,
|
| 1383 |
+
"logps/rejected": -1059.865234375,
|
| 1384 |
+
"loss": 0.2268,
|
| 1385 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 1386 |
+
"rewards/chosen": -2.962860584259033,
|
| 1387 |
+
"rewards/margins": 3.6696648597717285,
|
| 1388 |
+
"rewards/rejected": -6.6325249671936035,
|
| 1389 |
+
"step": 910
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.1233211233211233,
|
| 1393 |
+
"grad_norm": 8.632975577344476,
|
| 1394 |
+
"learning_rate": 2.3987890983461403e-07,
|
| 1395 |
+
"logits/chosen": -1.7676483392715454,
|
| 1396 |
+
"logits/rejected": -1.7390410900115967,
|
| 1397 |
+
"logps/chosen": -696.1847534179688,
|
| 1398 |
+
"logps/rejected": -1042.147216796875,
|
| 1399 |
+
"loss": 0.2198,
|
| 1400 |
+
"rewards/accuracies": 0.90625,
|
| 1401 |
+
"rewards/chosen": -2.94394850730896,
|
| 1402 |
+
"rewards/margins": 3.708522319793701,
|
| 1403 |
+
"rewards/rejected": -6.652470588684082,
|
| 1404 |
+
"step": 920
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 1.1355311355311355,
|
| 1408 |
+
"grad_norm": 9.916019501072466,
|
| 1409 |
+
"learning_rate": 2.3455763419867544e-07,
|
| 1410 |
+
"logits/chosen": -1.7698522806167603,
|
| 1411 |
+
"logits/rejected": -1.8097903728485107,
|
| 1412 |
+
"logps/chosen": -661.6279907226562,
|
| 1413 |
+
"logps/rejected": -993.2236328125,
|
| 1414 |
+
"loss": 0.2322,
|
| 1415 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 1416 |
+
"rewards/chosen": -2.7180135250091553,
|
| 1417 |
+
"rewards/margins": 3.294512987136841,
|
| 1418 |
+
"rewards/rejected": -6.012526512145996,
|
| 1419 |
+
"step": 930
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 1.1477411477411477,
|
| 1423 |
+
"grad_norm": 8.842361688694279,
|
| 1424 |
+
"learning_rate": 2.2924337315083353e-07,
|
| 1425 |
+
"logits/chosen": -1.782261610031128,
|
| 1426 |
+
"logits/rejected": -1.8005173206329346,
|
| 1427 |
+
"logps/chosen": -626.9920043945312,
|
| 1428 |
+
"logps/rejected": -989.1697387695312,
|
| 1429 |
+
"loss": 0.2168,
|
| 1430 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1431 |
+
"rewards/chosen": -2.3607215881347656,
|
| 1432 |
+
"rewards/margins": 3.6655585765838623,
|
| 1433 |
+
"rewards/rejected": -6.026279926300049,
|
| 1434 |
+
"step": 940
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.1599511599511598,
|
| 1438 |
+
"grad_norm": 9.32555579290425,
|
| 1439 |
+
"learning_rate": 2.239385406574955e-07,
|
| 1440 |
+
"logits/chosen": -1.7827867269515991,
|
| 1441 |
+
"logits/rejected": -1.779762625694275,
|
| 1442 |
+
"logps/chosen": -655.2774047851562,
|
| 1443 |
+
"logps/rejected": -1022.1461791992188,
|
| 1444 |
+
"loss": 0.2233,
|
| 1445 |
+
"rewards/accuracies": 0.940625011920929,
|
| 1446 |
+
"rewards/chosen": -2.592012405395508,
|
| 1447 |
+
"rewards/margins": 3.816871166229248,
|
| 1448 |
+
"rewards/rejected": -6.408883571624756,
|
| 1449 |
+
"step": 950
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 1.1721611721611722,
|
| 1453 |
+
"grad_norm": 10.939107380292038,
|
| 1454 |
+
"learning_rate": 2.1864554640221244e-07,
|
| 1455 |
+
"logits/chosen": -1.7000805139541626,
|
| 1456 |
+
"logits/rejected": -1.688855767250061,
|
| 1457 |
+
"logps/chosen": -676.7574462890625,
|
| 1458 |
+
"logps/rejected": -1019.0291748046875,
|
| 1459 |
+
"loss": 0.2202,
|
| 1460 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 1461 |
+
"rewards/chosen": -2.7764394283294678,
|
| 1462 |
+
"rewards/margins": 3.648998260498047,
|
| 1463 |
+
"rewards/rejected": -6.425436973571777,
|
| 1464 |
+
"step": 960
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 1.1843711843711844,
|
| 1468 |
+
"grad_norm": 10.161202425220864,
|
| 1469 |
+
"learning_rate": 2.133667946910977e-07,
|
| 1470 |
+
"logits/chosen": -1.812464714050293,
|
| 1471 |
+
"logits/rejected": -1.8180309534072876,
|
| 1472 |
+
"logps/chosen": -680.7279052734375,
|
| 1473 |
+
"logps/rejected": -1037.299560546875,
|
| 1474 |
+
"loss": 0.2178,
|
| 1475 |
+
"rewards/accuracies": 0.90625,
|
| 1476 |
+
"rewards/chosen": -2.6717233657836914,
|
| 1477 |
+
"rewards/margins": 3.6986587047576904,
|
| 1478 |
+
"rewards/rejected": -6.370382308959961,
|
| 1479 |
+
"step": 970
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 1.1965811965811965,
|
| 1483 |
+
"grad_norm": 9.089009127660146,
|
| 1484 |
+
"learning_rate": 2.0810468336068697e-07,
|
| 1485 |
+
"logits/chosen": -1.752598524093628,
|
| 1486 |
+
"logits/rejected": -1.7276668548583984,
|
| 1487 |
+
"logps/chosen": -658.7349243164062,
|
| 1488 |
+
"logps/rejected": -977.19580078125,
|
| 1489 |
+
"loss": 0.2208,
|
| 1490 |
+
"rewards/accuracies": 0.90625,
|
| 1491 |
+
"rewards/chosen": -2.660342216491699,
|
| 1492 |
+
"rewards/margins": 3.40667462348938,
|
| 1493 |
+
"rewards/rejected": -6.0670166015625,
|
| 1494 |
+
"step": 980
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.2087912087912087,
|
| 1498 |
+
"grad_norm": 8.82346746989032,
|
| 1499 |
+
"learning_rate": 2.0286160268873826e-07,
|
| 1500 |
+
"logits/chosen": -1.7947756052017212,
|
| 1501 |
+
"logits/rejected": -1.7663764953613281,
|
| 1502 |
+
"logps/chosen": -677.0531005859375,
|
| 1503 |
+
"logps/rejected": -991.953125,
|
| 1504 |
+
"loss": 0.2212,
|
| 1505 |
+
"rewards/accuracies": 0.903124988079071,
|
| 1506 |
+
"rewards/chosen": -2.6913745403289795,
|
| 1507 |
+
"rewards/margins": 3.3568809032440186,
|
| 1508 |
+
"rewards/rejected": -6.048255920410156,
|
| 1509 |
+
"step": 990
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 1.221001221001221,
|
| 1513 |
+
"grad_norm": 8.94368139635424,
|
| 1514 |
+
"learning_rate": 1.9763993430846392e-07,
|
| 1515 |
+
"logits/chosen": -1.691082239151001,
|
| 1516 |
+
"logits/rejected": -1.7195104360580444,
|
| 1517 |
+
"logps/chosen": -689.7127075195312,
|
| 1518 |
+
"logps/rejected": -993.8372802734375,
|
| 1519 |
+
"loss": 0.2293,
|
| 1520 |
+
"rewards/accuracies": 0.893750011920929,
|
| 1521 |
+
"rewards/chosen": -2.915121555328369,
|
| 1522 |
+
"rewards/margins": 3.158756971359253,
|
| 1523 |
+
"rewards/rejected": -6.073878288269043,
|
| 1524 |
+
"step": 1000
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 1.221001221001221,
|
| 1528 |
+
"eval_logits/chosen": -1.457505226135254,
|
| 1529 |
+
"eval_logits/rejected": -1.4398448467254639,
|
| 1530 |
+
"eval_logps/chosen": -689.9403076171875,
|
| 1531 |
+
"eval_logps/rejected": -1048.287109375,
|
| 1532 |
+
"eval_loss": 0.32239073514938354,
|
| 1533 |
+
"eval_rewards/accuracies": 0.90625,
|
| 1534 |
+
"eval_rewards/chosen": -3.3080387115478516,
|
| 1535 |
+
"eval_rewards/margins": 3.7410764694213867,
|
| 1536 |
+
"eval_rewards/rejected": -7.049115180969238,
|
| 1537 |
+
"eval_runtime": 3.2056,
|
| 1538 |
+
"eval_samples_per_second": 62.391,
|
| 1539 |
+
"eval_steps_per_second": 1.248,
|
| 1540 |
+
"step": 1000
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.2332112332112333,
|
| 1544 |
+
"grad_norm": 10.431477967240758,
|
| 1545 |
+
"learning_rate": 1.9244205012669066e-07,
|
| 1546 |
+
"logits/chosen": -1.6947963237762451,
|
| 1547 |
+
"logits/rejected": -1.6955578327178955,
|
| 1548 |
+
"logps/chosen": -682.6785278320312,
|
| 1549 |
+
"logps/rejected": -1043.7890625,
|
| 1550 |
+
"loss": 0.2195,
|
| 1551 |
+
"rewards/accuracies": 0.909375011920929,
|
| 1552 |
+
"rewards/chosen": -2.892595052719116,
|
| 1553 |
+
"rewards/margins": 3.645073652267456,
|
| 1554 |
+
"rewards/rejected": -6.537668704986572,
|
| 1555 |
+
"step": 1010
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.2454212454212454,
|
| 1559 |
+
"grad_norm": 10.45964252043519,
|
| 1560 |
+
"learning_rate": 1.8727031124643738e-07,
|
| 1561 |
+
"logits/chosen": -1.7148557901382446,
|
| 1562 |
+
"logits/rejected": -1.730376958847046,
|
| 1563 |
+
"logps/chosen": -693.7335205078125,
|
| 1564 |
+
"logps/rejected": -1035.718505859375,
|
| 1565 |
+
"loss": 0.2235,
|
| 1566 |
+
"rewards/accuracies": 0.909375011920929,
|
| 1567 |
+
"rewards/chosen": -2.9769744873046875,
|
| 1568 |
+
"rewards/margins": 3.624099016189575,
|
| 1569 |
+
"rewards/rejected": -6.601072788238525,
|
| 1570 |
+
"step": 1020
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.2576312576312576,
|
| 1574 |
+
"grad_norm": 10.46971576068257,
|
| 1575 |
+
"learning_rate": 1.8212706689439993e-07,
|
| 1576 |
+
"logits/chosen": -1.6916805505752563,
|
| 1577 |
+
"logits/rejected": -1.693860411643982,
|
| 1578 |
+
"logps/chosen": -655.6392211914062,
|
| 1579 |
+
"logps/rejected": -1019.2525634765625,
|
| 1580 |
+
"loss": 0.2083,
|
| 1581 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1582 |
+
"rewards/chosen": -2.5844101905822754,
|
| 1583 |
+
"rewards/margins": 3.817833662033081,
|
| 1584 |
+
"rewards/rejected": -6.402244567871094,
|
| 1585 |
+
"step": 1030
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.2698412698412698,
|
| 1589 |
+
"grad_norm": 11.412713931602926,
|
| 1590 |
+
"learning_rate": 1.7701465335383148e-07,
|
| 1591 |
+
"logits/chosen": -1.7352428436279297,
|
| 1592 |
+
"logits/rejected": -1.712531328201294,
|
| 1593 |
+
"logps/chosen": -703.946533203125,
|
| 1594 |
+
"logps/rejected": -1039.840087890625,
|
| 1595 |
+
"loss": 0.2246,
|
| 1596 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 1597 |
+
"rewards/chosen": -2.9772469997406006,
|
| 1598 |
+
"rewards/margins": 3.4660041332244873,
|
| 1599 |
+
"rewards/rejected": -6.443251132965088,
|
| 1600 |
+
"step": 1040
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 1.282051282051282,
|
| 1604 |
+
"grad_norm": 9.59659946541285,
|
| 1605 |
+
"learning_rate": 1.7193539290330172e-07,
|
| 1606 |
+
"logits/chosen": -1.7665666341781616,
|
| 1607 |
+
"logits/rejected": -1.7859532833099365,
|
| 1608 |
+
"logps/chosen": -693.987548828125,
|
| 1609 |
+
"logps/rejected": -1081.337158203125,
|
| 1610 |
+
"loss": 0.2048,
|
| 1611 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1612 |
+
"rewards/chosen": -2.7806649208068848,
|
| 1613 |
+
"rewards/margins": 3.96142578125,
|
| 1614 |
+
"rewards/rejected": -6.742089748382568,
|
| 1615 |
+
"step": 1050
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 1.2942612942612943,
|
| 1619 |
+
"grad_norm": 7.833467318755542,
|
| 1620 |
+
"learning_rate": 1.668915927618183e-07,
|
| 1621 |
+
"logits/chosen": -1.6072509288787842,
|
| 1622 |
+
"logits/rejected": -1.5868537425994873,
|
| 1623 |
+
"logps/chosen": -644.87353515625,
|
| 1624 |
+
"logps/rejected": -992.3282470703125,
|
| 1625 |
+
"loss": 0.2139,
|
| 1626 |
+
"rewards/accuracies": 0.893750011920929,
|
| 1627 |
+
"rewards/chosen": -2.6021721363067627,
|
| 1628 |
+
"rewards/margins": 3.6462604999542236,
|
| 1629 |
+
"rewards/rejected": -6.248432636260986,
|
| 1630 |
+
"step": 1060
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 1.3064713064713065,
|
| 1634 |
+
"grad_norm": 8.944995143399364,
|
| 1635 |
+
"learning_rate": 1.618855440407878e-07,
|
| 1636 |
+
"logits/chosen": -1.6417900323867798,
|
| 1637 |
+
"logits/rejected": -1.645308494567871,
|
| 1638 |
+
"logps/chosen": -680.7823486328125,
|
| 1639 |
+
"logps/rejected": -984.8699951171875,
|
| 1640 |
+
"loss": 0.2206,
|
| 1641 |
+
"rewards/accuracies": 0.878125011920929,
|
| 1642 |
+
"rewards/chosen": -2.845156669616699,
|
| 1643 |
+
"rewards/margins": 3.2548797130584717,
|
| 1644 |
+
"rewards/rejected": -6.100037097930908,
|
| 1645 |
+
"step": 1070
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 1.3186813186813187,
|
| 1649 |
+
"grad_norm": 8.538205860728207,
|
| 1650 |
+
"learning_rate": 1.5691952070329493e-07,
|
| 1651 |
+
"logits/chosen": -1.5984406471252441,
|
| 1652 |
+
"logits/rejected": -1.5525627136230469,
|
| 1653 |
+
"logps/chosen": -696.5716552734375,
|
| 1654 |
+
"logps/rejected": -1034.56640625,
|
| 1655 |
+
"loss": 0.2183,
|
| 1656 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 1657 |
+
"rewards/chosen": -2.9366581439971924,
|
| 1658 |
+
"rewards/margins": 3.6637485027313232,
|
| 1659 |
+
"rewards/rejected": -6.600405693054199,
|
| 1660 |
+
"step": 1080
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 1.3308913308913308,
|
| 1664 |
+
"grad_norm": 12.006671385110291,
|
| 1665 |
+
"learning_rate": 1.519957785311698e-07,
|
| 1666 |
+
"logits/chosen": -1.6169030666351318,
|
| 1667 |
+
"logits/rejected": -1.630112648010254,
|
| 1668 |
+
"logps/chosen": -687.6176147460938,
|
| 1669 |
+
"logps/rejected": -1042.2376708984375,
|
| 1670 |
+
"loss": 0.2274,
|
| 1671 |
+
"rewards/accuracies": 0.940625011920929,
|
| 1672 |
+
"rewards/chosen": -2.9744813442230225,
|
| 1673 |
+
"rewards/margins": 3.7546725273132324,
|
| 1674 |
+
"rewards/rejected": -6.729154109954834,
|
| 1675 |
+
"step": 1090
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.3431013431013432,
|
| 1679 |
+
"grad_norm": 9.403413375132386,
|
| 1680 |
+
"learning_rate": 1.4711655410031536e-07,
|
| 1681 |
+
"logits/chosen": -1.752509355545044,
|
| 1682 |
+
"logits/rejected": -1.7398014068603516,
|
| 1683 |
+
"logps/chosen": -701.8548583984375,
|
| 1684 |
+
"logps/rejected": -1074.74609375,
|
| 1685 |
+
"loss": 0.2107,
|
| 1686 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 1687 |
+
"rewards/chosen": -2.93528413772583,
|
| 1688 |
+
"rewards/margins": 3.7751307487487793,
|
| 1689 |
+
"rewards/rejected": -6.710414886474609,
|
| 1690 |
+
"step": 1100
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.3553113553113554,
|
| 1694 |
+
"grad_norm": 9.067370494527864,
|
| 1695 |
+
"learning_rate": 1.422840637647574e-07,
|
| 1696 |
+
"logits/chosen": -1.6540132761001587,
|
| 1697 |
+
"logits/rejected": -1.6520767211914062,
|
| 1698 |
+
"logps/chosen": -648.3968505859375,
|
| 1699 |
+
"logps/rejected": -1009.6447143554688,
|
| 1700 |
+
"loss": 0.2154,
|
| 1701 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 1702 |
+
"rewards/chosen": -2.6349334716796875,
|
| 1703 |
+
"rewards/margins": 3.7205891609191895,
|
| 1704 |
+
"rewards/rejected": -6.355522632598877,
|
| 1705 |
+
"step": 1110
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 1.3675213675213675,
|
| 1709 |
+
"grad_norm": 8.8218853819329,
|
| 1710 |
+
"learning_rate": 1.3750050264988172e-07,
|
| 1711 |
+
"logits/chosen": -1.599161148071289,
|
| 1712 |
+
"logits/rejected": -1.6590015888214111,
|
| 1713 |
+
"logps/chosen": -683.4578857421875,
|
| 1714 |
+
"logps/rejected": -1025.697998046875,
|
| 1715 |
+
"loss": 0.2113,
|
| 1716 |
+
"rewards/accuracies": 0.918749988079071,
|
| 1717 |
+
"rewards/chosen": -2.938286542892456,
|
| 1718 |
+
"rewards/margins": 3.5185725688934326,
|
| 1719 |
+
"rewards/rejected": -6.456859588623047,
|
| 1720 |
+
"step": 1120
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 1.3797313797313797,
|
| 1724 |
+
"grad_norm": 7.729524060255651,
|
| 1725 |
+
"learning_rate": 1.3276804365531303e-07,
|
| 1726 |
+
"logits/chosen": -1.6903009414672852,
|
| 1727 |
+
"logits/rejected": -1.6954196691513062,
|
| 1728 |
+
"logps/chosen": -684.0424194335938,
|
| 1729 |
+
"logps/rejected": -1034.60205078125,
|
| 1730 |
+
"loss": 0.1965,
|
| 1731 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 1732 |
+
"rewards/chosen": -2.8384480476379395,
|
| 1733 |
+
"rewards/margins": 3.558556079864502,
|
| 1734 |
+
"rewards/rejected": -6.3970046043396,
|
| 1735 |
+
"step": 1130
|
| 1736 |
+
},
|
| 1737 |
+
{
|
| 1738 |
+
"epoch": 1.3919413919413919,
|
| 1739 |
+
"grad_norm": 10.111068058521244,
|
| 1740 |
+
"learning_rate": 1.2808883646789088e-07,
|
| 1741 |
+
"logits/chosen": -1.6597654819488525,
|
| 1742 |
+
"logits/rejected": -1.6594810485839844,
|
| 1743 |
+
"logps/chosen": -678.0111083984375,
|
| 1744 |
+
"logps/rejected": -1045.977783203125,
|
| 1745 |
+
"loss": 0.2093,
|
| 1746 |
+
"rewards/accuracies": 0.9437500238418579,
|
| 1747 |
+
"rewards/chosen": -2.8265321254730225,
|
| 1748 |
+
"rewards/margins": 3.8290627002716064,
|
| 1749 |
+
"rewards/rejected": -6.655595302581787,
|
| 1750 |
+
"step": 1140
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 1.404151404151404,
|
| 1754 |
+
"grad_norm": 9.778269285200919,
|
| 1755 |
+
"learning_rate": 1.2346500658518864e-07,
|
| 1756 |
+
"logits/chosen": -1.6785917282104492,
|
| 1757 |
+
"logits/rejected": -1.6684529781341553,
|
| 1758 |
+
"logps/chosen": -682.1680908203125,
|
| 1759 |
+
"logps/rejected": -1051.3212890625,
|
| 1760 |
+
"loss": 0.2278,
|
| 1761 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1762 |
+
"rewards/chosen": -2.902517080307007,
|
| 1763 |
+
"rewards/margins": 3.8359665870666504,
|
| 1764 |
+
"rewards/rejected": -6.738483428955078,
|
| 1765 |
+
"step": 1150
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 1.4163614163614164,
|
| 1769 |
+
"grad_norm": 8.55576204479013,
|
| 1770 |
+
"learning_rate": 1.1889865435002117e-07,
|
| 1771 |
+
"logits/chosen": -1.7918224334716797,
|
| 1772 |
+
"logits/rejected": -1.7642894983291626,
|
| 1773 |
+
"logps/chosen": -692.8043823242188,
|
| 1774 |
+
"logps/rejected": -1036.125244140625,
|
| 1775 |
+
"loss": 0.2091,
|
| 1776 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 1777 |
+
"rewards/chosen": -2.8053765296936035,
|
| 1778 |
+
"rewards/margins": 3.668884754180908,
|
| 1779 |
+
"rewards/rejected": -6.474261283874512,
|
| 1780 |
+
"step": 1160
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 1.4285714285714286,
|
| 1784 |
+
"grad_norm": 9.250758880138703,
|
| 1785 |
+
"learning_rate": 1.1439185399637888e-07,
|
| 1786 |
+
"logits/chosen": -1.809579849243164,
|
| 1787 |
+
"logits/rejected": -1.820180892944336,
|
| 1788 |
+
"logps/chosen": -689.2313232421875,
|
| 1789 |
+
"logps/rejected": -1069.072021484375,
|
| 1790 |
+
"loss": 0.2072,
|
| 1791 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 1792 |
+
"rewards/chosen": -2.8422112464904785,
|
| 1793 |
+
"rewards/margins": 3.8554039001464844,
|
| 1794 |
+
"rewards/rejected": -6.697615623474121,
|
| 1795 |
+
"step": 1170
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.4407814407814408,
|
| 1799 |
+
"grad_norm": 11.66443113180578,
|
| 1800 |
+
"learning_rate": 1.099466527072207e-07,
|
| 1801 |
+
"logits/chosen": -1.7422889471054077,
|
| 1802 |
+
"logits/rejected": -1.6775966882705688,
|
| 1803 |
+
"logps/chosen": -724.9702758789062,
|
| 1804 |
+
"logps/rejected": -1071.587890625,
|
| 1805 |
+
"loss": 0.2124,
|
| 1806 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 1807 |
+
"rewards/chosen": -3.1015143394470215,
|
| 1808 |
+
"rewards/margins": 3.6922173500061035,
|
| 1809 |
+
"rewards/rejected": -6.793731689453125,
|
| 1810 |
+
"step": 1180
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 1.452991452991453,
|
| 1814 |
+
"grad_norm": 9.357661665183807,
|
| 1815 |
+
"learning_rate": 1.0556506968455556e-07,
|
| 1816 |
+
"logits/chosen": -1.6604530811309814,
|
| 1817 |
+
"logits/rejected": -1.598813533782959,
|
| 1818 |
+
"logps/chosen": -708.178955078125,
|
| 1819 |
+
"logps/rejected": -1078.4857177734375,
|
| 1820 |
+
"loss": 0.2191,
|
| 1821 |
+
"rewards/accuracies": 0.953125,
|
| 1822 |
+
"rewards/chosen": -2.9419569969177246,
|
| 1823 |
+
"rewards/margins": 4.017121315002441,
|
| 1824 |
+
"rewards/rejected": -6.959078311920166,
|
| 1825 |
+
"step": 1190
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 1.4652014652014653,
|
| 1829 |
+
"grad_norm": 8.971265853956833,
|
| 1830 |
+
"learning_rate": 1.0124909523223418e-07,
|
| 1831 |
+
"logits/chosen": -1.8122241497039795,
|
| 1832 |
+
"logits/rejected": -1.7499420642852783,
|
| 1833 |
+
"logps/chosen": -676.4575805664062,
|
| 1834 |
+
"logps/rejected": -1026.417724609375,
|
| 1835 |
+
"loss": 0.226,
|
| 1836 |
+
"rewards/accuracies": 0.903124988079071,
|
| 1837 |
+
"rewards/chosen": -2.6309123039245605,
|
| 1838 |
+
"rewards/margins": 3.7246639728546143,
|
| 1839 |
+
"rewards/rejected": -6.355576038360596,
|
| 1840 |
+
"step": 1200
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 1.4774114774114775,
|
| 1844 |
+
"grad_norm": 11.17454543526871,
|
| 1845 |
+
"learning_rate": 9.700068985186677e-08,
|
| 1846 |
+
"logits/chosen": -1.7836052179336548,
|
| 1847 |
+
"logits/rejected": -1.7842352390289307,
|
| 1848 |
+
"logps/chosen": -679.8863525390625,
|
| 1849 |
+
"logps/rejected": -1052.3326416015625,
|
| 1850 |
+
"loss": 0.2076,
|
| 1851 |
+
"rewards/accuracies": 0.890625,
|
| 1852 |
+
"rewards/chosen": -2.7537055015563965,
|
| 1853 |
+
"rewards/margins": 3.8916175365448,
|
| 1854 |
+
"rewards/rejected": -6.645322322845459,
|
| 1855 |
+
"step": 1210
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 1.4896214896214897,
|
| 1859 |
+
"grad_norm": 8.470663050298606,
|
| 1860 |
+
"learning_rate": 9.282178335227883e-08,
|
| 1861 |
+
"logits/chosen": -1.6586374044418335,
|
| 1862 |
+
"logits/rejected": -1.6548519134521484,
|
| 1863 |
+
"logps/chosen": -673.2451171875,
|
| 1864 |
+
"logps/rejected": -1058.727294921875,
|
| 1865 |
+
"loss": 0.1957,
|
| 1866 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1867 |
+
"rewards/chosen": -2.7371268272399902,
|
| 1868 |
+
"rewards/margins": 3.8924622535705566,
|
| 1869 |
+
"rewards/rejected": -6.629590034484863,
|
| 1870 |
+
"step": 1220
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 1.5018315018315018,
|
| 1874 |
+
"grad_norm": 8.98808268846931,
|
| 1875 |
+
"learning_rate": 8.871427397290893e-08,
|
| 1876 |
+
"logits/chosen": -1.6249803304672241,
|
| 1877 |
+
"logits/rejected": -1.6458851099014282,
|
| 1878 |
+
"logps/chosen": -654.238037109375,
|
| 1879 |
+
"logps/rejected": -1073.6612548828125,
|
| 1880 |
+
"loss": 0.1921,
|
| 1881 |
+
"rewards/accuracies": 0.940625011920929,
|
| 1882 |
+
"rewards/chosen": -2.555995464324951,
|
| 1883 |
+
"rewards/margins": 4.273428440093994,
|
| 1884 |
+
"rewards/rejected": -6.8294243812561035,
|
| 1885 |
+
"step": 1230
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 1.514041514041514,
|
| 1889 |
+
"grad_norm": 10.028503381886956,
|
| 1890 |
+
"learning_rate": 8.468002752154671e-08,
|
| 1891 |
+
"logits/chosen": -1.7111746072769165,
|
| 1892 |
+
"logits/rejected": -1.7021448612213135,
|
| 1893 |
+
"logps/chosen": -696.2750854492188,
|
| 1894 |
+
"logps/rejected": -1062.916748046875,
|
| 1895 |
+
"loss": 0.2054,
|
| 1896 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1897 |
+
"rewards/chosen": -2.8526463508605957,
|
| 1898 |
+
"rewards/margins": 3.87031626701355,
|
| 1899 |
+
"rewards/rejected": -6.722962856292725,
|
| 1900 |
+
"step": 1240
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.5262515262515262,
|
| 1904 |
+
"grad_norm": 9.587289277274007,
|
| 1905 |
+
"learning_rate": 8.07208765268021e-08,
|
| 1906 |
+
"logits/chosen": -1.6548601388931274,
|
| 1907 |
+
"logits/rejected": -1.6278877258300781,
|
| 1908 |
+
"logps/chosen": -686.55126953125,
|
| 1909 |
+
"logps/rejected": -1056.9136962890625,
|
| 1910 |
+
"loss": 0.2031,
|
| 1911 |
+
"rewards/accuracies": 0.909375011920929,
|
| 1912 |
+
"rewards/chosen": -2.857362985610962,
|
| 1913 |
+
"rewards/margins": 3.842681407928467,
|
| 1914 |
+
"rewards/rejected": -6.700045108795166,
|
| 1915 |
+
"step": 1250
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 1.5384615384615383,
|
| 1919 |
+
"grad_norm": 9.113469988640157,
|
| 1920 |
+
"learning_rate": 7.683861940569217e-08,
|
| 1921 |
+
"logits/chosen": -1.6518821716308594,
|
| 1922 |
+
"logits/rejected": -1.6446526050567627,
|
| 1923 |
+
"logps/chosen": -680.0530395507812,
|
| 1924 |
+
"logps/rejected": -1071.1126708984375,
|
| 1925 |
+
"loss": 0.1992,
|
| 1926 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 1927 |
+
"rewards/chosen": -2.8840444087982178,
|
| 1928 |
+
"rewards/margins": 4.099475383758545,
|
| 1929 |
+
"rewards/rejected": -6.983519077301025,
|
| 1930 |
+
"step": 1260
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 1.5506715506715507,
|
| 1934 |
+
"grad_norm": 9.40788034177994,
|
| 1935 |
+
"learning_rate": 7.303501964672246e-08,
|
| 1936 |
+
"logits/chosen": -1.6062475442886353,
|
| 1937 |
+
"logits/rejected": -1.5758030414581299,
|
| 1938 |
+
"logps/chosen": -676.6715087890625,
|
| 1939 |
+
"logps/rejected": -1019.4368896484375,
|
| 1940 |
+
"loss": 0.2274,
|
| 1941 |
+
"rewards/accuracies": 0.875,
|
| 1942 |
+
"rewards/chosen": -2.716747760772705,
|
| 1943 |
+
"rewards/margins": 3.6902427673339844,
|
| 1944 |
+
"rewards/rejected": -6.406990051269531,
|
| 1945 |
+
"step": 1270
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 1.5628815628815629,
|
| 1949 |
+
"grad_norm": 10.234752942213406,
|
| 1950 |
+
"learning_rate": 6.931180500883484e-08,
|
| 1951 |
+
"logits/chosen": -1.6409828662872314,
|
| 1952 |
+
"logits/rejected": -1.6017128229141235,
|
| 1953 |
+
"logps/chosen": -686.1589965820312,
|
| 1954 |
+
"logps/rejected": -1049.4881591796875,
|
| 1955 |
+
"loss": 0.2011,
|
| 1956 |
+
"rewards/accuracies": 0.918749988079071,
|
| 1957 |
+
"rewards/chosen": -2.8092753887176514,
|
| 1958 |
+
"rewards/margins": 3.778611421585083,
|
| 1959 |
+
"rewards/rejected": -6.587886810302734,
|
| 1960 |
+
"step": 1280
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 1.575091575091575,
|
| 1964 |
+
"grad_norm": 10.648822630887103,
|
| 1965 |
+
"learning_rate": 6.567066673658442e-08,
|
| 1966 |
+
"logits/chosen": -1.5854926109313965,
|
| 1967 |
+
"logits/rejected": -1.5248379707336426,
|
| 1968 |
+
"logps/chosen": -675.6094360351562,
|
| 1969 |
+
"logps/rejected": -1013.9142456054688,
|
| 1970 |
+
"loss": 0.2134,
|
| 1971 |
+
"rewards/accuracies": 0.871874988079071,
|
| 1972 |
+
"rewards/chosen": -2.8324193954467773,
|
| 1973 |
+
"rewards/margins": 3.5213024616241455,
|
| 1974 |
+
"rewards/rejected": -6.353722095489502,
|
| 1975 |
+
"step": 1290
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 1.5873015873015874,
|
| 1979 |
+
"grad_norm": 10.089650308836681,
|
| 1980 |
+
"learning_rate": 6.21132587919036e-08,
|
| 1981 |
+
"logits/chosen": -1.593980312347412,
|
| 1982 |
+
"logits/rejected": -1.577946662902832,
|
| 1983 |
+
"logps/chosen": -666.9020385742188,
|
| 1984 |
+
"logps/rejected": -1011.3902587890625,
|
| 1985 |
+
"loss": 0.2021,
|
| 1986 |
+
"rewards/accuracies": 0.893750011920929,
|
| 1987 |
+
"rewards/chosen": -2.7358155250549316,
|
| 1988 |
+
"rewards/margins": 3.5997562408447266,
|
| 1989 |
+
"rewards/rejected": -6.335571765899658,
|
| 1990 |
+
"step": 1300
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 1.5995115995115996,
|
| 1994 |
+
"grad_norm": 10.857356047559415,
|
| 1995 |
+
"learning_rate": 5.864119710280158e-08,
|
| 1996 |
+
"logits/chosen": -1.6040117740631104,
|
| 1997 |
+
"logits/rejected": -1.6094341278076172,
|
| 1998 |
+
"logps/chosen": -694.5184326171875,
|
| 1999 |
+
"logps/rejected": -1058.58740234375,
|
| 2000 |
+
"loss": 0.2002,
|
| 2001 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 2002 |
+
"rewards/chosen": -2.9541563987731934,
|
| 2003 |
+
"rewards/margins": 3.760449171066284,
|
| 2004 |
+
"rewards/rejected": -6.714605808258057,
|
| 2005 |
+
"step": 1310
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 1.6117216117216118,
|
| 2009 |
+
"grad_norm": 14.265677703397381,
|
| 2010 |
+
"learning_rate": 5.525605882933965e-08,
|
| 2011 |
+
"logits/chosen": -1.623427391052246,
|
| 2012 |
+
"logits/rejected": -1.6019861698150635,
|
| 2013 |
+
"logps/chosen": -692.2745971679688,
|
| 2014 |
+
"logps/rejected": -1085.745361328125,
|
| 2015 |
+
"loss": 0.1973,
|
| 2016 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2017 |
+
"rewards/chosen": -2.8990519046783447,
|
| 2018 |
+
"rewards/margins": 4.028355598449707,
|
| 2019 |
+
"rewards/rejected": -6.927407741546631,
|
| 2020 |
+
"step": 1320
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 1.623931623931624,
|
| 2024 |
+
"grad_norm": 9.076988920336841,
|
| 2025 |
+
"learning_rate": 5.1959381647217665e-08,
|
| 2026 |
+
"logits/chosen": -1.5620160102844238,
|
| 2027 |
+
"logits/rejected": -1.5626099109649658,
|
| 2028 |
+
"logps/chosen": -688.0601806640625,
|
| 2029 |
+
"logps/rejected": -1034.831298828125,
|
| 2030 |
+
"loss": 0.205,
|
| 2031 |
+
"rewards/accuracies": 0.90625,
|
| 2032 |
+
"rewards/chosen": -2.8957722187042236,
|
| 2033 |
+
"rewards/margins": 3.635453701019287,
|
| 2034 |
+
"rewards/rejected": -6.53122615814209,
|
| 2035 |
+
"step": 1330
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 1.636141636141636,
|
| 2039 |
+
"grad_norm": 11.741004645166157,
|
| 2040 |
+
"learning_rate": 4.875266304929496e-08,
|
| 2041 |
+
"logits/chosen": -1.5946216583251953,
|
| 2042 |
+
"logits/rejected": -1.552394151687622,
|
| 2043 |
+
"logps/chosen": -699.3309326171875,
|
| 2044 |
+
"logps/rejected": -1056.736083984375,
|
| 2045 |
+
"loss": 0.203,
|
| 2046 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 2047 |
+
"rewards/chosen": -3.076303005218506,
|
| 2048 |
+
"rewards/margins": 3.665733814239502,
|
| 2049 |
+
"rewards/rejected": -6.74203634262085,
|
| 2050 |
+
"step": 1340
|
| 2051 |
+
},
|
| 2052 |
+
{
|
| 2053 |
+
"epoch": 1.6483516483516483,
|
| 2054 |
+
"grad_norm": 11.38093239647967,
|
| 2055 |
+
"learning_rate": 4.5637359665365025e-08,
|
| 2056 |
+
"logits/chosen": -1.5630689859390259,
|
| 2057 |
+
"logits/rejected": -1.5610148906707764,
|
| 2058 |
+
"logps/chosen": -688.2283325195312,
|
| 2059 |
+
"logps/rejected": -1071.9249267578125,
|
| 2060 |
+
"loss": 0.1984,
|
| 2061 |
+
"rewards/accuracies": 0.921875,
|
| 2062 |
+
"rewards/chosen": -2.906219244003296,
|
| 2063 |
+
"rewards/margins": 3.8943076133728027,
|
| 2064 |
+
"rewards/rejected": -6.8005266189575195,
|
| 2065 |
+
"step": 1350
|
| 2066 |
+
},
|
| 2067 |
+
{
|
| 2068 |
+
"epoch": 1.6605616605616604,
|
| 2069 |
+
"grad_norm": 9.41029423304621,
|
| 2070 |
+
"learning_rate": 4.2614886600491115e-08,
|
| 2071 |
+
"logits/chosen": -1.521751046180725,
|
| 2072 |
+
"logits/rejected": -1.528058409690857,
|
| 2073 |
+
"logps/chosen": -666.5167236328125,
|
| 2074 |
+
"logps/rejected": -1023.7020263671875,
|
| 2075 |
+
"loss": 0.2189,
|
| 2076 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2077 |
+
"rewards/chosen": -2.8758678436279297,
|
| 2078 |
+
"rewards/margins": 3.6217830181121826,
|
| 2079 |
+
"rewards/rejected": -6.497651100158691,
|
| 2080 |
+
"step": 1360
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 1.6727716727716728,
|
| 2084 |
+
"grad_norm": 10.63202381564192,
|
| 2085 |
+
"learning_rate": 3.968661679220467e-08,
|
| 2086 |
+
"logits/chosen": -1.7034133672714233,
|
| 2087 |
+
"logits/rejected": -1.7073980569839478,
|
| 2088 |
+
"logps/chosen": -703.4854736328125,
|
| 2089 |
+
"logps/rejected": -1088.0830078125,
|
| 2090 |
+
"loss": 0.2113,
|
| 2091 |
+
"rewards/accuracies": 0.878125011920929,
|
| 2092 |
+
"rewards/chosen": -3.028465747833252,
|
| 2093 |
+
"rewards/margins": 3.8890113830566406,
|
| 2094 |
+
"rewards/rejected": -6.917477607727051,
|
| 2095 |
+
"step": 1370
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 1.684981684981685,
|
| 2099 |
+
"grad_norm": 8.59601070770463,
|
| 2100 |
+
"learning_rate": 3.685388038685811e-08,
|
| 2101 |
+
"logits/chosen": -1.572533369064331,
|
| 2102 |
+
"logits/rejected": -1.5645427703857422,
|
| 2103 |
+
"logps/chosen": -689.0482177734375,
|
| 2104 |
+
"logps/rejected": -1043.1580810546875,
|
| 2105 |
+
"loss": 0.1907,
|
| 2106 |
+
"rewards/accuracies": 0.918749988079071,
|
| 2107 |
+
"rewards/chosen": -2.8478474617004395,
|
| 2108 |
+
"rewards/margins": 3.773597240447998,
|
| 2109 |
+
"rewards/rejected": -6.621443748474121,
|
| 2110 |
+
"step": 1380
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.6971916971916972,
|
| 2114 |
+
"grad_norm": 8.096290604656218,
|
| 2115 |
+
"learning_rate": 3.41179641354146e-08,
|
| 2116 |
+
"logits/chosen": -1.5869381427764893,
|
| 2117 |
+
"logits/rejected": -1.5721813440322876,
|
| 2118 |
+
"logps/chosen": -709.0938110351562,
|
| 2119 |
+
"logps/rejected": -1064.73681640625,
|
| 2120 |
+
"loss": 0.2101,
|
| 2121 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2122 |
+
"rewards/chosen": -3.0553669929504395,
|
| 2123 |
+
"rewards/margins": 3.6293067932128906,
|
| 2124 |
+
"rewards/rejected": -6.684674263000488,
|
| 2125 |
+
"step": 1390
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 1.7094017094017095,
|
| 2129 |
+
"grad_norm": 10.34775868024534,
|
| 2130 |
+
"learning_rate": 3.1480110808950746e-08,
|
| 2131 |
+
"logits/chosen": -1.6079189777374268,
|
| 2132 |
+
"logits/rejected": -1.5532805919647217,
|
| 2133 |
+
"logps/chosen": -701.6239624023438,
|
| 2134 |
+
"logps/rejected": -1058.663330078125,
|
| 2135 |
+
"loss": 0.202,
|
| 2136 |
+
"rewards/accuracies": 0.921875,
|
| 2137 |
+
"rewards/chosen": -2.9393839836120605,
|
| 2138 |
+
"rewards/margins": 3.915294647216797,
|
| 2139 |
+
"rewards/rejected": -6.854678153991699,
|
| 2140 |
+
"step": 1400
|
| 2141 |
+
},
|
| 2142 |
+
{
|
| 2143 |
+
"epoch": 1.7216117216117217,
|
| 2144 |
+
"grad_norm": 8.551636954434649,
|
| 2145 |
+
"learning_rate": 2.8941518634136047e-08,
|
| 2146 |
+
"logits/chosen": -1.5729877948760986,
|
| 2147 |
+
"logits/rejected": -1.5950627326965332,
|
| 2148 |
+
"logps/chosen": -667.7506713867188,
|
| 2149 |
+
"logps/rejected": -1075.386962890625,
|
| 2150 |
+
"loss": 0.2036,
|
| 2151 |
+
"rewards/accuracies": 0.921875,
|
| 2152 |
+
"rewards/chosen": -2.7873880863189697,
|
| 2153 |
+
"rewards/margins": 4.09014892578125,
|
| 2154 |
+
"rewards/rejected": -6.877536773681641,
|
| 2155 |
+
"step": 1410
|
| 2156 |
+
},
|
| 2157 |
+
{
|
| 2158 |
+
"epoch": 1.7338217338217339,
|
| 2159 |
+
"grad_norm": 11.293363203538648,
|
| 2160 |
+
"learning_rate": 2.6503340748947083e-08,
|
| 2161 |
+
"logits/chosen": -1.6414867639541626,
|
| 2162 |
+
"logits/rejected": -1.6144354343414307,
|
| 2163 |
+
"logps/chosen": -710.8462524414062,
|
| 2164 |
+
"logps/rejected": -1076.911865234375,
|
| 2165 |
+
"loss": 0.2096,
|
| 2166 |
+
"rewards/accuracies": 0.934374988079071,
|
| 2167 |
+
"rewards/chosen": -3.167412519454956,
|
| 2168 |
+
"rewards/margins": 3.803969621658325,
|
| 2169 |
+
"rewards/rejected": -6.971382141113281,
|
| 2170 |
+
"step": 1420
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 1.746031746031746,
|
| 2174 |
+
"grad_norm": 9.537502992918926,
|
| 2175 |
+
"learning_rate": 2.4166684678862208e-08,
|
| 2176 |
+
"logits/chosen": -1.5303064584732056,
|
| 2177 |
+
"logits/rejected": -1.523916482925415,
|
| 2178 |
+
"logps/chosen": -673.16748046875,
|
| 2179 |
+
"logps/rejected": -1046.1734619140625,
|
| 2180 |
+
"loss": 0.2136,
|
| 2181 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 2182 |
+
"rewards/chosen": -2.9092774391174316,
|
| 2183 |
+
"rewards/margins": 3.6622211933135986,
|
| 2184 |
+
"rewards/rejected": -6.571497917175293,
|
| 2185 |
+
"step": 1430
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 1.7582417582417582,
|
| 2189 |
+
"grad_norm": 11.313708406928527,
|
| 2190 |
+
"learning_rate": 2.1932611833775843e-08,
|
| 2191 |
+
"logits/chosen": -1.6144897937774658,
|
| 2192 |
+
"logits/rejected": -1.5737859010696411,
|
| 2193 |
+
"logps/chosen": -713.6920166015625,
|
| 2194 |
+
"logps/rejected": -1063.091064453125,
|
| 2195 |
+
"loss": 0.2143,
|
| 2196 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2197 |
+
"rewards/chosen": -3.0494279861450195,
|
| 2198 |
+
"rewards/margins": 3.7339680194854736,
|
| 2199 |
+
"rewards/rejected": -6.7833967208862305,
|
| 2200 |
+
"step": 1440
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 1.7704517704517704,
|
| 2204 |
+
"grad_norm": 12.659032905074808,
|
| 2205 |
+
"learning_rate": 1.9802137025860394e-08,
|
| 2206 |
+
"logits/chosen": -1.541495680809021,
|
| 2207 |
+
"logits/rejected": -1.508460283279419,
|
| 2208 |
+
"logps/chosen": -672.1906127929688,
|
| 2209 |
+
"logps/rejected": -1020.8779296875,
|
| 2210 |
+
"loss": 0.2044,
|
| 2211 |
+
"rewards/accuracies": 0.9375,
|
| 2212 |
+
"rewards/chosen": -2.761587142944336,
|
| 2213 |
+
"rewards/margins": 3.6672203540802,
|
| 2214 |
+
"rewards/rejected": -6.428807258605957,
|
| 2215 |
+
"step": 1450
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 1.7826617826617825,
|
| 2219 |
+
"grad_norm": 8.561486281617269,
|
| 2220 |
+
"learning_rate": 1.7776228008594962e-08,
|
| 2221 |
+
"logits/chosen": -1.5731874704360962,
|
| 2222 |
+
"logits/rejected": -1.6000001430511475,
|
| 2223 |
+
"logps/chosen": -699.12060546875,
|
| 2224 |
+
"logps/rejected": -1043.0130615234375,
|
| 2225 |
+
"loss": 0.1974,
|
| 2226 |
+
"rewards/accuracies": 0.909375011920929,
|
| 2227 |
+
"rewards/chosen": -2.973306894302368,
|
| 2228 |
+
"rewards/margins": 3.508462429046631,
|
| 2229 |
+
"rewards/rejected": -6.481769561767578,
|
| 2230 |
+
"step": 1460
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 1.7948717948717947,
|
| 2234 |
+
"grad_norm": 9.72206499571952,
|
| 2235 |
+
"learning_rate": 1.5855805037169682e-08,
|
| 2236 |
+
"logits/chosen": -1.6329854726791382,
|
| 2237 |
+
"logits/rejected": -1.6412932872772217,
|
| 2238 |
+
"logps/chosen": -683.8406372070312,
|
| 2239 |
+
"logps/rejected": -1081.219482421875,
|
| 2240 |
+
"loss": 0.2127,
|
| 2241 |
+
"rewards/accuracies": 0.903124988079071,
|
| 2242 |
+
"rewards/chosen": -2.91072678565979,
|
| 2243 |
+
"rewards/margins": 4.06986665725708,
|
| 2244 |
+
"rewards/rejected": -6.980593681335449,
|
| 2245 |
+
"step": 1470
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 1.807081807081807,
|
| 2249 |
+
"grad_norm": 9.226812102547623,
|
| 2250 |
+
"learning_rate": 1.4041740450466383e-08,
|
| 2251 |
+
"logits/chosen": -1.667553186416626,
|
| 2252 |
+
"logits/rejected": -1.6831690073013306,
|
| 2253 |
+
"logps/chosen": -684.6526489257812,
|
| 2254 |
+
"logps/rejected": -1098.877197265625,
|
| 2255 |
+
"loss": 0.2142,
|
| 2256 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 2257 |
+
"rewards/chosen": -2.8064825534820557,
|
| 2258 |
+
"rewards/margins": 4.242220878601074,
|
| 2259 |
+
"rewards/rejected": -7.048703670501709,
|
| 2260 |
+
"step": 1480
|
| 2261 |
+
},
|
| 2262 |
+
{
|
| 2263 |
+
"epoch": 1.8192918192918193,
|
| 2264 |
+
"grad_norm": 10.957646684560045,
|
| 2265 |
+
"learning_rate": 1.2334858274804655e-08,
|
| 2266 |
+
"logits/chosen": -1.5813372135162354,
|
| 2267 |
+
"logits/rejected": -1.5335392951965332,
|
| 2268 |
+
"logps/chosen": -666.4457397460938,
|
| 2269 |
+
"logps/rejected": -1035.133544921875,
|
| 2270 |
+
"loss": 0.1955,
|
| 2271 |
+
"rewards/accuracies": 0.953125,
|
| 2272 |
+
"rewards/chosen": -2.7881767749786377,
|
| 2273 |
+
"rewards/margins": 3.741603136062622,
|
| 2274 |
+
"rewards/rejected": -6.529780387878418,
|
| 2275 |
+
"step": 1490
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 1.8315018315018317,
|
| 2279 |
+
"grad_norm": 9.22473440765943,
|
| 2280 |
+
"learning_rate": 1.0735933849633561e-08,
|
| 2281 |
+
"logits/chosen": -1.5356563329696655,
|
| 2282 |
+
"logits/rejected": -1.5444982051849365,
|
| 2283 |
+
"logps/chosen": -656.3917236328125,
|
| 2284 |
+
"logps/rejected": -1014.2698974609375,
|
| 2285 |
+
"loss": 0.2173,
|
| 2286 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2287 |
+
"rewards/chosen": -2.679664134979248,
|
| 2288 |
+
"rewards/margins": 3.69049072265625,
|
| 2289 |
+
"rewards/rejected": -6.370155334472656,
|
| 2290 |
+
"step": 1500
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 1.8437118437118438,
|
| 2294 |
+
"grad_norm": 10.099888495331802,
|
| 2295 |
+
"learning_rate": 9.245693475338906e-09,
|
| 2296 |
+
"logits/chosen": -1.6426986455917358,
|
| 2297 |
+
"logits/rejected": -1.636652946472168,
|
| 2298 |
+
"logps/chosen": -706.7011108398438,
|
| 2299 |
+
"logps/rejected": -1085.6732177734375,
|
| 2300 |
+
"loss": 0.2117,
|
| 2301 |
+
"rewards/accuracies": 0.918749988079071,
|
| 2302 |
+
"rewards/chosen": -2.9595284461975098,
|
| 2303 |
+
"rewards/margins": 3.9240341186523438,
|
| 2304 |
+
"rewards/rejected": -6.8835625648498535,
|
| 2305 |
+
"step": 1510
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 1.855921855921856,
|
| 2309 |
+
"grad_norm": 11.718296845103126,
|
| 2310 |
+
"learning_rate": 7.86481408332651e-09,
|
| 2311 |
+
"logits/chosen": -1.63693368434906,
|
| 2312 |
+
"logits/rejected": -1.6097259521484375,
|
| 2313 |
+
"logps/chosen": -700.6412353515625,
|
| 2314 |
+
"logps/rejected": -1046.5230712890625,
|
| 2315 |
+
"loss": 0.2173,
|
| 2316 |
+
"rewards/accuracies": 0.90625,
|
| 2317 |
+
"rewards/chosen": -3.012932300567627,
|
| 2318 |
+
"rewards/margins": 3.689767837524414,
|
| 2319 |
+
"rewards/rejected": -6.702700614929199,
|
| 2320 |
+
"step": 1520
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 1.8681318681318682,
|
| 2324 |
+
"grad_norm": 11.61323707998988,
|
| 2325 |
+
"learning_rate": 6.593922928530754e-09,
|
| 2326 |
+
"logits/chosen": -1.6446201801300049,
|
| 2327 |
+
"logits/rejected": -1.648411750793457,
|
| 2328 |
+
"logps/chosen": -701.7564697265625,
|
| 2329 |
+
"logps/rejected": -1085.4237060546875,
|
| 2330 |
+
"loss": 0.2141,
|
| 2331 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2332 |
+
"rewards/chosen": -2.927586317062378,
|
| 2333 |
+
"rewards/margins": 3.9655041694641113,
|
| 2334 |
+
"rewards/rejected": -6.893090724945068,
|
| 2335 |
+
"step": 1530
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 1.8803418803418803,
|
| 2339 |
+
"grad_norm": 10.971983293596454,
|
| 2340 |
+
"learning_rate": 5.433597304488113e-09,
|
| 2341 |
+
"logits/chosen": -1.602278709411621,
|
| 2342 |
+
"logits/rejected": -1.5537140369415283,
|
| 2343 |
+
"logps/chosen": -681.44677734375,
|
| 2344 |
+
"logps/rejected": -1042.9771728515625,
|
| 2345 |
+
"loss": 0.2152,
|
| 2346 |
+
"rewards/accuracies": 0.934374988079071,
|
| 2347 |
+
"rewards/chosen": -2.83951997756958,
|
| 2348 |
+
"rewards/margins": 3.8127970695495605,
|
| 2349 |
+
"rewards/rejected": -6.652318000793457,
|
| 2350 |
+
"step": 1540
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 1.8925518925518925,
|
| 2354 |
+
"grad_norm": 9.023723155797857,
|
| 2355 |
+
"learning_rate": 4.384364281105973e-09,
|
| 2356 |
+
"logits/chosen": -1.5327403545379639,
|
| 2357 |
+
"logits/rejected": -1.5472204685211182,
|
| 2358 |
+
"logps/chosen": -680.0806274414062,
|
| 2359 |
+
"logps/rejected": -1091.057861328125,
|
| 2360 |
+
"loss": 0.2025,
|
| 2361 |
+
"rewards/accuracies": 0.90625,
|
| 2362 |
+
"rewards/chosen": -2.8416030406951904,
|
| 2363 |
+
"rewards/margins": 4.13409948348999,
|
| 2364 |
+
"rewards/rejected": -6.97570276260376,
|
| 2365 |
+
"step": 1550
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 1.9047619047619047,
|
| 2369 |
+
"grad_norm": 9.01098707409564,
|
| 2370 |
+
"learning_rate": 3.4467004652442842e-09,
|
| 2371 |
+
"logits/chosen": -1.4875050783157349,
|
| 2372 |
+
"logits/rejected": -1.4984132051467896,
|
| 2373 |
+
"logps/chosen": -679.8426513671875,
|
| 2374 |
+
"logps/rejected": -1051.27392578125,
|
| 2375 |
+
"loss": 0.1913,
|
| 2376 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2377 |
+
"rewards/chosen": -2.8839166164398193,
|
| 2378 |
+
"rewards/margins": 3.8126988410949707,
|
| 2379 |
+
"rewards/rejected": -6.696614742279053,
|
| 2380 |
+
"step": 1560
|
| 2381 |
+
},
|
| 2382 |
+
{
|
| 2383 |
+
"epoch": 1.9169719169719168,
|
| 2384 |
+
"grad_norm": 9.97369900546846,
|
| 2385 |
+
"learning_rate": 2.6210317842206565e-09,
|
| 2386 |
+
"logits/chosen": -1.5761525630950928,
|
| 2387 |
+
"logits/rejected": -1.5815128087997437,
|
| 2388 |
+
"logps/chosen": -681.6040649414062,
|
| 2389 |
+
"logps/rejected": -1054.3304443359375,
|
| 2390 |
+
"loss": 0.2155,
|
| 2391 |
+
"rewards/accuracies": 0.875,
|
| 2392 |
+
"rewards/chosen": -2.8990960121154785,
|
| 2393 |
+
"rewards/margins": 3.823047637939453,
|
| 2394 |
+
"rewards/rejected": -6.722143650054932,
|
| 2395 |
+
"step": 1570
|
| 2396 |
+
},
|
| 2397 |
+
{
|
| 2398 |
+
"epoch": 1.9291819291819292,
|
| 2399 |
+
"grad_norm": 9.735400391893615,
|
| 2400 |
+
"learning_rate": 1.9077332923353728e-09,
|
| 2401 |
+
"logits/chosen": -1.554095983505249,
|
| 2402 |
+
"logits/rejected": -1.505615234375,
|
| 2403 |
+
"logps/chosen": -667.6380004882812,
|
| 2404 |
+
"logps/rejected": -1052.5147705078125,
|
| 2405 |
+
"loss": 0.1958,
|
| 2406 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 2407 |
+
"rewards/chosen": -2.7255098819732666,
|
| 2408 |
+
"rewards/margins": 4.0490617752075195,
|
| 2409 |
+
"rewards/rejected": -6.774571895599365,
|
| 2410 |
+
"step": 1580
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 1.9413919413919414,
|
| 2414 |
+
"grad_norm": 11.38486846510476,
|
| 2415 |
+
"learning_rate": 1.307129000505891e-09,
|
| 2416 |
+
"logits/chosen": -1.550065279006958,
|
| 2417 |
+
"logits/rejected": -1.5272270441055298,
|
| 2418 |
+
"logps/chosen": -644.3756103515625,
|
| 2419 |
+
"logps/rejected": -1003.22314453125,
|
| 2420 |
+
"loss": 0.2359,
|
| 2421 |
+
"rewards/accuracies": 0.921875,
|
| 2422 |
+
"rewards/chosen": -2.641263484954834,
|
| 2423 |
+
"rewards/margins": 3.7743351459503174,
|
| 2424 |
+
"rewards/rejected": -6.415598392486572,
|
| 2425 |
+
"step": 1590
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 1.9536019536019538,
|
| 2429 |
+
"grad_norm": 12.537603692123119,
|
| 2430 |
+
"learning_rate": 8.194917290869907e-10,
|
| 2431 |
+
"logits/chosen": -1.638336420059204,
|
| 2432 |
+
"logits/rejected": -1.6147558689117432,
|
| 2433 |
+
"logps/chosen": -701.1859130859375,
|
| 2434 |
+
"logps/rejected": -1065.903076171875,
|
| 2435 |
+
"loss": 0.2184,
|
| 2436 |
+
"rewards/accuracies": 0.921875,
|
| 2437 |
+
"rewards/chosen": -2.792646884918213,
|
| 2438 |
+
"rewards/margins": 3.8054585456848145,
|
| 2439 |
+
"rewards/rejected": -6.598104953765869,
|
| 2440 |
+
"step": 1600
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 1.965811965811966,
|
| 2444 |
+
"grad_norm": 9.457547390988431,
|
| 2445 |
+
"learning_rate": 4.450429839439884e-10,
|
| 2446 |
+
"logits/chosen": -1.52217435836792,
|
| 2447 |
+
"logits/rejected": -1.5371259450912476,
|
| 2448 |
+
"logps/chosen": -676.3851318359375,
|
| 2449 |
+
"logps/rejected": -1045.146728515625,
|
| 2450 |
+
"loss": 0.2196,
|
| 2451 |
+
"rewards/accuracies": 0.918749988079071,
|
| 2452 |
+
"rewards/chosen": -2.9359381198883057,
|
| 2453 |
+
"rewards/margins": 3.676692247390747,
|
| 2454 |
+
"rewards/rejected": -6.612630367279053,
|
| 2455 |
+
"step": 1610
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 1.978021978021978,
|
| 2459 |
+
"grad_norm": 9.281764750094489,
|
| 2460 |
+
"learning_rate": 1.8395285583530652e-10,
|
| 2461 |
+
"logits/chosen": -1.5917165279388428,
|
| 2462 |
+
"logits/rejected": -1.591761827468872,
|
| 2463 |
+
"logps/chosen": -704.8922729492188,
|
| 2464 |
+
"logps/rejected": -1085.7891845703125,
|
| 2465 |
+
"loss": 0.2179,
|
| 2466 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 2467 |
+
"rewards/chosen": -2.95951509475708,
|
| 2468 |
+
"rewards/margins": 3.9156622886657715,
|
| 2469 |
+
"rewards/rejected": -6.87517786026001,
|
| 2470 |
+
"step": 1620
|
| 2471 |
+
},
|
| 2472 |
+
{
|
| 2473 |
+
"epoch": 1.9902319902319903,
|
| 2474 |
+
"grad_norm": 9.654531338082007,
|
| 2475 |
+
"learning_rate": 3.63399431498046e-11,
|
| 2476 |
+
"logits/chosen": -1.483194351196289,
|
| 2477 |
+
"logits/rejected": -1.4703377485275269,
|
| 2478 |
+
"logps/chosen": -654.5201416015625,
|
| 2479 |
+
"logps/rejected": -1018.1467895507812,
|
| 2480 |
+
"loss": 0.2121,
|
| 2481 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2482 |
+
"rewards/chosen": -2.7714340686798096,
|
| 2483 |
+
"rewards/margins": 3.737165927886963,
|
| 2484 |
+
"rewards/rejected": -6.50860071182251,
|
| 2485 |
+
"step": 1630
|
| 2486 |
+
},
|
| 2487 |
+
{
|
| 2488 |
+
"epoch": 2.0,
|
| 2489 |
+
"step": 1638,
|
| 2490 |
+
"total_flos": 0.0,
|
| 2491 |
+
"train_loss": 0.2817063624168927,
|
| 2492 |
+
"train_runtime": 11455.6006,
|
| 2493 |
+
"train_samples_per_second": 36.602,
|
| 2494 |
+
"train_steps_per_second": 0.143
|
| 2495 |
+
}
|
| 2496 |
+
],
|
| 2497 |
+
"logging_steps": 10,
|
| 2498 |
+
"max_steps": 1638,
|
| 2499 |
+
"num_input_tokens_seen": 0,
|
| 2500 |
+
"num_train_epochs": 2,
|
| 2501 |
+
"save_steps": 100,
|
| 2502 |
+
"stateful_callbacks": {
|
| 2503 |
+
"TrainerControl": {
|
| 2504 |
+
"args": {
|
| 2505 |
+
"should_epoch_stop": false,
|
| 2506 |
+
"should_evaluate": false,
|
| 2507 |
+
"should_log": false,
|
| 2508 |
+
"should_save": true,
|
| 2509 |
+
"should_training_stop": true
|
| 2510 |
+
},
|
| 2511 |
+
"attributes": {}
|
| 2512 |
+
}
|
| 2513 |
+
},
|
| 2514 |
+
"total_flos": 0.0,
|
| 2515 |
+
"train_batch_size": 8,
|
| 2516 |
+
"trial_name": null,
|
| 2517 |
+
"trial_params": null
|
| 2518 |
+
}
|