Model save
Browse files- README.md +78 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +992 -0
README.md
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- dpo
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: zephyr-7b-align-scan-9e-07-0.86-linear-2.0
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# zephyr-7b-align-scan-9e-07-0.86-linear-2.0
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 1.2247
|
| 22 |
+
- Rewards/chosen: 0.1937
|
| 23 |
+
- Rewards/rejected: -1.8484
|
| 24 |
+
- Rewards/accuracies: 0.3214
|
| 25 |
+
- Rewards/margins: 2.0420
|
| 26 |
+
- Logps/rejected: -83.2776
|
| 27 |
+
- Logps/chosen: -74.2660
|
| 28 |
+
- Logits/rejected: -2.5909
|
| 29 |
+
- Logits/chosen: -2.6070
|
| 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: 9e-07
|
| 49 |
+
- train_batch_size: 8
|
| 50 |
+
- eval_batch_size: 8
|
| 51 |
+
- seed: 42
|
| 52 |
+
- distributed_type: multi-GPU
|
| 53 |
+
- num_devices: 4
|
| 54 |
+
- gradient_accumulation_steps: 2
|
| 55 |
+
- total_train_batch_size: 64
|
| 56 |
+
- total_eval_batch_size: 32
|
| 57 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 58 |
+
- lr_scheduler_type: linear
|
| 59 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 60 |
+
- num_epochs: 2
|
| 61 |
+
|
| 62 |
+
### Training results
|
| 63 |
+
|
| 64 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 65 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 66 |
+
| 0.9734 | 0.3484 | 100 | 0.9203 | 2.0652 | 1.3351 | 0.3294 | 0.7302 | -79.5760 | -72.0898 | -2.5695 | -2.5853 |
|
| 67 |
+
| 0.9883 | 0.6969 | 200 | 1.0967 | 2.4271 | 1.1885 | 0.3373 | 1.2386 | -79.7464 | -71.6691 | -2.5708 | -2.5875 |
|
| 68 |
+
| 0.4215 | 1.0453 | 300 | 1.1234 | 3.0876 | 1.7905 | 0.3313 | 1.2970 | -79.0463 | -70.9010 | -2.6403 | -2.6560 |
|
| 69 |
+
| 0.393 | 1.3937 | 400 | 1.2234 | -0.1343 | -1.8934 | 0.3234 | 1.7591 | -83.3299 | -74.6474 | -2.6093 | -2.6250 |
|
| 70 |
+
| 0.3986 | 1.7422 | 500 | 1.2247 | 0.1937 | -1.8484 | 0.3214 | 2.0420 | -83.2776 | -74.2660 | -2.5909 | -2.6070 |
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
### Framework versions
|
| 74 |
+
|
| 75 |
+
- Transformers 4.44.2
|
| 76 |
+
- Pytorch 2.4.0
|
| 77 |
+
- Datasets 2.21.0
|
| 78 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.6794358088995106,
|
| 5 |
+
"train_runtime": 6413.3344,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.719,
|
| 8 |
+
"train_steps_per_second": 0.09
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.44.2"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.6794358088995106,
|
| 5 |
+
"train_runtime": 6413.3344,
|
| 6 |
+
"train_samples": 18340,
|
| 7 |
+
"train_samples_per_second": 5.719,
|
| 8 |
+
"train_steps_per_second": 0.09
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,992 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 574,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.003484320557491289,
|
| 13 |
+
"grad_norm": 497.3575626103625,
|
| 14 |
+
"learning_rate": 1.5517241379310344e-08,
|
| 15 |
+
"logits/chosen": -2.5345611572265625,
|
| 16 |
+
"logits/rejected": -2.581700563430786,
|
| 17 |
+
"logps/chosen": -60.002105712890625,
|
| 18 |
+
"logps/rejected": -99.98374938964844,
|
| 19 |
+
"loss": 0.6931,
|
| 20 |
+
"rewards/accuracies": 0.0,
|
| 21 |
+
"rewards/chosen": 0.0,
|
| 22 |
+
"rewards/margins": 0.0,
|
| 23 |
+
"rewards/rejected": 0.0,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.03484320557491289,
|
| 28 |
+
"grad_norm": 462.4792154122565,
|
| 29 |
+
"learning_rate": 1.5517241379310344e-07,
|
| 30 |
+
"logits/chosen": -2.563889503479004,
|
| 31 |
+
"logits/rejected": -2.562492609024048,
|
| 32 |
+
"logps/chosen": -59.64662551879883,
|
| 33 |
+
"logps/rejected": -73.37532043457031,
|
| 34 |
+
"loss": 0.6974,
|
| 35 |
+
"rewards/accuracies": 0.2083333283662796,
|
| 36 |
+
"rewards/chosen": 0.011204251088202,
|
| 37 |
+
"rewards/margins": 0.00639230664819479,
|
| 38 |
+
"rewards/rejected": 0.004811946302652359,
|
| 39 |
+
"step": 10
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.06968641114982578,
|
| 43 |
+
"grad_norm": 537.1826669031543,
|
| 44 |
+
"learning_rate": 3.103448275862069e-07,
|
| 45 |
+
"logits/chosen": -2.6071362495422363,
|
| 46 |
+
"logits/rejected": -2.565964698791504,
|
| 47 |
+
"logps/chosen": -103.9787826538086,
|
| 48 |
+
"logps/rejected": -94.84336853027344,
|
| 49 |
+
"loss": 0.6775,
|
| 50 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 51 |
+
"rewards/chosen": 0.12751302123069763,
|
| 52 |
+
"rewards/margins": 0.08386950194835663,
|
| 53 |
+
"rewards/rejected": 0.0436435230076313,
|
| 54 |
+
"step": 20
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.10452961672473868,
|
| 58 |
+
"grad_norm": 586.0340590697643,
|
| 59 |
+
"learning_rate": 4.6551724137931035e-07,
|
| 60 |
+
"logits/chosen": -2.5952959060668945,
|
| 61 |
+
"logits/rejected": -2.5755515098571777,
|
| 62 |
+
"logps/chosen": -81.9345932006836,
|
| 63 |
+
"logps/rejected": -91.35636901855469,
|
| 64 |
+
"loss": 0.6673,
|
| 65 |
+
"rewards/accuracies": 0.33125001192092896,
|
| 66 |
+
"rewards/chosen": 0.48329487442970276,
|
| 67 |
+
"rewards/margins": 0.3252241015434265,
|
| 68 |
+
"rewards/rejected": 0.15807083249092102,
|
| 69 |
+
"step": 30
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13937282229965156,
|
| 73 |
+
"grad_norm": 508.2061569040689,
|
| 74 |
+
"learning_rate": 6.206896551724138e-07,
|
| 75 |
+
"logits/chosen": -2.495903491973877,
|
| 76 |
+
"logits/rejected": -2.4941375255584717,
|
| 77 |
+
"logps/chosen": -78.31864929199219,
|
| 78 |
+
"logps/rejected": -74.04932403564453,
|
| 79 |
+
"loss": 0.6704,
|
| 80 |
+
"rewards/accuracies": 0.3187499940395355,
|
| 81 |
+
"rewards/chosen": -0.3442053496837616,
|
| 82 |
+
"rewards/margins": 0.5570718050003052,
|
| 83 |
+
"rewards/rejected": -0.9012771844863892,
|
| 84 |
+
"step": 40
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.17421602787456447,
|
| 88 |
+
"grad_norm": 371.6317110957719,
|
| 89 |
+
"learning_rate": 7.758620689655172e-07,
|
| 90 |
+
"logits/chosen": -2.522507667541504,
|
| 91 |
+
"logits/rejected": -2.5270426273345947,
|
| 92 |
+
"logps/chosen": -63.30371856689453,
|
| 93 |
+
"logps/rejected": -75.92526245117188,
|
| 94 |
+
"loss": 0.7487,
|
| 95 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 96 |
+
"rewards/chosen": 0.5464689135551453,
|
| 97 |
+
"rewards/margins": 0.3932306468486786,
|
| 98 |
+
"rewards/rejected": 0.15323826670646667,
|
| 99 |
+
"step": 50
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.20905923344947736,
|
| 103 |
+
"grad_norm": 345.88352506142644,
|
| 104 |
+
"learning_rate": 8.965116279069767e-07,
|
| 105 |
+
"logits/chosen": -2.493328094482422,
|
| 106 |
+
"logits/rejected": -2.4874939918518066,
|
| 107 |
+
"logps/chosen": -70.6145248413086,
|
| 108 |
+
"logps/rejected": -66.24369049072266,
|
| 109 |
+
"loss": 0.7394,
|
| 110 |
+
"rewards/accuracies": 0.3125,
|
| 111 |
+
"rewards/chosen": 2.2291111946105957,
|
| 112 |
+
"rewards/margins": 0.3673287332057953,
|
| 113 |
+
"rewards/rejected": 1.8617826700210571,
|
| 114 |
+
"step": 60
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.24390243902439024,
|
| 118 |
+
"grad_norm": 461.9022420517148,
|
| 119 |
+
"learning_rate": 8.790697674418603e-07,
|
| 120 |
+
"logits/chosen": -2.5233232975006104,
|
| 121 |
+
"logits/rejected": -2.518942356109619,
|
| 122 |
+
"logps/chosen": -61.38080978393555,
|
| 123 |
+
"logps/rejected": -66.21412658691406,
|
| 124 |
+
"loss": 0.8029,
|
| 125 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 126 |
+
"rewards/chosen": 2.4556429386138916,
|
| 127 |
+
"rewards/margins": 0.5528993606567383,
|
| 128 |
+
"rewards/rejected": 1.9027433395385742,
|
| 129 |
+
"step": 70
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.2787456445993031,
|
| 133 |
+
"grad_norm": 445.0623855333619,
|
| 134 |
+
"learning_rate": 8.616279069767441e-07,
|
| 135 |
+
"logits/chosen": -2.464928388595581,
|
| 136 |
+
"logits/rejected": -2.4558563232421875,
|
| 137 |
+
"logps/chosen": -72.87749481201172,
|
| 138 |
+
"logps/rejected": -75.74545288085938,
|
| 139 |
+
"loss": 0.8699,
|
| 140 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 141 |
+
"rewards/chosen": 2.397139310836792,
|
| 142 |
+
"rewards/margins": 0.8753674626350403,
|
| 143 |
+
"rewards/rejected": 1.521772027015686,
|
| 144 |
+
"step": 80
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.313588850174216,
|
| 148 |
+
"grad_norm": 523.0685685145581,
|
| 149 |
+
"learning_rate": 8.441860465116279e-07,
|
| 150 |
+
"logits/chosen": -2.514057159423828,
|
| 151 |
+
"logits/rejected": -2.529362201690674,
|
| 152 |
+
"logps/chosen": -63.073158264160156,
|
| 153 |
+
"logps/rejected": -67.64947509765625,
|
| 154 |
+
"loss": 0.9667,
|
| 155 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 156 |
+
"rewards/chosen": 2.5018441677093506,
|
| 157 |
+
"rewards/margins": 0.38231611251831055,
|
| 158 |
+
"rewards/rejected": 2.11952805519104,
|
| 159 |
+
"step": 90
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.34843205574912894,
|
| 163 |
+
"grad_norm": 538.3596051073214,
|
| 164 |
+
"learning_rate": 8.267441860465116e-07,
|
| 165 |
+
"logits/chosen": -2.5051705837249756,
|
| 166 |
+
"logits/rejected": -2.5038981437683105,
|
| 167 |
+
"logps/chosen": -72.73633575439453,
|
| 168 |
+
"logps/rejected": -79.80201721191406,
|
| 169 |
+
"loss": 0.9734,
|
| 170 |
+
"rewards/accuracies": 0.3125,
|
| 171 |
+
"rewards/chosen": 2.179957389831543,
|
| 172 |
+
"rewards/margins": 1.1052770614624023,
|
| 173 |
+
"rewards/rejected": 1.0746803283691406,
|
| 174 |
+
"step": 100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.34843205574912894,
|
| 178 |
+
"eval_logits/chosen": -2.5853304862976074,
|
| 179 |
+
"eval_logits/rejected": -2.569460868835449,
|
| 180 |
+
"eval_logps/chosen": -72.08978271484375,
|
| 181 |
+
"eval_logps/rejected": -79.57595825195312,
|
| 182 |
+
"eval_loss": 0.9202933311462402,
|
| 183 |
+
"eval_rewards/accuracies": 0.329365074634552,
|
| 184 |
+
"eval_rewards/chosen": 2.065248489379883,
|
| 185 |
+
"eval_rewards/margins": 0.730187714099884,
|
| 186 |
+
"eval_rewards/rejected": 1.3350608348846436,
|
| 187 |
+
"eval_runtime": 113.407,
|
| 188 |
+
"eval_samples_per_second": 17.636,
|
| 189 |
+
"eval_steps_per_second": 0.556,
|
| 190 |
+
"step": 100
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.3832752613240418,
|
| 194 |
+
"grad_norm": 565.1538822918781,
|
| 195 |
+
"learning_rate": 8.093023255813954e-07,
|
| 196 |
+
"logits/chosen": -2.517366409301758,
|
| 197 |
+
"logits/rejected": -2.486677646636963,
|
| 198 |
+
"logps/chosen": -71.63627624511719,
|
| 199 |
+
"logps/rejected": -61.792930603027344,
|
| 200 |
+
"loss": 0.9796,
|
| 201 |
+
"rewards/accuracies": 0.2562499940395355,
|
| 202 |
+
"rewards/chosen": 1.3190182447433472,
|
| 203 |
+
"rewards/margins": 0.27183371782302856,
|
| 204 |
+
"rewards/rejected": 1.0471845865249634,
|
| 205 |
+
"step": 110
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.4181184668989547,
|
| 209 |
+
"grad_norm": 415.3172615536407,
|
| 210 |
+
"learning_rate": 7.91860465116279e-07,
|
| 211 |
+
"logits/chosen": -2.5475993156433105,
|
| 212 |
+
"logits/rejected": -2.518019437789917,
|
| 213 |
+
"logps/chosen": -75.4908676147461,
|
| 214 |
+
"logps/rejected": -65.90876770019531,
|
| 215 |
+
"loss": 0.9023,
|
| 216 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 217 |
+
"rewards/chosen": 1.6948095560073853,
|
| 218 |
+
"rewards/margins": 1.0387893915176392,
|
| 219 |
+
"rewards/rejected": 0.6560201644897461,
|
| 220 |
+
"step": 120
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.4529616724738676,
|
| 224 |
+
"grad_norm": 654.650382698047,
|
| 225 |
+
"learning_rate": 7.744186046511627e-07,
|
| 226 |
+
"logits/chosen": -2.57270884513855,
|
| 227 |
+
"logits/rejected": -2.5545814037323,
|
| 228 |
+
"logps/chosen": -82.57270812988281,
|
| 229 |
+
"logps/rejected": -87.82828521728516,
|
| 230 |
+
"loss": 1.0166,
|
| 231 |
+
"rewards/accuracies": 0.34375,
|
| 232 |
+
"rewards/chosen": 1.1249371767044067,
|
| 233 |
+
"rewards/margins": 1.3999314308166504,
|
| 234 |
+
"rewards/rejected": -0.2749941945075989,
|
| 235 |
+
"step": 130
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.4878048780487805,
|
| 239 |
+
"grad_norm": 392.83640221764773,
|
| 240 |
+
"learning_rate": 7.569767441860465e-07,
|
| 241 |
+
"logits/chosen": -2.4829936027526855,
|
| 242 |
+
"logits/rejected": -2.474769353866577,
|
| 243 |
+
"logps/chosen": -78.97613525390625,
|
| 244 |
+
"logps/rejected": -70.12017059326172,
|
| 245 |
+
"loss": 0.9314,
|
| 246 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 247 |
+
"rewards/chosen": 1.8655914068222046,
|
| 248 |
+
"rewards/margins": 1.3116776943206787,
|
| 249 |
+
"rewards/rejected": 0.5539135932922363,
|
| 250 |
+
"step": 140
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.5226480836236934,
|
| 254 |
+
"grad_norm": 496.0616334612732,
|
| 255 |
+
"learning_rate": 7.395348837209303e-07,
|
| 256 |
+
"logits/chosen": -2.5784010887145996,
|
| 257 |
+
"logits/rejected": -2.5391345024108887,
|
| 258 |
+
"logps/chosen": -77.13214111328125,
|
| 259 |
+
"logps/rejected": -78.45569610595703,
|
| 260 |
+
"loss": 1.0681,
|
| 261 |
+
"rewards/accuracies": 0.28125,
|
| 262 |
+
"rewards/chosen": 2.022892713546753,
|
| 263 |
+
"rewards/margins": 1.0811107158660889,
|
| 264 |
+
"rewards/rejected": 0.9417816400527954,
|
| 265 |
+
"step": 150
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.5574912891986062,
|
| 269 |
+
"grad_norm": 572.3657252794802,
|
| 270 |
+
"learning_rate": 7.220930232558139e-07,
|
| 271 |
+
"logits/chosen": -2.579862117767334,
|
| 272 |
+
"logits/rejected": -2.5984551906585693,
|
| 273 |
+
"logps/chosen": -62.170875549316406,
|
| 274 |
+
"logps/rejected": -70.7127914428711,
|
| 275 |
+
"loss": 0.996,
|
| 276 |
+
"rewards/accuracies": 0.28125,
|
| 277 |
+
"rewards/chosen": 2.024418354034424,
|
| 278 |
+
"rewards/margins": 0.9407089948654175,
|
| 279 |
+
"rewards/rejected": 1.0837091207504272,
|
| 280 |
+
"step": 160
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.5923344947735192,
|
| 284 |
+
"grad_norm": 568.8903494255211,
|
| 285 |
+
"learning_rate": 7.046511627906976e-07,
|
| 286 |
+
"logits/chosen": -2.5988354682922363,
|
| 287 |
+
"logits/rejected": -2.5848004817962646,
|
| 288 |
+
"logps/chosen": -66.76710510253906,
|
| 289 |
+
"logps/rejected": -75.21239471435547,
|
| 290 |
+
"loss": 0.9987,
|
| 291 |
+
"rewards/accuracies": 0.25,
|
| 292 |
+
"rewards/chosen": 1.8922548294067383,
|
| 293 |
+
"rewards/margins": 0.944841206073761,
|
| 294 |
+
"rewards/rejected": 0.9474137425422668,
|
| 295 |
+
"step": 170
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.627177700348432,
|
| 299 |
+
"grad_norm": 521.2606597836349,
|
| 300 |
+
"learning_rate": 6.872093023255814e-07,
|
| 301 |
+
"logits/chosen": -2.6132636070251465,
|
| 302 |
+
"logits/rejected": -2.602483034133911,
|
| 303 |
+
"logps/chosen": -89.0395736694336,
|
| 304 |
+
"logps/rejected": -84.7410659790039,
|
| 305 |
+
"loss": 1.0952,
|
| 306 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 307 |
+
"rewards/chosen": 2.4231231212615967,
|
| 308 |
+
"rewards/margins": 0.9014472961425781,
|
| 309 |
+
"rewards/rejected": 1.5216760635375977,
|
| 310 |
+
"step": 180
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.662020905923345,
|
| 314 |
+
"grad_norm": 416.32272877425527,
|
| 315 |
+
"learning_rate": 6.697674418604651e-07,
|
| 316 |
+
"logits/chosen": -2.6020960807800293,
|
| 317 |
+
"logits/rejected": -2.594320297241211,
|
| 318 |
+
"logps/chosen": -68.85299682617188,
|
| 319 |
+
"logps/rejected": -79.95862579345703,
|
| 320 |
+
"loss": 1.0472,
|
| 321 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 322 |
+
"rewards/chosen": 1.8906943798065186,
|
| 323 |
+
"rewards/margins": 0.5408729314804077,
|
| 324 |
+
"rewards/rejected": 1.3498214483261108,
|
| 325 |
+
"step": 190
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.6968641114982579,
|
| 329 |
+
"grad_norm": 736.2428171588583,
|
| 330 |
+
"learning_rate": 6.523255813953487e-07,
|
| 331 |
+
"logits/chosen": -2.6261801719665527,
|
| 332 |
+
"logits/rejected": -2.6321043968200684,
|
| 333 |
+
"logps/chosen": -86.841064453125,
|
| 334 |
+
"logps/rejected": -90.11552429199219,
|
| 335 |
+
"loss": 0.9883,
|
| 336 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 337 |
+
"rewards/chosen": 3.2811522483825684,
|
| 338 |
+
"rewards/margins": 1.6169086694717407,
|
| 339 |
+
"rewards/rejected": 1.6642436981201172,
|
| 340 |
+
"step": 200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.6968641114982579,
|
| 344 |
+
"eval_logits/chosen": -2.5874733924865723,
|
| 345 |
+
"eval_logits/rejected": -2.5707614421844482,
|
| 346 |
+
"eval_logps/chosen": -71.6690673828125,
|
| 347 |
+
"eval_logps/rejected": -79.74640655517578,
|
| 348 |
+
"eval_loss": 1.0967082977294922,
|
| 349 |
+
"eval_rewards/accuracies": 0.3373015820980072,
|
| 350 |
+
"eval_rewards/chosen": 2.4270544052124023,
|
| 351 |
+
"eval_rewards/margins": 1.2385817766189575,
|
| 352 |
+
"eval_rewards/rejected": 1.1884726285934448,
|
| 353 |
+
"eval_runtime": 113.2711,
|
| 354 |
+
"eval_samples_per_second": 17.657,
|
| 355 |
+
"eval_steps_per_second": 0.556,
|
| 356 |
+
"step": 200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.7317073170731707,
|
| 360 |
+
"grad_norm": 760.3151343554975,
|
| 361 |
+
"learning_rate": 6.348837209302325e-07,
|
| 362 |
+
"logits/chosen": -2.594675064086914,
|
| 363 |
+
"logits/rejected": -2.570430278778076,
|
| 364 |
+
"logps/chosen": -66.980224609375,
|
| 365 |
+
"logps/rejected": -62.78594207763672,
|
| 366 |
+
"loss": 1.0072,
|
| 367 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 368 |
+
"rewards/chosen": 2.011408805847168,
|
| 369 |
+
"rewards/margins": 1.4795125722885132,
|
| 370 |
+
"rewards/rejected": 0.5318960547447205,
|
| 371 |
+
"step": 210
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.7665505226480837,
|
| 375 |
+
"grad_norm": 481.77954463638264,
|
| 376 |
+
"learning_rate": 6.174418604651163e-07,
|
| 377 |
+
"logits/chosen": -2.645392894744873,
|
| 378 |
+
"logits/rejected": -2.6270740032196045,
|
| 379 |
+
"logps/chosen": -70.76057434082031,
|
| 380 |
+
"logps/rejected": -69.80204772949219,
|
| 381 |
+
"loss": 1.1057,
|
| 382 |
+
"rewards/accuracies": 0.24375000596046448,
|
| 383 |
+
"rewards/chosen": 2.319939136505127,
|
| 384 |
+
"rewards/margins": 0.7193705439567566,
|
| 385 |
+
"rewards/rejected": 1.6005685329437256,
|
| 386 |
+
"step": 220
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.8013937282229965,
|
| 390 |
+
"grad_norm": 765.046306381631,
|
| 391 |
+
"learning_rate": 6e-07,
|
| 392 |
+
"logits/chosen": -2.6732935905456543,
|
| 393 |
+
"logits/rejected": -2.65371036529541,
|
| 394 |
+
"logps/chosen": -86.65088653564453,
|
| 395 |
+
"logps/rejected": -87.35969543457031,
|
| 396 |
+
"loss": 1.1788,
|
| 397 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 398 |
+
"rewards/chosen": 2.8807787895202637,
|
| 399 |
+
"rewards/margins": 2.20210599899292,
|
| 400 |
+
"rewards/rejected": 0.6786726713180542,
|
| 401 |
+
"step": 230
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.8362369337979094,
|
| 405 |
+
"grad_norm": 548.8932956260087,
|
| 406 |
+
"learning_rate": 5.825581395348838e-07,
|
| 407 |
+
"logits/chosen": -2.6739866733551025,
|
| 408 |
+
"logits/rejected": -2.6417675018310547,
|
| 409 |
+
"logps/chosen": -83.05066680908203,
|
| 410 |
+
"logps/rejected": -77.82533264160156,
|
| 411 |
+
"loss": 1.1372,
|
| 412 |
+
"rewards/accuracies": 0.34375,
|
| 413 |
+
"rewards/chosen": 3.1742072105407715,
|
| 414 |
+
"rewards/margins": 1.1869542598724365,
|
| 415 |
+
"rewards/rejected": 1.9872528314590454,
|
| 416 |
+
"step": 240
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.8710801393728222,
|
| 420 |
+
"grad_norm": 616.7927885159357,
|
| 421 |
+
"learning_rate": 5.651162790697674e-07,
|
| 422 |
+
"logits/chosen": -2.671238899230957,
|
| 423 |
+
"logits/rejected": -2.637321949005127,
|
| 424 |
+
"logps/chosen": -92.645263671875,
|
| 425 |
+
"logps/rejected": -88.57808685302734,
|
| 426 |
+
"loss": 0.8737,
|
| 427 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 428 |
+
"rewards/chosen": 2.4114432334899902,
|
| 429 |
+
"rewards/margins": 1.2348735332489014,
|
| 430 |
+
"rewards/rejected": 1.1765693426132202,
|
| 431 |
+
"step": 250
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.9059233449477352,
|
| 435 |
+
"grad_norm": 478.21518650366113,
|
| 436 |
+
"learning_rate": 5.476744186046511e-07,
|
| 437 |
+
"logits/chosen": -2.555266857147217,
|
| 438 |
+
"logits/rejected": -2.5690598487854004,
|
| 439 |
+
"logps/chosen": -57.567100524902344,
|
| 440 |
+
"logps/rejected": -65.19673919677734,
|
| 441 |
+
"loss": 1.1295,
|
| 442 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 443 |
+
"rewards/chosen": 1.2747440338134766,
|
| 444 |
+
"rewards/margins": 0.9036104083061218,
|
| 445 |
+
"rewards/rejected": 0.37113362550735474,
|
| 446 |
+
"step": 260
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.9407665505226481,
|
| 450 |
+
"grad_norm": 730.4342654590606,
|
| 451 |
+
"learning_rate": 5.302325581395349e-07,
|
| 452 |
+
"logits/chosen": -2.6346664428710938,
|
| 453 |
+
"logits/rejected": -2.633521556854248,
|
| 454 |
+
"logps/chosen": -67.13279724121094,
|
| 455 |
+
"logps/rejected": -82.27578735351562,
|
| 456 |
+
"loss": 1.0712,
|
| 457 |
+
"rewards/accuracies": 0.3062500059604645,
|
| 458 |
+
"rewards/chosen": 1.5027388334274292,
|
| 459 |
+
"rewards/margins": 1.282596468925476,
|
| 460 |
+
"rewards/rejected": 0.2201424092054367,
|
| 461 |
+
"step": 270
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.975609756097561,
|
| 465 |
+
"grad_norm": 595.5490958688126,
|
| 466 |
+
"learning_rate": 5.127906976744186e-07,
|
| 467 |
+
"logits/chosen": -2.5305819511413574,
|
| 468 |
+
"logits/rejected": -2.510344982147217,
|
| 469 |
+
"logps/chosen": -65.64894104003906,
|
| 470 |
+
"logps/rejected": -70.33460998535156,
|
| 471 |
+
"loss": 0.9263,
|
| 472 |
+
"rewards/accuracies": 0.32499998807907104,
|
| 473 |
+
"rewards/chosen": 1.8794810771942139,
|
| 474 |
+
"rewards/margins": 1.4496572017669678,
|
| 475 |
+
"rewards/rejected": 0.42982417345046997,
|
| 476 |
+
"step": 280
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.0104529616724738,
|
| 480 |
+
"grad_norm": 23.979212075374758,
|
| 481 |
+
"learning_rate": 4.953488372093023e-07,
|
| 482 |
+
"logits/chosen": -2.573585033416748,
|
| 483 |
+
"logits/rejected": -2.5460705757141113,
|
| 484 |
+
"logps/chosen": -66.83503723144531,
|
| 485 |
+
"logps/rejected": -65.80335998535156,
|
| 486 |
+
"loss": 0.7697,
|
| 487 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 488 |
+
"rewards/chosen": 4.500079154968262,
|
| 489 |
+
"rewards/margins": 5.2650299072265625,
|
| 490 |
+
"rewards/rejected": -0.7649505734443665,
|
| 491 |
+
"step": 290
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.0452961672473868,
|
| 495 |
+
"grad_norm": 44.13210545620442,
|
| 496 |
+
"learning_rate": 4.779069767441861e-07,
|
| 497 |
+
"logits/chosen": -2.6247591972351074,
|
| 498 |
+
"logits/rejected": -2.610288381576538,
|
| 499 |
+
"logps/chosen": -56.94911575317383,
|
| 500 |
+
"logps/rejected": -78.56959533691406,
|
| 501 |
+
"loss": 0.4215,
|
| 502 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 503 |
+
"rewards/chosen": 8.255823135375977,
|
| 504 |
+
"rewards/margins": 16.339366912841797,
|
| 505 |
+
"rewards/rejected": -8.08354377746582,
|
| 506 |
+
"step": 300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.0452961672473868,
|
| 510 |
+
"eval_logits/chosen": -2.656022548675537,
|
| 511 |
+
"eval_logits/rejected": -2.6403489112854004,
|
| 512 |
+
"eval_logps/chosen": -70.90100860595703,
|
| 513 |
+
"eval_logps/rejected": -79.04632568359375,
|
| 514 |
+
"eval_loss": 1.1233958005905151,
|
| 515 |
+
"eval_rewards/accuracies": 0.3313491940498352,
|
| 516 |
+
"eval_rewards/chosen": 3.0875895023345947,
|
| 517 |
+
"eval_rewards/margins": 1.2970443964004517,
|
| 518 |
+
"eval_rewards/rejected": 1.790545105934143,
|
| 519 |
+
"eval_runtime": 113.4658,
|
| 520 |
+
"eval_samples_per_second": 17.626,
|
| 521 |
+
"eval_steps_per_second": 0.555,
|
| 522 |
+
"step": 300
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.0801393728222997,
|
| 526 |
+
"grad_norm": 40.17101348318783,
|
| 527 |
+
"learning_rate": 4.604651162790698e-07,
|
| 528 |
+
"logits/chosen": -2.611358165740967,
|
| 529 |
+
"logits/rejected": -2.6116092205047607,
|
| 530 |
+
"logps/chosen": -58.17817306518555,
|
| 531 |
+
"logps/rejected": -85.24242401123047,
|
| 532 |
+
"loss": 0.4163,
|
| 533 |
+
"rewards/accuracies": 0.4375,
|
| 534 |
+
"rewards/chosen": 8.134294509887695,
|
| 535 |
+
"rewards/margins": 16.730432510375977,
|
| 536 |
+
"rewards/rejected": -8.596138000488281,
|
| 537 |
+
"step": 310
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.1149825783972125,
|
| 541 |
+
"grad_norm": 153.73635089362784,
|
| 542 |
+
"learning_rate": 4.4302325581395346e-07,
|
| 543 |
+
"logits/chosen": -2.647315740585327,
|
| 544 |
+
"logits/rejected": -2.6344220638275146,
|
| 545 |
+
"logps/chosen": -64.29966735839844,
|
| 546 |
+
"logps/rejected": -88.07272338867188,
|
| 547 |
+
"loss": 0.3886,
|
| 548 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 549 |
+
"rewards/chosen": 8.3077974319458,
|
| 550 |
+
"rewards/margins": 16.785139083862305,
|
| 551 |
+
"rewards/rejected": -8.477343559265137,
|
| 552 |
+
"step": 320
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.1498257839721253,
|
| 556 |
+
"grad_norm": 196.76703226808354,
|
| 557 |
+
"learning_rate": 4.255813953488372e-07,
|
| 558 |
+
"logits/chosen": -2.6384284496307373,
|
| 559 |
+
"logits/rejected": -2.6142385005950928,
|
| 560 |
+
"logps/chosen": -73.04707336425781,
|
| 561 |
+
"logps/rejected": -88.73326873779297,
|
| 562 |
+
"loss": 0.4236,
|
| 563 |
+
"rewards/accuracies": 0.518750011920929,
|
| 564 |
+
"rewards/chosen": 9.5239896774292,
|
| 565 |
+
"rewards/margins": 18.215595245361328,
|
| 566 |
+
"rewards/rejected": -8.691606521606445,
|
| 567 |
+
"step": 330
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.1846689895470384,
|
| 571 |
+
"grad_norm": 16.30501354867415,
|
| 572 |
+
"learning_rate": 4.081395348837209e-07,
|
| 573 |
+
"logits/chosen": -2.6152892112731934,
|
| 574 |
+
"logits/rejected": -2.617434024810791,
|
| 575 |
+
"logps/chosen": -72.81196594238281,
|
| 576 |
+
"logps/rejected": -107.9176254272461,
|
| 577 |
+
"loss": 0.4009,
|
| 578 |
+
"rewards/accuracies": 0.550000011920929,
|
| 579 |
+
"rewards/chosen": 9.404792785644531,
|
| 580 |
+
"rewards/margins": 20.270700454711914,
|
| 581 |
+
"rewards/rejected": -10.8659086227417,
|
| 582 |
+
"step": 340
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.2195121951219512,
|
| 586 |
+
"grad_norm": 253.70590340563848,
|
| 587 |
+
"learning_rate": 3.9069767441860464e-07,
|
| 588 |
+
"logits/chosen": -2.627079486846924,
|
| 589 |
+
"logits/rejected": -2.594177007675171,
|
| 590 |
+
"logps/chosen": -57.3041877746582,
|
| 591 |
+
"logps/rejected": -77.43646240234375,
|
| 592 |
+
"loss": 0.4007,
|
| 593 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 594 |
+
"rewards/chosen": 9.71183967590332,
|
| 595 |
+
"rewards/margins": 19.823144912719727,
|
| 596 |
+
"rewards/rejected": -10.111306190490723,
|
| 597 |
+
"step": 350
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.254355400696864,
|
| 601 |
+
"grad_norm": 137.93033306808474,
|
| 602 |
+
"learning_rate": 3.7325581395348837e-07,
|
| 603 |
+
"logits/chosen": -2.603165626525879,
|
| 604 |
+
"logits/rejected": -2.573925495147705,
|
| 605 |
+
"logps/chosen": -62.46228790283203,
|
| 606 |
+
"logps/rejected": -73.45440673828125,
|
| 607 |
+
"loss": 0.374,
|
| 608 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 609 |
+
"rewards/chosen": 8.073513984680176,
|
| 610 |
+
"rewards/margins": 14.77312183380127,
|
| 611 |
+
"rewards/rejected": -6.699607849121094,
|
| 612 |
+
"step": 360
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.289198606271777,
|
| 616 |
+
"grad_norm": 2.3357390678870833,
|
| 617 |
+
"learning_rate": 3.5581395348837204e-07,
|
| 618 |
+
"logits/chosen": -2.562800407409668,
|
| 619 |
+
"logits/rejected": -2.5790553092956543,
|
| 620 |
+
"logps/chosen": -61.26555633544922,
|
| 621 |
+
"logps/rejected": -86.61184692382812,
|
| 622 |
+
"loss": 0.4449,
|
| 623 |
+
"rewards/accuracies": 0.46875,
|
| 624 |
+
"rewards/chosen": 8.428425788879395,
|
| 625 |
+
"rewards/margins": 18.987430572509766,
|
| 626 |
+
"rewards/rejected": -10.559003829956055,
|
| 627 |
+
"step": 370
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 1.32404181184669,
|
| 631 |
+
"grad_norm": 240.07418332327254,
|
| 632 |
+
"learning_rate": 3.383720930232558e-07,
|
| 633 |
+
"logits/chosen": -2.55873966217041,
|
| 634 |
+
"logits/rejected": -2.556434392929077,
|
| 635 |
+
"logps/chosen": -77.422607421875,
|
| 636 |
+
"logps/rejected": -101.52474212646484,
|
| 637 |
+
"loss": 0.4144,
|
| 638 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 639 |
+
"rewards/chosen": 8.128644943237305,
|
| 640 |
+
"rewards/margins": 21.078929901123047,
|
| 641 |
+
"rewards/rejected": -12.950288772583008,
|
| 642 |
+
"step": 380
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 1.3588850174216027,
|
| 646 |
+
"grad_norm": 26.553648537118264,
|
| 647 |
+
"learning_rate": 3.2093023255813955e-07,
|
| 648 |
+
"logits/chosen": -2.6296944618225098,
|
| 649 |
+
"logits/rejected": -2.6129908561706543,
|
| 650 |
+
"logps/chosen": -56.90552520751953,
|
| 651 |
+
"logps/rejected": -84.36430358886719,
|
| 652 |
+
"loss": 0.378,
|
| 653 |
+
"rewards/accuracies": 0.4375,
|
| 654 |
+
"rewards/chosen": 8.677267074584961,
|
| 655 |
+
"rewards/margins": 19.946393966674805,
|
| 656 |
+
"rewards/rejected": -11.269128799438477,
|
| 657 |
+
"step": 390
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.3937282229965158,
|
| 661 |
+
"grad_norm": 11.915044642787674,
|
| 662 |
+
"learning_rate": 3.034883720930232e-07,
|
| 663 |
+
"logits/chosen": -2.6192545890808105,
|
| 664 |
+
"logits/rejected": -2.591175079345703,
|
| 665 |
+
"logps/chosen": -77.75074768066406,
|
| 666 |
+
"logps/rejected": -112.25565338134766,
|
| 667 |
+
"loss": 0.393,
|
| 668 |
+
"rewards/accuracies": 0.518750011920929,
|
| 669 |
+
"rewards/chosen": 7.220816135406494,
|
| 670 |
+
"rewards/margins": 19.133440017700195,
|
| 671 |
+
"rewards/rejected": -11.912622451782227,
|
| 672 |
+
"step": 400
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.3937282229965158,
|
| 676 |
+
"eval_logits/chosen": -2.625042676925659,
|
| 677 |
+
"eval_logits/rejected": -2.609307050704956,
|
| 678 |
+
"eval_logps/chosen": -74.64740753173828,
|
| 679 |
+
"eval_logps/rejected": -83.32994842529297,
|
| 680 |
+
"eval_loss": 1.223362684249878,
|
| 681 |
+
"eval_rewards/accuracies": 0.32341268658638,
|
| 682 |
+
"eval_rewards/chosen": -0.1343114674091339,
|
| 683 |
+
"eval_rewards/margins": 1.759063720703125,
|
| 684 |
+
"eval_rewards/rejected": -1.893375277519226,
|
| 685 |
+
"eval_runtime": 113.3813,
|
| 686 |
+
"eval_samples_per_second": 17.64,
|
| 687 |
+
"eval_steps_per_second": 0.556,
|
| 688 |
+
"step": 400
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 1.4285714285714286,
|
| 692 |
+
"grad_norm": 265.4416418102539,
|
| 693 |
+
"learning_rate": 2.8604651162790695e-07,
|
| 694 |
+
"logits/chosen": -2.6114742755889893,
|
| 695 |
+
"logits/rejected": -2.601598024368286,
|
| 696 |
+
"logps/chosen": -73.08380889892578,
|
| 697 |
+
"logps/rejected": -94.53117370605469,
|
| 698 |
+
"loss": 0.6924,
|
| 699 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 700 |
+
"rewards/chosen": 8.195378303527832,
|
| 701 |
+
"rewards/margins": 20.368560791015625,
|
| 702 |
+
"rewards/rejected": -12.17318344116211,
|
| 703 |
+
"step": 410
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.4634146341463414,
|
| 707 |
+
"grad_norm": 110.44055155377453,
|
| 708 |
+
"learning_rate": 2.6860465116279067e-07,
|
| 709 |
+
"logits/chosen": -2.651491641998291,
|
| 710 |
+
"logits/rejected": -2.65342378616333,
|
| 711 |
+
"logps/chosen": -69.68745422363281,
|
| 712 |
+
"logps/rejected": -97.7309341430664,
|
| 713 |
+
"loss": 0.4257,
|
| 714 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 715 |
+
"rewards/chosen": 5.508909702301025,
|
| 716 |
+
"rewards/margins": 16.567991256713867,
|
| 717 |
+
"rewards/rejected": -11.059081077575684,
|
| 718 |
+
"step": 420
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.4982578397212545,
|
| 722 |
+
"grad_norm": 0.6603991710359594,
|
| 723 |
+
"learning_rate": 2.511627906976744e-07,
|
| 724 |
+
"logits/chosen": -2.606182336807251,
|
| 725 |
+
"logits/rejected": -2.5925304889678955,
|
| 726 |
+
"logps/chosen": -62.015174865722656,
|
| 727 |
+
"logps/rejected": -85.03865814208984,
|
| 728 |
+
"loss": 0.3861,
|
| 729 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 730 |
+
"rewards/chosen": 5.7165632247924805,
|
| 731 |
+
"rewards/margins": 16.509204864501953,
|
| 732 |
+
"rewards/rejected": -10.792640686035156,
|
| 733 |
+
"step": 430
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.533101045296167,
|
| 737 |
+
"grad_norm": 171.49719572579374,
|
| 738 |
+
"learning_rate": 2.3372093023255815e-07,
|
| 739 |
+
"logits/chosen": -2.5812370777130127,
|
| 740 |
+
"logits/rejected": -2.5880465507507324,
|
| 741 |
+
"logps/chosen": -62.13433074951172,
|
| 742 |
+
"logps/rejected": -84.38856506347656,
|
| 743 |
+
"loss": 0.4158,
|
| 744 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 745 |
+
"rewards/chosen": 6.746127128601074,
|
| 746 |
+
"rewards/margins": 16.110437393188477,
|
| 747 |
+
"rewards/rejected": -9.364312171936035,
|
| 748 |
+
"step": 440
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.5679442508710801,
|
| 752 |
+
"grad_norm": 28.499265540380463,
|
| 753 |
+
"learning_rate": 2.1627906976744185e-07,
|
| 754 |
+
"logits/chosen": -2.641430139541626,
|
| 755 |
+
"logits/rejected": -2.6030945777893066,
|
| 756 |
+
"logps/chosen": -79.95354461669922,
|
| 757 |
+
"logps/rejected": -103.62571716308594,
|
| 758 |
+
"loss": 0.4177,
|
| 759 |
+
"rewards/accuracies": 0.512499988079071,
|
| 760 |
+
"rewards/chosen": 9.364060401916504,
|
| 761 |
+
"rewards/margins": 24.554065704345703,
|
| 762 |
+
"rewards/rejected": -15.1900053024292,
|
| 763 |
+
"step": 450
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.6027874564459932,
|
| 767 |
+
"grad_norm": 208.9071332592839,
|
| 768 |
+
"learning_rate": 1.9883720930232558e-07,
|
| 769 |
+
"logits/chosen": -2.6206653118133545,
|
| 770 |
+
"logits/rejected": -2.597672939300537,
|
| 771 |
+
"logps/chosen": -66.5615463256836,
|
| 772 |
+
"logps/rejected": -93.40229797363281,
|
| 773 |
+
"loss": 0.4043,
|
| 774 |
+
"rewards/accuracies": 0.5,
|
| 775 |
+
"rewards/chosen": 8.271448135375977,
|
| 776 |
+
"rewards/margins": 22.373382568359375,
|
| 777 |
+
"rewards/rejected": -14.101933479309082,
|
| 778 |
+
"step": 460
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.6376306620209058,
|
| 782 |
+
"grad_norm": 0.3844536104459883,
|
| 783 |
+
"learning_rate": 1.8139534883720928e-07,
|
| 784 |
+
"logits/chosen": -2.5973286628723145,
|
| 785 |
+
"logits/rejected": -2.5908195972442627,
|
| 786 |
+
"logps/chosen": -54.78232955932617,
|
| 787 |
+
"logps/rejected": -81.273681640625,
|
| 788 |
+
"loss": 0.4316,
|
| 789 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 790 |
+
"rewards/chosen": 5.607256889343262,
|
| 791 |
+
"rewards/margins": 13.865686416625977,
|
| 792 |
+
"rewards/rejected": -8.258430480957031,
|
| 793 |
+
"step": 470
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.6724738675958188,
|
| 797 |
+
"grad_norm": 150.7317542423224,
|
| 798 |
+
"learning_rate": 1.63953488372093e-07,
|
| 799 |
+
"logits/chosen": -2.599632501602173,
|
| 800 |
+
"logits/rejected": -2.584341049194336,
|
| 801 |
+
"logps/chosen": -49.672950744628906,
|
| 802 |
+
"logps/rejected": -57.88109588623047,
|
| 803 |
+
"loss": 0.4093,
|
| 804 |
+
"rewards/accuracies": 0.35624998807907104,
|
| 805 |
+
"rewards/chosen": 3.7741661071777344,
|
| 806 |
+
"rewards/margins": 8.84811782836914,
|
| 807 |
+
"rewards/rejected": -5.073951721191406,
|
| 808 |
+
"step": 480
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 1.7073170731707317,
|
| 812 |
+
"grad_norm": 51.065123641891695,
|
| 813 |
+
"learning_rate": 1.4651162790697676e-07,
|
| 814 |
+
"logits/chosen": -2.571542978286743,
|
| 815 |
+
"logits/rejected": -2.561433792114258,
|
| 816 |
+
"logps/chosen": -66.06110382080078,
|
| 817 |
+
"logps/rejected": -81.25625610351562,
|
| 818 |
+
"loss": 0.5041,
|
| 819 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 820 |
+
"rewards/chosen": 6.158200740814209,
|
| 821 |
+
"rewards/margins": 15.785115242004395,
|
| 822 |
+
"rewards/rejected": -9.626913070678711,
|
| 823 |
+
"step": 490
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.7421602787456445,
|
| 827 |
+
"grad_norm": 9.459943139553493,
|
| 828 |
+
"learning_rate": 1.2906976744186046e-07,
|
| 829 |
+
"logits/chosen": -2.496741771697998,
|
| 830 |
+
"logits/rejected": -2.4906914234161377,
|
| 831 |
+
"logps/chosen": -68.98286437988281,
|
| 832 |
+
"logps/rejected": -96.44841003417969,
|
| 833 |
+
"loss": 0.3986,
|
| 834 |
+
"rewards/accuracies": 0.46875,
|
| 835 |
+
"rewards/chosen": 6.282340049743652,
|
| 836 |
+
"rewards/margins": 17.225933074951172,
|
| 837 |
+
"rewards/rejected": -10.943593978881836,
|
| 838 |
+
"step": 500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.7421602787456445,
|
| 842 |
+
"eval_logits/chosen": -2.607004404067993,
|
| 843 |
+
"eval_logits/rejected": -2.5908732414245605,
|
| 844 |
+
"eval_logps/chosen": -74.26602935791016,
|
| 845 |
+
"eval_logps/rejected": -83.2776107788086,
|
| 846 |
+
"eval_loss": 1.2247282266616821,
|
| 847 |
+
"eval_rewards/accuracies": 0.3214285671710968,
|
| 848 |
+
"eval_rewards/chosen": 0.1936810314655304,
|
| 849 |
+
"eval_rewards/margins": 2.0420379638671875,
|
| 850 |
+
"eval_rewards/rejected": -1.848357081413269,
|
| 851 |
+
"eval_runtime": 113.3979,
|
| 852 |
+
"eval_samples_per_second": 17.637,
|
| 853 |
+
"eval_steps_per_second": 0.556,
|
| 854 |
+
"step": 500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.7770034843205575,
|
| 858 |
+
"grad_norm": 22.381813019053016,
|
| 859 |
+
"learning_rate": 1.1162790697674417e-07,
|
| 860 |
+
"logits/chosen": -2.572063446044922,
|
| 861 |
+
"logits/rejected": -2.551802635192871,
|
| 862 |
+
"logps/chosen": -62.93878173828125,
|
| 863 |
+
"logps/rejected": -80.62296295166016,
|
| 864 |
+
"loss": 0.4109,
|
| 865 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 866 |
+
"rewards/chosen": 6.723170280456543,
|
| 867 |
+
"rewards/margins": 16.216602325439453,
|
| 868 |
+
"rewards/rejected": -9.49343204498291,
|
| 869 |
+
"step": 510
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.8118466898954704,
|
| 873 |
+
"grad_norm": 34.84061860816185,
|
| 874 |
+
"learning_rate": 9.418604651162791e-08,
|
| 875 |
+
"logits/chosen": -2.5822184085845947,
|
| 876 |
+
"logits/rejected": -2.5742483139038086,
|
| 877 |
+
"logps/chosen": -66.30496978759766,
|
| 878 |
+
"logps/rejected": -88.59693908691406,
|
| 879 |
+
"loss": 0.3956,
|
| 880 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 881 |
+
"rewards/chosen": 6.850976467132568,
|
| 882 |
+
"rewards/margins": 16.641326904296875,
|
| 883 |
+
"rewards/rejected": -9.790349006652832,
|
| 884 |
+
"step": 520
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 1.8466898954703832,
|
| 888 |
+
"grad_norm": 26.726116225430992,
|
| 889 |
+
"learning_rate": 7.674418604651163e-08,
|
| 890 |
+
"logits/chosen": -2.5606324672698975,
|
| 891 |
+
"logits/rejected": -2.5552823543548584,
|
| 892 |
+
"logps/chosen": -62.88715744018555,
|
| 893 |
+
"logps/rejected": -84.60001373291016,
|
| 894 |
+
"loss": 0.3962,
|
| 895 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 896 |
+
"rewards/chosen": 7.2661237716674805,
|
| 897 |
+
"rewards/margins": 18.649456024169922,
|
| 898 |
+
"rewards/rejected": -11.383334159851074,
|
| 899 |
+
"step": 530
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.8815331010452963,
|
| 903 |
+
"grad_norm": 746.1912709837381,
|
| 904 |
+
"learning_rate": 5.930232558139535e-08,
|
| 905 |
+
"logits/chosen": -2.5763189792633057,
|
| 906 |
+
"logits/rejected": -2.589749336242676,
|
| 907 |
+
"logps/chosen": -54.900634765625,
|
| 908 |
+
"logps/rejected": -89.78245544433594,
|
| 909 |
+
"loss": 0.4804,
|
| 910 |
+
"rewards/accuracies": 0.45625001192092896,
|
| 911 |
+
"rewards/chosen": 7.59392786026001,
|
| 912 |
+
"rewards/margins": 21.733322143554688,
|
| 913 |
+
"rewards/rejected": -14.139394760131836,
|
| 914 |
+
"step": 540
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.916376306620209,
|
| 918 |
+
"grad_norm": 0.6251683362748027,
|
| 919 |
+
"learning_rate": 4.1860465116279067e-08,
|
| 920 |
+
"logits/chosen": -2.5647377967834473,
|
| 921 |
+
"logits/rejected": -2.5423054695129395,
|
| 922 |
+
"logps/chosen": -79.24058532714844,
|
| 923 |
+
"logps/rejected": -106.3379898071289,
|
| 924 |
+
"loss": 0.4013,
|
| 925 |
+
"rewards/accuracies": 0.5,
|
| 926 |
+
"rewards/chosen": 9.971015930175781,
|
| 927 |
+
"rewards/margins": 26.699609756469727,
|
| 928 |
+
"rewards/rejected": -16.728595733642578,
|
| 929 |
+
"step": 550
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 1.951219512195122,
|
| 933 |
+
"grad_norm": 252.84212758390947,
|
| 934 |
+
"learning_rate": 2.441860465116279e-08,
|
| 935 |
+
"logits/chosen": -2.560770034790039,
|
| 936 |
+
"logits/rejected": -2.567601442337036,
|
| 937 |
+
"logps/chosen": -57.37895584106445,
|
| 938 |
+
"logps/rejected": -80.76757049560547,
|
| 939 |
+
"loss": 0.4091,
|
| 940 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 941 |
+
"rewards/chosen": 5.453306674957275,
|
| 942 |
+
"rewards/margins": 14.458730697631836,
|
| 943 |
+
"rewards/rejected": -9.005423545837402,
|
| 944 |
+
"step": 560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.986062717770035,
|
| 948 |
+
"grad_norm": 110.2396617228431,
|
| 949 |
+
"learning_rate": 6.976744186046511e-09,
|
| 950 |
+
"logits/chosen": -2.637852191925049,
|
| 951 |
+
"logits/rejected": -2.608257293701172,
|
| 952 |
+
"logps/chosen": -58.60048294067383,
|
| 953 |
+
"logps/rejected": -71.56431579589844,
|
| 954 |
+
"loss": 0.3934,
|
| 955 |
+
"rewards/accuracies": 0.4375,
|
| 956 |
+
"rewards/chosen": 5.686958312988281,
|
| 957 |
+
"rewards/margins": 13.428431510925293,
|
| 958 |
+
"rewards/rejected": -7.7414727210998535,
|
| 959 |
+
"step": 570
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 2.0,
|
| 963 |
+
"step": 574,
|
| 964 |
+
"total_flos": 0.0,
|
| 965 |
+
"train_loss": 0.6794358088995106,
|
| 966 |
+
"train_runtime": 6413.3344,
|
| 967 |
+
"train_samples_per_second": 5.719,
|
| 968 |
+
"train_steps_per_second": 0.09
|
| 969 |
+
}
|
| 970 |
+
],
|
| 971 |
+
"logging_steps": 10,
|
| 972 |
+
"max_steps": 574,
|
| 973 |
+
"num_input_tokens_seen": 0,
|
| 974 |
+
"num_train_epochs": 2,
|
| 975 |
+
"save_steps": 100,
|
| 976 |
+
"stateful_callbacks": {
|
| 977 |
+
"TrainerControl": {
|
| 978 |
+
"args": {
|
| 979 |
+
"should_epoch_stop": false,
|
| 980 |
+
"should_evaluate": false,
|
| 981 |
+
"should_log": false,
|
| 982 |
+
"should_save": true,
|
| 983 |
+
"should_training_stop": true
|
| 984 |
+
},
|
| 985 |
+
"attributes": {}
|
| 986 |
+
}
|
| 987 |
+
},
|
| 988 |
+
"total_flos": 0.0,
|
| 989 |
+
"train_batch_size": 8,
|
| 990 |
+
"trial_name": null,
|
| 991 |
+
"trial_params": null
|
| 992 |
+
}
|