Model save
Browse files- README.md +77 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +2497 -0
README.md
ADDED
|
@@ -0,0 +1,77 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
tags:
|
| 4 |
+
- trl
|
| 5 |
+
- dpo
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
base_model: allenai/tulu-2-13b
|
| 8 |
+
model-index:
|
| 9 |
+
- name: tulu2-13b-cost-UI-UF-7bjudge
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# tulu2-13b-cost-UI-UF-7bjudge
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [allenai/tulu-2-13b](https://huggingface.co/allenai/tulu-2-13b) on the None dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.6924
|
| 21 |
+
- Rewards/chosen: 0.0238
|
| 22 |
+
- Rewards/rejected: 0.0217
|
| 23 |
+
- Rewards/accuracies: 0.5640
|
| 24 |
+
- Rewards/margins: 0.0021
|
| 25 |
+
- Rewards/margins Max: 0.0393
|
| 26 |
+
- Rewards/margins Min: -0.0362
|
| 27 |
+
- Rewards/margins Std: 0.0335
|
| 28 |
+
- Logps/rejected: -313.6311
|
| 29 |
+
- Logps/chosen: -321.7533
|
| 30 |
+
- Logits/rejected: -1.0117
|
| 31 |
+
- Logits/chosen: -1.1541
|
| 32 |
+
|
| 33 |
+
## Model description
|
| 34 |
+
|
| 35 |
+
More information needed
|
| 36 |
+
|
| 37 |
+
## Intended uses & limitations
|
| 38 |
+
|
| 39 |
+
More information needed
|
| 40 |
+
|
| 41 |
+
## Training and evaluation data
|
| 42 |
+
|
| 43 |
+
More information needed
|
| 44 |
+
|
| 45 |
+
## Training procedure
|
| 46 |
+
|
| 47 |
+
### Training hyperparameters
|
| 48 |
+
|
| 49 |
+
The following hyperparameters were used during training:
|
| 50 |
+
- learning_rate: 5e-07
|
| 51 |
+
- train_batch_size: 1
|
| 52 |
+
- eval_batch_size: 4
|
| 53 |
+
- seed: 42
|
| 54 |
+
- distributed_type: multi-GPU
|
| 55 |
+
- num_devices: 8
|
| 56 |
+
- gradient_accumulation_steps: 2
|
| 57 |
+
- total_train_batch_size: 16
|
| 58 |
+
- total_eval_batch_size: 32
|
| 59 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 60 |
+
- lr_scheduler_type: cosine
|
| 61 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 62 |
+
- num_epochs: 1
|
| 63 |
+
|
| 64 |
+
### Training results
|
| 65 |
+
|
| 66 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Rewards/margins Max | Rewards/margins Min | Rewards/margins Std | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 67 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:-------------------:|:-------------------:|:-------------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 68 |
+
| 0.6643 | 1.0 | 1350 | 0.6924 | 0.0238 | 0.0217 | 0.5640 | 0.0021 | 0.0393 | -0.0362 | 0.0335 | -313.6311 | -321.7533 | -1.0117 | -1.1541 |
|
| 69 |
+
|
| 70 |
+
|
| 71 |
+
### Framework versions
|
| 72 |
+
|
| 73 |
+
- PEFT 0.7.1
|
| 74 |
+
- Transformers 4.39.0.dev0
|
| 75 |
+
- Pytorch 2.1.2+cu121
|
| 76 |
+
- Datasets 2.14.6
|
| 77 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1001466944
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bae62720fc889439a11cb21e195e4d29ed4d4b9a47127a12c7dd4e2b1f7b0409
|
| 3 |
size 1001466944
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.6709109875007911,
|
| 4 |
+
"train_runtime": 7657.6869,
|
| 5 |
+
"train_samples": 21603,
|
| 6 |
+
"train_samples_per_second": 2.821,
|
| 7 |
+
"train_steps_per_second": 0.176
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.6709109875007911,
|
| 4 |
+
"train_runtime": 7657.6869,
|
| 5 |
+
"train_samples": 21603,
|
| 6 |
+
"train_samples_per_second": 2.821,
|
| 7 |
+
"train_steps_per_second": 0.176
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2497 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9996297667530544,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 1350,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"grad_norm": 0.3203125,
|
| 14 |
+
"learning_rate": 3.7037037037037036e-09,
|
| 15 |
+
"logits/chosen": -1.8868012428283691,
|
| 16 |
+
"logits/rejected": -1.5128765106201172,
|
| 17 |
+
"logps/chosen": -211.82923889160156,
|
| 18 |
+
"logps/rejected": -174.22210693359375,
|
| 19 |
+
"loss": 0.6931,
|
| 20 |
+
"rewards/accuracies": 0.0,
|
| 21 |
+
"rewards/chosen": 0.0,
|
| 22 |
+
"rewards/margins": 0.0,
|
| 23 |
+
"rewards/margins_max": 0.0,
|
| 24 |
+
"rewards/margins_min": 0.0,
|
| 25 |
+
"rewards/margins_std": NaN,
|
| 26 |
+
"rewards/rejected": 0.0,
|
| 27 |
+
"step": 1
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.01,
|
| 31 |
+
"grad_norm": 0.291015625,
|
| 32 |
+
"learning_rate": 3.7037037037037036e-08,
|
| 33 |
+
"logits/chosen": -1.9202638864517212,
|
| 34 |
+
"logits/rejected": -1.4150642156600952,
|
| 35 |
+
"logps/chosen": -264.96990966796875,
|
| 36 |
+
"logps/rejected": -187.2726593017578,
|
| 37 |
+
"loss": 0.693,
|
| 38 |
+
"rewards/accuracies": 0.3888888955116272,
|
| 39 |
+
"rewards/chosen": 0.0007235632510855794,
|
| 40 |
+
"rewards/margins": -0.0005523893050849438,
|
| 41 |
+
"rewards/margins_max": -0.0005523893050849438,
|
| 42 |
+
"rewards/margins_min": -0.0005523893050849438,
|
| 43 |
+
"rewards/margins_std": NaN,
|
| 44 |
+
"rewards/rejected": 0.0012759524397552013,
|
| 45 |
+
"step": 10
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.01,
|
| 49 |
+
"grad_norm": 0.328125,
|
| 50 |
+
"learning_rate": 7.407407407407407e-08,
|
| 51 |
+
"logits/chosen": -2.004002809524536,
|
| 52 |
+
"logits/rejected": -1.4588892459869385,
|
| 53 |
+
"logps/chosen": -264.7344970703125,
|
| 54 |
+
"logps/rejected": -264.04217529296875,
|
| 55 |
+
"loss": 0.6931,
|
| 56 |
+
"rewards/accuracies": 0.5,
|
| 57 |
+
"rewards/chosen": 0.00047433850704692304,
|
| 58 |
+
"rewards/margins": 8.134842209983617e-05,
|
| 59 |
+
"rewards/margins_max": 8.134842209983617e-05,
|
| 60 |
+
"rewards/margins_min": 8.134842209983617e-05,
|
| 61 |
+
"rewards/margins_std": NaN,
|
| 62 |
+
"rewards/rejected": 0.0003929900412913412,
|
| 63 |
+
"step": 20
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.02,
|
| 67 |
+
"grad_norm": 0.359375,
|
| 68 |
+
"learning_rate": 1.111111111111111e-07,
|
| 69 |
+
"logits/chosen": -2.0108284950256348,
|
| 70 |
+
"logits/rejected": -1.3832088708877563,
|
| 71 |
+
"logps/chosen": -253.376953125,
|
| 72 |
+
"logps/rejected": -235.9996337890625,
|
| 73 |
+
"loss": 0.6931,
|
| 74 |
+
"rewards/accuracies": 0.550000011920929,
|
| 75 |
+
"rewards/chosen": 0.0005588417407125235,
|
| 76 |
+
"rewards/margins": 0.0008742943173274398,
|
| 77 |
+
"rewards/margins_max": 0.0008742943173274398,
|
| 78 |
+
"rewards/margins_min": 0.0008742943173274398,
|
| 79 |
+
"rewards/margins_std": NaN,
|
| 80 |
+
"rewards/rejected": -0.0003154525766149163,
|
| 81 |
+
"step": 30
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.03,
|
| 85 |
+
"grad_norm": 0.310546875,
|
| 86 |
+
"learning_rate": 1.4814814814814815e-07,
|
| 87 |
+
"logits/chosen": -1.9831428527832031,
|
| 88 |
+
"logits/rejected": -1.3419771194458008,
|
| 89 |
+
"logps/chosen": -249.38821411132812,
|
| 90 |
+
"logps/rejected": -314.62939453125,
|
| 91 |
+
"loss": 0.6929,
|
| 92 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 93 |
+
"rewards/chosen": 0.0013472252758219838,
|
| 94 |
+
"rewards/margins": 0.001804034924134612,
|
| 95 |
+
"rewards/margins_max": 0.001804034924134612,
|
| 96 |
+
"rewards/margins_min": 0.001804034924134612,
|
| 97 |
+
"rewards/margins_std": NaN,
|
| 98 |
+
"rewards/rejected": -0.00045680999755859375,
|
| 99 |
+
"step": 40
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.04,
|
| 103 |
+
"grad_norm": 0.314453125,
|
| 104 |
+
"learning_rate": 1.8518518518518516e-07,
|
| 105 |
+
"logits/chosen": -2.0799918174743652,
|
| 106 |
+
"logits/rejected": -1.2689168453216553,
|
| 107 |
+
"logps/chosen": -269.61492919921875,
|
| 108 |
+
"logps/rejected": -192.31704711914062,
|
| 109 |
+
"loss": 0.6929,
|
| 110 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 111 |
+
"rewards/chosen": 0.0003735256032086909,
|
| 112 |
+
"rewards/margins": 0.0009886417537927628,
|
| 113 |
+
"rewards/margins_max": 0.0009886417537927628,
|
| 114 |
+
"rewards/margins_min": 0.0009886417537927628,
|
| 115 |
+
"rewards/margins_std": NaN,
|
| 116 |
+
"rewards/rejected": -0.000615116092376411,
|
| 117 |
+
"step": 50
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.04,
|
| 121 |
+
"grad_norm": 0.310546875,
|
| 122 |
+
"learning_rate": 2.222222222222222e-07,
|
| 123 |
+
"logits/chosen": -1.9307972192764282,
|
| 124 |
+
"logits/rejected": -1.5086009502410889,
|
| 125 |
+
"logps/chosen": -200.31387329101562,
|
| 126 |
+
"logps/rejected": -219.31478881835938,
|
| 127 |
+
"loss": 0.6926,
|
| 128 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 129 |
+
"rewards/chosen": 0.00043567089596763253,
|
| 130 |
+
"rewards/margins": 0.0002629738301038742,
|
| 131 |
+
"rewards/margins_max": 0.0002629738301038742,
|
| 132 |
+
"rewards/margins_min": 0.0002629738301038742,
|
| 133 |
+
"rewards/margins_std": NaN,
|
| 134 |
+
"rewards/rejected": 0.0001726970513118431,
|
| 135 |
+
"step": 60
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.05,
|
| 139 |
+
"grad_norm": 0.32421875,
|
| 140 |
+
"learning_rate": 2.5925925925925923e-07,
|
| 141 |
+
"logits/chosen": -2.057626724243164,
|
| 142 |
+
"logits/rejected": -1.3474286794662476,
|
| 143 |
+
"logps/chosen": -214.9297637939453,
|
| 144 |
+
"logps/rejected": -265.07855224609375,
|
| 145 |
+
"loss": 0.6922,
|
| 146 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 147 |
+
"rewards/chosen": 0.00036124425241723657,
|
| 148 |
+
"rewards/margins": 0.002079057740047574,
|
| 149 |
+
"rewards/margins_max": 0.002079057740047574,
|
| 150 |
+
"rewards/margins_min": 0.002079057740047574,
|
| 151 |
+
"rewards/margins_std": NaN,
|
| 152 |
+
"rewards/rejected": -0.0017178135458379984,
|
| 153 |
+
"step": 70
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.06,
|
| 157 |
+
"grad_norm": 0.33203125,
|
| 158 |
+
"learning_rate": 2.962962962962963e-07,
|
| 159 |
+
"logits/chosen": -2.0998992919921875,
|
| 160 |
+
"logits/rejected": -1.2106634378433228,
|
| 161 |
+
"logps/chosen": -266.18182373046875,
|
| 162 |
+
"logps/rejected": -211.5884552001953,
|
| 163 |
+
"loss": 0.6926,
|
| 164 |
+
"rewards/accuracies": 0.75,
|
| 165 |
+
"rewards/chosen": 0.0019807242788374424,
|
| 166 |
+
"rewards/margins": 0.003271686378866434,
|
| 167 |
+
"rewards/margins_max": 0.003271686378866434,
|
| 168 |
+
"rewards/margins_min": 0.003271686378866434,
|
| 169 |
+
"rewards/margins_std": NaN,
|
| 170 |
+
"rewards/rejected": -0.0012909623328596354,
|
| 171 |
+
"step": 80
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.07,
|
| 175 |
+
"grad_norm": 0.337890625,
|
| 176 |
+
"learning_rate": 3.333333333333333e-07,
|
| 177 |
+
"logits/chosen": -2.0147571563720703,
|
| 178 |
+
"logits/rejected": -1.2995073795318604,
|
| 179 |
+
"logps/chosen": -247.42263793945312,
|
| 180 |
+
"logps/rejected": -228.5316619873047,
|
| 181 |
+
"loss": 0.6919,
|
| 182 |
+
"rewards/accuracies": 0.5,
|
| 183 |
+
"rewards/chosen": 0.001941970898769796,
|
| 184 |
+
"rewards/margins": 0.0006094207055866718,
|
| 185 |
+
"rewards/margins_max": 0.0006094207055866718,
|
| 186 |
+
"rewards/margins_min": 0.0006094207055866718,
|
| 187 |
+
"rewards/margins_std": NaN,
|
| 188 |
+
"rewards/rejected": 0.0013325500767678022,
|
| 189 |
+
"step": 90
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.07,
|
| 193 |
+
"grad_norm": 0.322265625,
|
| 194 |
+
"learning_rate": 3.703703703703703e-07,
|
| 195 |
+
"logits/chosen": -1.983267068862915,
|
| 196 |
+
"logits/rejected": -1.3546688556671143,
|
| 197 |
+
"logps/chosen": -202.480224609375,
|
| 198 |
+
"logps/rejected": -174.5247344970703,
|
| 199 |
+
"loss": 0.6916,
|
| 200 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 201 |
+
"rewards/chosen": 0.0025401152670383453,
|
| 202 |
+
"rewards/margins": 0.002821914618834853,
|
| 203 |
+
"rewards/margins_max": 0.002821914618834853,
|
| 204 |
+
"rewards/margins_min": 0.002821914618834853,
|
| 205 |
+
"rewards/margins_std": NaN,
|
| 206 |
+
"rewards/rejected": -0.0002817992935888469,
|
| 207 |
+
"step": 100
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.08,
|
| 211 |
+
"grad_norm": 0.3125,
|
| 212 |
+
"learning_rate": 4.0740740740740737e-07,
|
| 213 |
+
"logits/chosen": -2.0803046226501465,
|
| 214 |
+
"logits/rejected": -1.4716895818710327,
|
| 215 |
+
"logps/chosen": -225.15560913085938,
|
| 216 |
+
"logps/rejected": -256.0934753417969,
|
| 217 |
+
"loss": 0.6912,
|
| 218 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 219 |
+
"rewards/chosen": 0.003935464657843113,
|
| 220 |
+
"rewards/margins": 0.0047843605279922485,
|
| 221 |
+
"rewards/margins_max": 0.0047843605279922485,
|
| 222 |
+
"rewards/margins_min": 0.0047843605279922485,
|
| 223 |
+
"rewards/margins_std": NaN,
|
| 224 |
+
"rewards/rejected": -0.0008488958701491356,
|
| 225 |
+
"step": 110
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.09,
|
| 229 |
+
"grad_norm": 0.298828125,
|
| 230 |
+
"learning_rate": 4.444444444444444e-07,
|
| 231 |
+
"logits/chosen": -1.8733142614364624,
|
| 232 |
+
"logits/rejected": -1.7225223779678345,
|
| 233 |
+
"logps/chosen": -172.91468811035156,
|
| 234 |
+
"logps/rejected": -199.7720489501953,
|
| 235 |
+
"loss": 0.6909,
|
| 236 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 237 |
+
"rewards/chosen": 0.00166155060287565,
|
| 238 |
+
"rewards/margins": 0.0015941675519570708,
|
| 239 |
+
"rewards/margins_max": 0.0015941675519570708,
|
| 240 |
+
"rewards/margins_min": 0.0015941675519570708,
|
| 241 |
+
"rewards/margins_std": NaN,
|
| 242 |
+
"rewards/rejected": 6.73827962600626e-05,
|
| 243 |
+
"step": 120
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.1,
|
| 247 |
+
"grad_norm": 0.326171875,
|
| 248 |
+
"learning_rate": 4.814814814814814e-07,
|
| 249 |
+
"logits/chosen": -1.9976625442504883,
|
| 250 |
+
"logits/rejected": -1.2138649225234985,
|
| 251 |
+
"logps/chosen": -239.32485961914062,
|
| 252 |
+
"logps/rejected": -240.15597534179688,
|
| 253 |
+
"loss": 0.6899,
|
| 254 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 255 |
+
"rewards/chosen": 0.0042144907638430595,
|
| 256 |
+
"rewards/margins": 0.005392442457377911,
|
| 257 |
+
"rewards/margins_max": 0.005392442457377911,
|
| 258 |
+
"rewards/margins_min": 0.005392442457377911,
|
| 259 |
+
"rewards/margins_std": NaN,
|
| 260 |
+
"rewards/rejected": -0.001177951809950173,
|
| 261 |
+
"step": 130
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.1,
|
| 265 |
+
"grad_norm": 0.291015625,
|
| 266 |
+
"learning_rate": 4.999791074638593e-07,
|
| 267 |
+
"logits/chosen": -2.195509433746338,
|
| 268 |
+
"logits/rejected": -1.5012781620025635,
|
| 269 |
+
"logps/chosen": -216.47579956054688,
|
| 270 |
+
"logps/rejected": -207.82186889648438,
|
| 271 |
+
"loss": 0.6904,
|
| 272 |
+
"rewards/accuracies": 0.800000011920929,
|
| 273 |
+
"rewards/chosen": 0.005214618518948555,
|
| 274 |
+
"rewards/margins": 0.006003814749419689,
|
| 275 |
+
"rewards/margins_max": 0.006003814749419689,
|
| 276 |
+
"rewards/margins_min": 0.006003814749419689,
|
| 277 |
+
"rewards/margins_std": NaN,
|
| 278 |
+
"rewards/rejected": -0.0007891958812251687,
|
| 279 |
+
"step": 140
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.11,
|
| 283 |
+
"grad_norm": 0.30078125,
|
| 284 |
+
"learning_rate": 4.998119881260575e-07,
|
| 285 |
+
"logits/chosen": -2.0202810764312744,
|
| 286 |
+
"logits/rejected": -1.4483054876327515,
|
| 287 |
+
"logps/chosen": -212.3457489013672,
|
| 288 |
+
"logps/rejected": -211.62197875976562,
|
| 289 |
+
"loss": 0.689,
|
| 290 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 291 |
+
"rewards/chosen": 0.004780811257660389,
|
| 292 |
+
"rewards/margins": 0.008003057911992073,
|
| 293 |
+
"rewards/margins_max": 0.008003057911992073,
|
| 294 |
+
"rewards/margins_min": 0.008003057911992073,
|
| 295 |
+
"rewards/margins_std": NaN,
|
| 296 |
+
"rewards/rejected": -0.003222248051315546,
|
| 297 |
+
"step": 150
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.12,
|
| 301 |
+
"grad_norm": 0.30859375,
|
| 302 |
+
"learning_rate": 4.99477861175283e-07,
|
| 303 |
+
"logits/chosen": -2.0366463661193848,
|
| 304 |
+
"logits/rejected": -1.3878607749938965,
|
| 305 |
+
"logps/chosen": -231.7950439453125,
|
| 306 |
+
"logps/rejected": -239.493896484375,
|
| 307 |
+
"loss": 0.6886,
|
| 308 |
+
"rewards/accuracies": 1.0,
|
| 309 |
+
"rewards/chosen": 0.007523299660533667,
|
| 310 |
+
"rewards/margins": 0.0094174575060606,
|
| 311 |
+
"rewards/margins_max": 0.0094174575060606,
|
| 312 |
+
"rewards/margins_min": 0.0094174575060606,
|
| 313 |
+
"rewards/margins_std": NaN,
|
| 314 |
+
"rewards/rejected": -0.0018941573798656464,
|
| 315 |
+
"step": 160
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.13,
|
| 319 |
+
"grad_norm": 0.30078125,
|
| 320 |
+
"learning_rate": 4.989769499865024e-07,
|
| 321 |
+
"logits/chosen": -1.8927358388900757,
|
| 322 |
+
"logits/rejected": -1.4275254011154175,
|
| 323 |
+
"logps/chosen": -239.3990936279297,
|
| 324 |
+
"logps/rejected": -282.657958984375,
|
| 325 |
+
"loss": 0.6885,
|
| 326 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 327 |
+
"rewards/chosen": 0.009550971910357475,
|
| 328 |
+
"rewards/margins": 0.011694083921611309,
|
| 329 |
+
"rewards/margins_max": 0.011694083921611309,
|
| 330 |
+
"rewards/margins_min": 0.011694083921611309,
|
| 331 |
+
"rewards/margins_std": NaN,
|
| 332 |
+
"rewards/rejected": -0.0021431122440844774,
|
| 333 |
+
"step": 170
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.13,
|
| 337 |
+
"grad_norm": 0.32421875,
|
| 338 |
+
"learning_rate": 4.983095894354857e-07,
|
| 339 |
+
"logits/chosen": -2.0323657989501953,
|
| 340 |
+
"logits/rejected": -1.417519211769104,
|
| 341 |
+
"logps/chosen": -321.04107666015625,
|
| 342 |
+
"logps/rejected": -304.40753173828125,
|
| 343 |
+
"loss": 0.6874,
|
| 344 |
+
"rewards/accuracies": 0.949999988079071,
|
| 345 |
+
"rewards/chosen": 0.006479477044194937,
|
| 346 |
+
"rewards/margins": 0.010976837947964668,
|
| 347 |
+
"rewards/margins_max": 0.010976837947964668,
|
| 348 |
+
"rewards/margins_min": 0.010976837947964668,
|
| 349 |
+
"rewards/margins_std": NaN,
|
| 350 |
+
"rewards/rejected": -0.00449735950678587,
|
| 351 |
+
"step": 180
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.14,
|
| 355 |
+
"grad_norm": 0.3046875,
|
| 356 |
+
"learning_rate": 4.974762256749318e-07,
|
| 357 |
+
"logits/chosen": -2.036602020263672,
|
| 358 |
+
"logits/rejected": -1.5517603158950806,
|
| 359 |
+
"logps/chosen": -231.96826171875,
|
| 360 |
+
"logps/rejected": -237.5748291015625,
|
| 361 |
+
"loss": 0.6869,
|
| 362 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 363 |
+
"rewards/chosen": 0.005657371133565903,
|
| 364 |
+
"rewards/margins": 0.008646010421216488,
|
| 365 |
+
"rewards/margins_max": 0.008646010421216488,
|
| 366 |
+
"rewards/margins_min": 0.008646010421216488,
|
| 367 |
+
"rewards/margins_std": NaN,
|
| 368 |
+
"rewards/rejected": -0.0029886397533118725,
|
| 369 |
+
"step": 190
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.15,
|
| 373 |
+
"grad_norm": 0.326171875,
|
| 374 |
+
"learning_rate": 4.96477415836199e-07,
|
| 375 |
+
"logits/chosen": -1.938863754272461,
|
| 376 |
+
"logits/rejected": -1.2694227695465088,
|
| 377 |
+
"logps/chosen": -244.3894805908203,
|
| 378 |
+
"logps/rejected": -208.3003387451172,
|
| 379 |
+
"loss": 0.6864,
|
| 380 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 381 |
+
"rewards/chosen": 0.005547370761632919,
|
| 382 |
+
"rewards/margins": 0.012681066989898682,
|
| 383 |
+
"rewards/margins_max": 0.012681066989898682,
|
| 384 |
+
"rewards/margins_min": 0.012681066989898682,
|
| 385 |
+
"rewards/margins_std": NaN,
|
| 386 |
+
"rewards/rejected": -0.007133697625249624,
|
| 387 |
+
"step": 200
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.16,
|
| 391 |
+
"grad_norm": 0.267578125,
|
| 392 |
+
"learning_rate": 4.953138276568461e-07,
|
| 393 |
+
"logits/chosen": -1.997301697731018,
|
| 394 |
+
"logits/rejected": -1.3399479389190674,
|
| 395 |
+
"logps/chosen": -301.72332763671875,
|
| 396 |
+
"logps/rejected": -227.8885955810547,
|
| 397 |
+
"loss": 0.6853,
|
| 398 |
+
"rewards/accuracies": 0.949999988079071,
|
| 399 |
+
"rewards/chosen": 0.008855095133185387,
|
| 400 |
+
"rewards/margins": 0.019029559567570686,
|
| 401 |
+
"rewards/margins_max": 0.019029559567570686,
|
| 402 |
+
"rewards/margins_min": 0.019029559567570686,
|
| 403 |
+
"rewards/margins_std": NaN,
|
| 404 |
+
"rewards/rejected": -0.010174466297030449,
|
| 405 |
+
"step": 210
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.16,
|
| 409 |
+
"grad_norm": 0.353515625,
|
| 410 |
+
"learning_rate": 4.939862390342259e-07,
|
| 411 |
+
"logits/chosen": -1.9495147466659546,
|
| 412 |
+
"logits/rejected": -1.5158745050430298,
|
| 413 |
+
"logps/chosen": -216.283935546875,
|
| 414 |
+
"logps/rejected": -199.36839294433594,
|
| 415 |
+
"loss": 0.6843,
|
| 416 |
+
"rewards/accuracies": 0.949999988079071,
|
| 417 |
+
"rewards/chosen": 0.00766110559925437,
|
| 418 |
+
"rewards/margins": 0.01607651077210903,
|
| 419 |
+
"rewards/margins_max": 0.01607651077210903,
|
| 420 |
+
"rewards/margins_min": 0.01607651077210903,
|
| 421 |
+
"rewards/margins_std": NaN,
|
| 422 |
+
"rewards/rejected": -0.00841540563851595,
|
| 423 |
+
"step": 220
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.17,
|
| 427 |
+
"grad_norm": 0.29296875,
|
| 428 |
+
"learning_rate": 4.924955375054358e-07,
|
| 429 |
+
"logits/chosen": -2.16361665725708,
|
| 430 |
+
"logits/rejected": -1.274595022201538,
|
| 431 |
+
"logps/chosen": -189.79823303222656,
|
| 432 |
+
"logps/rejected": -198.684814453125,
|
| 433 |
+
"loss": 0.6845,
|
| 434 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 435 |
+
"rewards/chosen": 0.012809082865715027,
|
| 436 |
+
"rewards/margins": 0.015618041157722473,
|
| 437 |
+
"rewards/margins_max": 0.015618041157722473,
|
| 438 |
+
"rewards/margins_min": 0.015618041157722473,
|
| 439 |
+
"rewards/margins_std": NaN,
|
| 440 |
+
"rewards/rejected": -0.002808959688991308,
|
| 441 |
+
"step": 230
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.18,
|
| 445 |
+
"grad_norm": 0.337890625,
|
| 446 |
+
"learning_rate": 4.908427196539701e-07,
|
| 447 |
+
"logits/chosen": -2.0759100914001465,
|
| 448 |
+
"logits/rejected": -1.4362943172454834,
|
| 449 |
+
"logps/chosen": -224.01724243164062,
|
| 450 |
+
"logps/rejected": -210.9996337890625,
|
| 451 |
+
"loss": 0.6831,
|
| 452 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 453 |
+
"rewards/chosen": 0.011357143521308899,
|
| 454 |
+
"rewards/margins": 0.015575545839965343,
|
| 455 |
+
"rewards/margins_max": 0.015575545839965343,
|
| 456 |
+
"rewards/margins_min": 0.015575545839965343,
|
| 457 |
+
"rewards/margins_std": NaN,
|
| 458 |
+
"rewards/rejected": -0.0042184023186564445,
|
| 459 |
+
"step": 240
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.19,
|
| 463 |
+
"grad_norm": 0.337890625,
|
| 464 |
+
"learning_rate": 4.890288904434699e-07,
|
| 465 |
+
"logits/chosen": -2.0894532203674316,
|
| 466 |
+
"logits/rejected": -1.2666500806808472,
|
| 467 |
+
"logps/chosen": -258.41571044921875,
|
| 468 |
+
"logps/rejected": -249.0290985107422,
|
| 469 |
+
"loss": 0.6807,
|
| 470 |
+
"rewards/accuracies": 1.0,
|
| 471 |
+
"rewards/chosen": 0.026370421051979065,
|
| 472 |
+
"rewards/margins": 0.03443429246544838,
|
| 473 |
+
"rewards/margins_max": 0.03443429246544838,
|
| 474 |
+
"rewards/margins_min": 0.03443429246544838,
|
| 475 |
+
"rewards/margins_std": NaN,
|
| 476 |
+
"rewards/rejected": -0.008063869550824165,
|
| 477 |
+
"step": 250
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.19,
|
| 481 |
+
"grad_norm": 0.310546875,
|
| 482 |
+
"learning_rate": 4.87055262479019e-07,
|
| 483 |
+
"logits/chosen": -2.0931239128112793,
|
| 484 |
+
"logits/rejected": -1.5411086082458496,
|
| 485 |
+
"logps/chosen": -231.93203735351562,
|
| 486 |
+
"logps/rejected": -209.02432250976562,
|
| 487 |
+
"loss": 0.6831,
|
| 488 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 489 |
+
"rewards/chosen": 0.012660378590226173,
|
| 490 |
+
"rewards/margins": 0.020816108211874962,
|
| 491 |
+
"rewards/margins_max": 0.020816108211874962,
|
| 492 |
+
"rewards/margins_min": 0.020816108211874962,
|
| 493 |
+
"rewards/margins_std": NaN,
|
| 494 |
+
"rewards/rejected": -0.008155731484293938,
|
| 495 |
+
"step": 260
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.2,
|
| 499 |
+
"grad_norm": 0.337890625,
|
| 500 |
+
"learning_rate": 4.849231551964771e-07,
|
| 501 |
+
"logits/chosen": -2.129526138305664,
|
| 502 |
+
"logits/rejected": -1.3961608409881592,
|
| 503 |
+
"logps/chosen": -257.1175231933594,
|
| 504 |
+
"logps/rejected": -238.4982147216797,
|
| 505 |
+
"loss": 0.6812,
|
| 506 |
+
"rewards/accuracies": 0.949999988079071,
|
| 507 |
+
"rewards/chosen": 0.022793341428041458,
|
| 508 |
+
"rewards/margins": 0.02573349140584469,
|
| 509 |
+
"rewards/margins_max": 0.02573349140584469,
|
| 510 |
+
"rewards/margins_min": 0.02573349140584469,
|
| 511 |
+
"rewards/margins_std": NaN,
|
| 512 |
+
"rewards/rejected": -0.00294014741666615,
|
| 513 |
+
"step": 270
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.21,
|
| 517 |
+
"grad_norm": 0.34765625,
|
| 518 |
+
"learning_rate": 4.826339939803922e-07,
|
| 519 |
+
"logits/chosen": -2.051805257797241,
|
| 520 |
+
"logits/rejected": -1.4157917499542236,
|
| 521 |
+
"logps/chosen": -226.42919921875,
|
| 522 |
+
"logps/rejected": -268.73980712890625,
|
| 523 |
+
"loss": 0.6803,
|
| 524 |
+
"rewards/accuracies": 0.949999988079071,
|
| 525 |
+
"rewards/chosen": 0.016779452562332153,
|
| 526 |
+
"rewards/margins": 0.027814749628305435,
|
| 527 |
+
"rewards/margins_max": 0.027814749628305435,
|
| 528 |
+
"rewards/margins_min": 0.027814749628305435,
|
| 529 |
+
"rewards/margins_std": NaN,
|
| 530 |
+
"rewards/rejected": -0.011035297997295856,
|
| 531 |
+
"step": 280
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.21,
|
| 535 |
+
"grad_norm": 0.2890625,
|
| 536 |
+
"learning_rate": 4.801893092110846e-07,
|
| 537 |
+
"logits/chosen": -2.0515878200531006,
|
| 538 |
+
"logits/rejected": -1.403578519821167,
|
| 539 |
+
"logps/chosen": -200.06935119628906,
|
| 540 |
+
"logps/rejected": -222.7530059814453,
|
| 541 |
+
"loss": 0.6798,
|
| 542 |
+
"rewards/accuracies": 0.949999988079071,
|
| 543 |
+
"rewards/chosen": 0.02002173662185669,
|
| 544 |
+
"rewards/margins": 0.03395923227071762,
|
| 545 |
+
"rewards/margins_max": 0.03395923227071762,
|
| 546 |
+
"rewards/margins_min": 0.03395923227071762,
|
| 547 |
+
"rewards/margins_std": NaN,
|
| 548 |
+
"rewards/rejected": -0.013937492854893208,
|
| 549 |
+
"step": 290
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.22,
|
| 553 |
+
"grad_norm": 0.34375,
|
| 554 |
+
"learning_rate": 4.775907352415367e-07,
|
| 555 |
+
"logits/chosen": -2.056366443634033,
|
| 556 |
+
"logits/rejected": -1.4939775466918945,
|
| 557 |
+
"logps/chosen": -227.35498046875,
|
| 558 |
+
"logps/rejected": -234.95849609375,
|
| 559 |
+
"loss": 0.6792,
|
| 560 |
+
"rewards/accuracies": 1.0,
|
| 561 |
+
"rewards/chosen": 0.017659524455666542,
|
| 562 |
+
"rewards/margins": 0.028245586901903152,
|
| 563 |
+
"rewards/margins_max": 0.028245586901903152,
|
| 564 |
+
"rewards/margins_min": 0.028245586901903152,
|
| 565 |
+
"rewards/margins_std": NaN,
|
| 566 |
+
"rewards/rejected": -0.01058606244623661,
|
| 567 |
+
"step": 300
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.23,
|
| 571 |
+
"grad_norm": 0.359375,
|
| 572 |
+
"learning_rate": 4.7484000930477327e-07,
|
| 573 |
+
"logits/chosen": -2.1390914916992188,
|
| 574 |
+
"logits/rejected": -1.4254201650619507,
|
| 575 |
+
"logps/chosen": -258.49652099609375,
|
| 576 |
+
"logps/rejected": -261.0328063964844,
|
| 577 |
+
"loss": 0.6774,
|
| 578 |
+
"rewards/accuracies": 0.949999988079071,
|
| 579 |
+
"rewards/chosen": 0.02551867440342903,
|
| 580 |
+
"rewards/margins": 0.03578513488173485,
|
| 581 |
+
"rewards/margins_max": 0.03578513488173485,
|
| 582 |
+
"rewards/margins_min": 0.03578513488173485,
|
| 583 |
+
"rewards/margins_std": NaN,
|
| 584 |
+
"rewards/rejected": -0.010266458615660667,
|
| 585 |
+
"step": 310
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.24,
|
| 589 |
+
"grad_norm": 0.359375,
|
| 590 |
+
"learning_rate": 4.71938970352464e-07,
|
| 591 |
+
"logits/chosen": -2.0027215480804443,
|
| 592 |
+
"logits/rejected": -1.6378333568572998,
|
| 593 |
+
"logps/chosen": -177.40463256835938,
|
| 594 |
+
"logps/rejected": -213.1610107421875,
|
| 595 |
+
"loss": 0.6789,
|
| 596 |
+
"rewards/accuracies": 0.949999988079071,
|
| 597 |
+
"rewards/chosen": 0.01185548771172762,
|
| 598 |
+
"rewards/margins": 0.021204955875873566,
|
| 599 |
+
"rewards/margins_max": 0.021204955875873566,
|
| 600 |
+
"rewards/margins_min": 0.021204955875873566,
|
| 601 |
+
"rewards/margins_std": NaN,
|
| 602 |
+
"rewards/rejected": -0.009349468164145947,
|
| 603 |
+
"step": 320
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.24,
|
| 607 |
+
"grad_norm": 0.349609375,
|
| 608 |
+
"learning_rate": 4.688895578255227e-07,
|
| 609 |
+
"logits/chosen": -2.0879385471343994,
|
| 610 |
+
"logits/rejected": -1.248313546180725,
|
| 611 |
+
"logps/chosen": -262.8072509765625,
|
| 612 |
+
"logps/rejected": -248.21780395507812,
|
| 613 |
+
"loss": 0.6765,
|
| 614 |
+
"rewards/accuracies": 0.949999988079071,
|
| 615 |
+
"rewards/chosen": 0.01555053610354662,
|
| 616 |
+
"rewards/margins": 0.032046493142843246,
|
| 617 |
+
"rewards/margins_max": 0.032046493142843246,
|
| 618 |
+
"rewards/margins_min": 0.032046493142843246,
|
| 619 |
+
"rewards/margins_std": NaN,
|
| 620 |
+
"rewards/rejected": -0.016495956107974052,
|
| 621 |
+
"step": 330
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.25,
|
| 625 |
+
"grad_norm": 0.34765625,
|
| 626 |
+
"learning_rate": 4.656938103575271e-07,
|
| 627 |
+
"logits/chosen": -2.172673463821411,
|
| 628 |
+
"logits/rejected": -1.4102470874786377,
|
| 629 |
+
"logps/chosen": -270.31378173828125,
|
| 630 |
+
"logps/rejected": -231.1504364013672,
|
| 631 |
+
"loss": 0.6758,
|
| 632 |
+
"rewards/accuracies": 1.0,
|
| 633 |
+
"rewards/chosen": 0.023389603942632675,
|
| 634 |
+
"rewards/margins": 0.043514709919691086,
|
| 635 |
+
"rewards/margins_max": 0.043514709919691086,
|
| 636 |
+
"rewards/margins_min": 0.043514709919691086,
|
| 637 |
+
"rewards/margins_std": NaN,
|
| 638 |
+
"rewards/rejected": -0.02012510970234871,
|
| 639 |
+
"step": 340
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.26,
|
| 643 |
+
"grad_norm": 0.28125,
|
| 644 |
+
"learning_rate": 4.6235386441182434e-07,
|
| 645 |
+
"logits/chosen": -2.126311779022217,
|
| 646 |
+
"logits/rejected": -1.5584805011749268,
|
| 647 |
+
"logps/chosen": -217.91452026367188,
|
| 648 |
+
"logps/rejected": -242.47653198242188,
|
| 649 |
+
"loss": 0.6768,
|
| 650 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 651 |
+
"rewards/chosen": 0.02275378443300724,
|
| 652 |
+
"rewards/margins": 0.031895603984594345,
|
| 653 |
+
"rewards/margins_max": 0.031895603984594345,
|
| 654 |
+
"rewards/margins_min": 0.031895603984594345,
|
| 655 |
+
"rewards/margins_std": NaN,
|
| 656 |
+
"rewards/rejected": -0.009141822345554829,
|
| 657 |
+
"step": 350
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.27,
|
| 661 |
+
"grad_norm": 0.33203125,
|
| 662 |
+
"learning_rate": 4.588719528532341e-07,
|
| 663 |
+
"logits/chosen": -2.138247013092041,
|
| 664 |
+
"logits/rejected": -1.4281116724014282,
|
| 665 |
+
"logps/chosen": -227.18411254882812,
|
| 666 |
+
"logps/rejected": -220.9769287109375,
|
| 667 |
+
"loss": 0.6768,
|
| 668 |
+
"rewards/accuracies": 1.0,
|
| 669 |
+
"rewards/chosen": 0.020104510709643364,
|
| 670 |
+
"rewards/margins": 0.04013504832983017,
|
| 671 |
+
"rewards/margins_max": 0.04013504832983017,
|
| 672 |
+
"rewards/margins_min": 0.04013504832983017,
|
| 673 |
+
"rewards/margins_std": NaN,
|
| 674 |
+
"rewards/rejected": -0.020030539482831955,
|
| 675 |
+
"step": 360
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.27,
|
| 679 |
+
"grad_norm": 0.349609375,
|
| 680 |
+
"learning_rate": 4.5525040345530465e-07,
|
| 681 |
+
"logits/chosen": -2.065033435821533,
|
| 682 |
+
"logits/rejected": -1.3876053094863892,
|
| 683 |
+
"logps/chosen": -255.1400604248047,
|
| 684 |
+
"logps/rejected": -341.7924499511719,
|
| 685 |
+
"loss": 0.6745,
|
| 686 |
+
"rewards/accuracies": 1.0,
|
| 687 |
+
"rewards/chosen": 0.030053580179810524,
|
| 688 |
+
"rewards/margins": 0.04531164467334747,
|
| 689 |
+
"rewards/margins_max": 0.04531164467334747,
|
| 690 |
+
"rewards/margins_min": 0.04531164467334747,
|
| 691 |
+
"rewards/margins_std": NaN,
|
| 692 |
+
"rewards/rejected": -0.015258069150149822,
|
| 693 |
+
"step": 370
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.28,
|
| 697 |
+
"grad_norm": 0.357421875,
|
| 698 |
+
"learning_rate": 4.5149163734411855e-07,
|
| 699 |
+
"logits/chosen": -1.9663784503936768,
|
| 700 |
+
"logits/rejected": -1.4839365482330322,
|
| 701 |
+
"logps/chosen": -257.76336669921875,
|
| 702 |
+
"logps/rejected": -252.73291015625,
|
| 703 |
+
"loss": 0.6729,
|
| 704 |
+
"rewards/accuracies": 1.0,
|
| 705 |
+
"rewards/chosen": 0.01958070509135723,
|
| 706 |
+
"rewards/margins": 0.04934977740049362,
|
| 707 |
+
"rewards/margins_max": 0.04934977740049362,
|
| 708 |
+
"rewards/margins_min": 0.04934977740049362,
|
| 709 |
+
"rewards/margins_std": NaN,
|
| 710 |
+
"rewards/rejected": -0.02976907417178154,
|
| 711 |
+
"step": 380
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.29,
|
| 715 |
+
"grad_norm": 0.314453125,
|
| 716 |
+
"learning_rate": 4.475981673796898e-07,
|
| 717 |
+
"logits/chosen": -2.087552309036255,
|
| 718 |
+
"logits/rejected": -1.1439874172210693,
|
| 719 |
+
"logps/chosen": -238.96328735351562,
|
| 720 |
+
"logps/rejected": -177.21859741210938,
|
| 721 |
+
"loss": 0.6734,
|
| 722 |
+
"rewards/accuracies": 1.0,
|
| 723 |
+
"rewards/chosen": 0.028685113415122032,
|
| 724 |
+
"rewards/margins": 0.04815841466188431,
|
| 725 |
+
"rewards/margins_max": 0.04815841466188431,
|
| 726 |
+
"rewards/margins_min": 0.04815841466188431,
|
| 727 |
+
"rewards/margins_std": NaN,
|
| 728 |
+
"rewards/rejected": -0.019473304972052574,
|
| 729 |
+
"step": 390
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.3,
|
| 733 |
+
"grad_norm": 0.33203125,
|
| 734 |
+
"learning_rate": 4.435725964760331e-07,
|
| 735 |
+
"logits/chosen": -2.0537381172180176,
|
| 736 |
+
"logits/rejected": -1.5530219078063965,
|
| 737 |
+
"logps/chosen": -198.19747924804688,
|
| 738 |
+
"logps/rejected": -225.1847686767578,
|
| 739 |
+
"loss": 0.6738,
|
| 740 |
+
"rewards/accuracies": 1.0,
|
| 741 |
+
"rewards/chosen": 0.025270873680710793,
|
| 742 |
+
"rewards/margins": 0.04031512513756752,
|
| 743 |
+
"rewards/margins_max": 0.04031512513756752,
|
| 744 |
+
"rewards/margins_min": 0.04031512513756752,
|
| 745 |
+
"rewards/margins_std": NaN,
|
| 746 |
+
"rewards/rejected": -0.01504424773156643,
|
| 747 |
+
"step": 400
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.3,
|
| 751 |
+
"grad_norm": 0.337890625,
|
| 752 |
+
"learning_rate": 4.3941761586102946e-07,
|
| 753 |
+
"logits/chosen": -2.0906436443328857,
|
| 754 |
+
"logits/rejected": -1.287006139755249,
|
| 755 |
+
"logps/chosen": -254.64700317382812,
|
| 756 |
+
"logps/rejected": -214.3684539794922,
|
| 757 |
+
"loss": 0.6727,
|
| 758 |
+
"rewards/accuracies": 0.949999988079071,
|
| 759 |
+
"rewards/chosen": 0.029230216518044472,
|
| 760 |
+
"rewards/margins": 0.04698926955461502,
|
| 761 |
+
"rewards/margins_max": 0.04698926955461502,
|
| 762 |
+
"rewards/margins_min": 0.04698926955461502,
|
| 763 |
+
"rewards/margins_std": NaN,
|
| 764 |
+
"rewards/rejected": -0.0177590511739254,
|
| 765 |
+
"step": 410
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.31,
|
| 769 |
+
"grad_norm": 0.294921875,
|
| 770 |
+
"learning_rate": 4.3513600327725117e-07,
|
| 771 |
+
"logits/chosen": -1.9869972467422485,
|
| 772 |
+
"logits/rejected": -1.4108494520187378,
|
| 773 |
+
"logps/chosen": -261.79034423828125,
|
| 774 |
+
"logps/rejected": -244.15512084960938,
|
| 775 |
+
"loss": 0.6719,
|
| 776 |
+
"rewards/accuracies": 0.949999988079071,
|
| 777 |
+
"rewards/chosen": 0.02686556614935398,
|
| 778 |
+
"rewards/margins": 0.04367043077945709,
|
| 779 |
+
"rewards/margins_max": 0.04367043077945709,
|
| 780 |
+
"rewards/margins_min": 0.04367043077945709,
|
| 781 |
+
"rewards/margins_std": NaN,
|
| 782 |
+
"rewards/rejected": -0.01680486463010311,
|
| 783 |
+
"step": 420
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.32,
|
| 787 |
+
"grad_norm": 0.361328125,
|
| 788 |
+
"learning_rate": 4.307306211249482e-07,
|
| 789 |
+
"logits/chosen": -1.9715445041656494,
|
| 790 |
+
"logits/rejected": -1.4791746139526367,
|
| 791 |
+
"logps/chosen": -181.6294403076172,
|
| 792 |
+
"logps/rejected": -219.4192352294922,
|
| 793 |
+
"loss": 0.6697,
|
| 794 |
+
"rewards/accuracies": 1.0,
|
| 795 |
+
"rewards/chosen": 0.02723332680761814,
|
| 796 |
+
"rewards/margins": 0.05029752105474472,
|
| 797 |
+
"rewards/margins_max": 0.05029752105474472,
|
| 798 |
+
"rewards/margins_min": 0.05029752105474472,
|
| 799 |
+
"rewards/margins_std": NaN,
|
| 800 |
+
"rewards/rejected": -0.023064199835062027,
|
| 801 |
+
"step": 430
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.33,
|
| 805 |
+
"grad_norm": 0.322265625,
|
| 806 |
+
"learning_rate": 4.26204414548439e-07,
|
| 807 |
+
"logits/chosen": -2.0978024005889893,
|
| 808 |
+
"logits/rejected": -1.3412537574768066,
|
| 809 |
+
"logps/chosen": -257.7082824707031,
|
| 810 |
+
"logps/rejected": -223.86495971679688,
|
| 811 |
+
"loss": 0.6714,
|
| 812 |
+
"rewards/accuracies": 0.949999988079071,
|
| 813 |
+
"rewards/chosen": 0.025436753407120705,
|
| 814 |
+
"rewards/margins": 0.050586361438035965,
|
| 815 |
+
"rewards/margins_max": 0.050586361438035965,
|
| 816 |
+
"rewards/margins_min": 0.050586361438035965,
|
| 817 |
+
"rewards/margins_std": NaN,
|
| 818 |
+
"rewards/rejected": -0.025149613618850708,
|
| 819 |
+
"step": 440
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.33,
|
| 823 |
+
"grad_norm": 0.3046875,
|
| 824 |
+
"learning_rate": 4.2156040946718343e-07,
|
| 825 |
+
"logits/chosen": -2.0477805137634277,
|
| 826 |
+
"logits/rejected": -1.5934501886367798,
|
| 827 |
+
"logps/chosen": -228.3242950439453,
|
| 828 |
+
"logps/rejected": -272.94488525390625,
|
| 829 |
+
"loss": 0.6667,
|
| 830 |
+
"rewards/accuracies": 1.0,
|
| 831 |
+
"rewards/chosen": 0.03539489209651947,
|
| 832 |
+
"rewards/margins": 0.05500010401010513,
|
| 833 |
+
"rewards/margins_max": 0.05500010401010513,
|
| 834 |
+
"rewards/margins_min": 0.05500010401010513,
|
| 835 |
+
"rewards/margins_std": NaN,
|
| 836 |
+
"rewards/rejected": -0.019605208188295364,
|
| 837 |
+
"step": 450
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.34,
|
| 841 |
+
"grad_norm": 0.30859375,
|
| 842 |
+
"learning_rate": 4.1680171055285484e-07,
|
| 843 |
+
"logits/chosen": -2.0926973819732666,
|
| 844 |
+
"logits/rejected": -1.2051217555999756,
|
| 845 |
+
"logps/chosen": -291.447021484375,
|
| 846 |
+
"logps/rejected": -230.220947265625,
|
| 847 |
+
"loss": 0.6697,
|
| 848 |
+
"rewards/accuracies": 1.0,
|
| 849 |
+
"rewards/chosen": 0.027433469891548157,
|
| 850 |
+
"rewards/margins": 0.04781699925661087,
|
| 851 |
+
"rewards/margins_max": 0.04781699925661087,
|
| 852 |
+
"rewards/margins_min": 0.04781699925661087,
|
| 853 |
+
"rewards/margins_std": NaN,
|
| 854 |
+
"rewards/rejected": -0.020383525639772415,
|
| 855 |
+
"step": 460
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.35,
|
| 859 |
+
"grad_norm": 0.3203125,
|
| 860 |
+
"learning_rate": 4.119314991537647e-07,
|
| 861 |
+
"logits/chosen": -2.0807855129241943,
|
| 862 |
+
"logits/rejected": -1.3683123588562012,
|
| 863 |
+
"logps/chosen": -234.7393035888672,
|
| 864 |
+
"logps/rejected": -259.2615051269531,
|
| 865 |
+
"loss": 0.6693,
|
| 866 |
+
"rewards/accuracies": 0.949999988079071,
|
| 867 |
+
"rewards/chosen": 0.03391563892364502,
|
| 868 |
+
"rewards/margins": 0.06610719114542007,
|
| 869 |
+
"rewards/margins_max": 0.06610719114542007,
|
| 870 |
+
"rewards/margins_min": 0.06610719114542007,
|
| 871 |
+
"rewards/margins_std": NaN,
|
| 872 |
+
"rewards/rejected": -0.032191552221775055,
|
| 873 |
+
"step": 470
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.36,
|
| 877 |
+
"grad_norm": 0.328125,
|
| 878 |
+
"learning_rate": 4.0695303116802467e-07,
|
| 879 |
+
"logits/chosen": -1.9418798685073853,
|
| 880 |
+
"logits/rejected": -1.5194308757781982,
|
| 881 |
+
"logps/chosen": -239.70016479492188,
|
| 882 |
+
"logps/rejected": -230.9000701904297,
|
| 883 |
+
"loss": 0.6673,
|
| 884 |
+
"rewards/accuracies": 1.0,
|
| 885 |
+
"rewards/chosen": 0.029919007793068886,
|
| 886 |
+
"rewards/margins": 0.05285835266113281,
|
| 887 |
+
"rewards/margins_max": 0.05285835266113281,
|
| 888 |
+
"rewards/margins_min": 0.05285835266113281,
|
| 889 |
+
"rewards/margins_std": NaN,
|
| 890 |
+
"rewards/rejected": -0.022939343005418777,
|
| 891 |
+
"step": 480
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.36,
|
| 895 |
+
"grad_norm": 0.33203125,
|
| 896 |
+
"learning_rate": 4.0186963486687086e-07,
|
| 897 |
+
"logits/chosen": -2.1395695209503174,
|
| 898 |
+
"logits/rejected": -1.520246148109436,
|
| 899 |
+
"logps/chosen": -181.22784423828125,
|
| 900 |
+
"logps/rejected": -206.8192901611328,
|
| 901 |
+
"loss": 0.6682,
|
| 902 |
+
"rewards/accuracies": 1.0,
|
| 903 |
+
"rewards/chosen": 0.03531926870346069,
|
| 904 |
+
"rewards/margins": 0.05207179859280586,
|
| 905 |
+
"rewards/margins_max": 0.05207179859280586,
|
| 906 |
+
"rewards/margins_min": 0.05207179859280586,
|
| 907 |
+
"rewards/margins_std": NaN,
|
| 908 |
+
"rewards/rejected": -0.01675252802670002,
|
| 909 |
+
"step": 490
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.37,
|
| 913 |
+
"grad_norm": 0.349609375,
|
| 914 |
+
"learning_rate": 3.966847086696045e-07,
|
| 915 |
+
"logits/chosen": -2.017302989959717,
|
| 916 |
+
"logits/rejected": -1.3829624652862549,
|
| 917 |
+
"logps/chosen": -231.52651977539062,
|
| 918 |
+
"logps/rejected": -221.0729217529297,
|
| 919 |
+
"loss": 0.6692,
|
| 920 |
+
"rewards/accuracies": 1.0,
|
| 921 |
+
"rewards/chosen": 0.023424681276082993,
|
| 922 |
+
"rewards/margins": 0.05106498673558235,
|
| 923 |
+
"rewards/margins_max": 0.05106498673558235,
|
| 924 |
+
"rewards/margins_min": 0.05106498673558235,
|
| 925 |
+
"rewards/margins_std": NaN,
|
| 926 |
+
"rewards/rejected": -0.02764030732214451,
|
| 927 |
+
"step": 500
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.38,
|
| 931 |
+
"grad_norm": 0.271484375,
|
| 932 |
+
"learning_rate": 3.9140171887163466e-07,
|
| 933 |
+
"logits/chosen": -1.9985767602920532,
|
| 934 |
+
"logits/rejected": -1.6083755493164062,
|
| 935 |
+
"logps/chosen": -280.05303955078125,
|
| 936 |
+
"logps/rejected": -237.509521484375,
|
| 937 |
+
"loss": 0.6681,
|
| 938 |
+
"rewards/accuracies": 0.949999988079071,
|
| 939 |
+
"rewards/chosen": 0.016823355108499527,
|
| 940 |
+
"rewards/margins": 0.0439620167016983,
|
| 941 |
+
"rewards/margins_max": 0.0439620167016983,
|
| 942 |
+
"rewards/margins_min": 0.0439620167016983,
|
| 943 |
+
"rewards/margins_std": NaN,
|
| 944 |
+
"rewards/rejected": -0.027138665318489075,
|
| 945 |
+
"step": 510
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 0.39,
|
| 949 |
+
"grad_norm": 0.283203125,
|
| 950 |
+
"learning_rate": 3.8602419732714566e-07,
|
| 951 |
+
"logits/chosen": -2.036795139312744,
|
| 952 |
+
"logits/rejected": -1.471296787261963,
|
| 953 |
+
"logps/chosen": -256.37481689453125,
|
| 954 |
+
"logps/rejected": -199.65538024902344,
|
| 955 |
+
"loss": 0.6697,
|
| 956 |
+
"rewards/accuracies": 1.0,
|
| 957 |
+
"rewards/chosen": 0.026260962709784508,
|
| 958 |
+
"rewards/margins": 0.038443367928266525,
|
| 959 |
+
"rewards/margins_max": 0.038443367928266525,
|
| 960 |
+
"rewards/margins_min": 0.038443367928266525,
|
| 961 |
+
"rewards/margins_std": NaN,
|
| 962 |
+
"rewards/rejected": -0.012182402424514294,
|
| 963 |
+
"step": 520
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.39,
|
| 967 |
+
"grad_norm": 0.322265625,
|
| 968 |
+
"learning_rate": 3.805557390879346e-07,
|
| 969 |
+
"logits/chosen": -1.9529104232788086,
|
| 970 |
+
"logits/rejected": -1.5426490306854248,
|
| 971 |
+
"logps/chosen": -182.79574584960938,
|
| 972 |
+
"logps/rejected": -194.38156127929688,
|
| 973 |
+
"loss": 0.6679,
|
| 974 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 975 |
+
"rewards/chosen": 0.020470619201660156,
|
| 976 |
+
"rewards/margins": 0.049423329532146454,
|
| 977 |
+
"rewards/margins_max": 0.049423329532146454,
|
| 978 |
+
"rewards/margins_min": 0.049423329532146454,
|
| 979 |
+
"rewards/margins_std": NaN,
|
| 980 |
+
"rewards/rejected": -0.02895270846784115,
|
| 981 |
+
"step": 530
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 0.4,
|
| 985 |
+
"grad_norm": 0.314453125,
|
| 986 |
+
"learning_rate": 3.75e-07,
|
| 987 |
+
"logits/chosen": -2.0001742839813232,
|
| 988 |
+
"logits/rejected": -1.2963764667510986,
|
| 989 |
+
"logps/chosen": -230.01394653320312,
|
| 990 |
+
"logps/rejected": -232.5394287109375,
|
| 991 |
+
"loss": 0.6667,
|
| 992 |
+
"rewards/accuracies": 0.949999988079071,
|
| 993 |
+
"rewards/chosen": 0.037200070917606354,
|
| 994 |
+
"rewards/margins": 0.0586659200489521,
|
| 995 |
+
"rewards/margins_max": 0.0586659200489521,
|
| 996 |
+
"rewards/margins_min": 0.0586659200489521,
|
| 997 |
+
"rewards/margins_std": NaN,
|
| 998 |
+
"rewards/rejected": -0.0214658472687006,
|
| 999 |
+
"step": 540
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.41,
|
| 1003 |
+
"grad_norm": 0.318359375,
|
| 1004 |
+
"learning_rate": 3.693606942594872e-07,
|
| 1005 |
+
"logits/chosen": -2.0440757274627686,
|
| 1006 |
+
"logits/rejected": -1.3815096616744995,
|
| 1007 |
+
"logps/chosen": -217.4425811767578,
|
| 1008 |
+
"logps/rejected": -171.90591430664062,
|
| 1009 |
+
"loss": 0.6676,
|
| 1010 |
+
"rewards/accuracies": 1.0,
|
| 1011 |
+
"rewards/chosen": 0.02869860827922821,
|
| 1012 |
+
"rewards/margins": 0.0414133220911026,
|
| 1013 |
+
"rewards/margins_max": 0.0414133220911026,
|
| 1014 |
+
"rewards/margins_min": 0.0414133220911026,
|
| 1015 |
+
"rewards/margins_std": NaN,
|
| 1016 |
+
"rewards/rejected": -0.012714711017906666,
|
| 1017 |
+
"step": 550
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.41,
|
| 1021 |
+
"grad_norm": 0.345703125,
|
| 1022 |
+
"learning_rate": 3.636415919296251e-07,
|
| 1023 |
+
"logits/chosen": -1.9253566265106201,
|
| 1024 |
+
"logits/rejected": -1.6695436239242554,
|
| 1025 |
+
"logps/chosen": -171.96823120117188,
|
| 1026 |
+
"logps/rejected": -167.778076171875,
|
| 1027 |
+
"loss": 0.667,
|
| 1028 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1029 |
+
"rewards/chosen": 0.02604065276682377,
|
| 1030 |
+
"rewards/margins": 0.04568729177117348,
|
| 1031 |
+
"rewards/margins_max": 0.04568729177117348,
|
| 1032 |
+
"rewards/margins_min": 0.04568729177117348,
|
| 1033 |
+
"rewards/margins_std": NaN,
|
| 1034 |
+
"rewards/rejected": -0.01964663527905941,
|
| 1035 |
+
"step": 560
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.42,
|
| 1039 |
+
"grad_norm": 0.3359375,
|
| 1040 |
+
"learning_rate": 3.5784651642031337e-07,
|
| 1041 |
+
"logits/chosen": -1.8566385507583618,
|
| 1042 |
+
"logits/rejected": -1.6557972431182861,
|
| 1043 |
+
"logps/chosen": -224.4451904296875,
|
| 1044 |
+
"logps/rejected": -270.01776123046875,
|
| 1045 |
+
"loss": 0.6701,
|
| 1046 |
+
"rewards/accuracies": 1.0,
|
| 1047 |
+
"rewards/chosen": 0.004797924309968948,
|
| 1048 |
+
"rewards/margins": 0.041053663939237595,
|
| 1049 |
+
"rewards/margins_max": 0.041053663939237595,
|
| 1050 |
+
"rewards/margins_min": 0.041053663939237595,
|
| 1051 |
+
"rewards/margins_std": NaN,
|
| 1052 |
+
"rewards/rejected": -0.036255739629268646,
|
| 1053 |
+
"step": 570
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.43,
|
| 1057 |
+
"grad_norm": 0.353515625,
|
| 1058 |
+
"learning_rate": 3.519793419320459e-07,
|
| 1059 |
+
"logits/chosen": -1.9248411655426025,
|
| 1060 |
+
"logits/rejected": -1.1966019868850708,
|
| 1061 |
+
"logps/chosen": -284.9685974121094,
|
| 1062 |
+
"logps/rejected": -223.8994140625,
|
| 1063 |
+
"loss": 0.6664,
|
| 1064 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1065 |
+
"rewards/chosen": 0.024664491415023804,
|
| 1066 |
+
"rewards/margins": 0.05492042750120163,
|
| 1067 |
+
"rewards/margins_max": 0.05492042750120163,
|
| 1068 |
+
"rewards/margins_min": 0.05492042750120163,
|
| 1069 |
+
"rewards/margins_std": NaN,
|
| 1070 |
+
"rewards/rejected": -0.030255932360887527,
|
| 1071 |
+
"step": 580
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.44,
|
| 1075 |
+
"grad_norm": 0.333984375,
|
| 1076 |
+
"learning_rate": 3.460439908658794e-07,
|
| 1077 |
+
"logits/chosen": -2.007855176925659,
|
| 1078 |
+
"logits/rejected": -1.1754387617111206,
|
| 1079 |
+
"logps/chosen": -349.209228515625,
|
| 1080 |
+
"logps/rejected": -206.0282440185547,
|
| 1081 |
+
"loss": 0.6687,
|
| 1082 |
+
"rewards/accuracies": 1.0,
|
| 1083 |
+
"rewards/chosen": 0.0192419346421957,
|
| 1084 |
+
"rewards/margins": 0.04281260445713997,
|
| 1085 |
+
"rewards/margins_max": 0.04281260445713997,
|
| 1086 |
+
"rewards/margins_min": 0.04281260445713997,
|
| 1087 |
+
"rewards/margins_std": NaN,
|
| 1088 |
+
"rewards/rejected": -0.023570667952299118,
|
| 1089 |
+
"step": 590
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.44,
|
| 1093 |
+
"grad_norm": 0.34765625,
|
| 1094 |
+
"learning_rate": 3.400444312011776e-07,
|
| 1095 |
+
"logits/chosen": -2.059713840484619,
|
| 1096 |
+
"logits/rejected": -1.321000576019287,
|
| 1097 |
+
"logps/chosen": -265.93463134765625,
|
| 1098 |
+
"logps/rejected": -285.9882507324219,
|
| 1099 |
+
"loss": 0.6689,
|
| 1100 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1101 |
+
"rewards/chosen": 0.0264985803514719,
|
| 1102 |
+
"rewards/margins": 0.05372191593050957,
|
| 1103 |
+
"rewards/margins_max": 0.05372191593050957,
|
| 1104 |
+
"rewards/margins_min": 0.05372191593050957,
|
| 1105 |
+
"rewards/margins_std": NaN,
|
| 1106 |
+
"rewards/rejected": -0.027223343029618263,
|
| 1107 |
+
"step": 600
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 0.45,
|
| 1111 |
+
"grad_norm": 0.326171875,
|
| 1112 |
+
"learning_rate": 3.3398467384288554e-07,
|
| 1113 |
+
"logits/chosen": -2.0164129734039307,
|
| 1114 |
+
"logits/rejected": -1.428724765777588,
|
| 1115 |
+
"logps/chosen": -174.00619506835938,
|
| 1116 |
+
"logps/rejected": -235.1026153564453,
|
| 1117 |
+
"loss": 0.665,
|
| 1118 |
+
"rewards/accuracies": 1.0,
|
| 1119 |
+
"rewards/chosen": 0.018517404794692993,
|
| 1120 |
+
"rewards/margins": 0.06648163497447968,
|
| 1121 |
+
"rewards/margins_max": 0.06648163497447968,
|
| 1122 |
+
"rewards/margins_min": 0.06648163497447968,
|
| 1123 |
+
"rewards/margins_std": NaN,
|
| 1124 |
+
"rewards/rejected": -0.047964222729206085,
|
| 1125 |
+
"step": 610
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.46,
|
| 1129 |
+
"grad_norm": 0.37890625,
|
| 1130 |
+
"learning_rate": 3.2786876994010616e-07,
|
| 1131 |
+
"logits/chosen": -1.9753456115722656,
|
| 1132 |
+
"logits/rejected": -1.5021337270736694,
|
| 1133 |
+
"logps/chosen": -207.2568359375,
|
| 1134 |
+
"logps/rejected": -196.98239135742188,
|
| 1135 |
+
"loss": 0.6654,
|
| 1136 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1137 |
+
"rewards/chosen": 0.026976531371474266,
|
| 1138 |
+
"rewards/margins": 0.05444413423538208,
|
| 1139 |
+
"rewards/margins_max": 0.05444413423538208,
|
| 1140 |
+
"rewards/margins_min": 0.05444413423538208,
|
| 1141 |
+
"rewards/margins_std": NaN,
|
| 1142 |
+
"rewards/rejected": -0.027467604726552963,
|
| 1143 |
+
"step": 620
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.47,
|
| 1147 |
+
"grad_norm": 0.318359375,
|
| 1148 |
+
"learning_rate": 3.2170080817777257e-07,
|
| 1149 |
+
"logits/chosen": -2.0977303981781006,
|
| 1150 |
+
"logits/rejected": -1.3275058269500732,
|
| 1151 |
+
"logps/chosen": -280.2716979980469,
|
| 1152 |
+
"logps/rejected": -243.47763061523438,
|
| 1153 |
+
"loss": 0.6658,
|
| 1154 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1155 |
+
"rewards/chosen": 0.0296818558126688,
|
| 1156 |
+
"rewards/margins": 0.054945122450590134,
|
| 1157 |
+
"rewards/margins_max": 0.054945122450590134,
|
| 1158 |
+
"rewards/margins_min": 0.054945122450590134,
|
| 1159 |
+
"rewards/margins_std": NaN,
|
| 1160 |
+
"rewards/rejected": -0.025263268500566483,
|
| 1161 |
+
"step": 630
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.47,
|
| 1165 |
+
"grad_norm": 0.349609375,
|
| 1166 |
+
"learning_rate": 3.1548491204322647e-07,
|
| 1167 |
+
"logits/chosen": -2.0041770935058594,
|
| 1168 |
+
"logits/rejected": -1.5506256818771362,
|
| 1169 |
+
"logps/chosen": -211.4202423095703,
|
| 1170 |
+
"logps/rejected": -374.6882629394531,
|
| 1171 |
+
"loss": 0.6674,
|
| 1172 |
+
"rewards/accuracies": 1.0,
|
| 1173 |
+
"rewards/chosen": 0.022962290793657303,
|
| 1174 |
+
"rewards/margins": 0.05601681396365166,
|
| 1175 |
+
"rewards/margins_max": 0.05601681396365166,
|
| 1176 |
+
"rewards/margins_min": 0.05601681396365166,
|
| 1177 |
+
"rewards/margins_std": NaN,
|
| 1178 |
+
"rewards/rejected": -0.033054523169994354,
|
| 1179 |
+
"step": 640
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.48,
|
| 1183 |
+
"grad_norm": 0.35546875,
|
| 1184 |
+
"learning_rate": 3.0922523706952976e-07,
|
| 1185 |
+
"logits/chosen": -2.059875249862671,
|
| 1186 |
+
"logits/rejected": -1.407546877861023,
|
| 1187 |
+
"logps/chosen": -208.85415649414062,
|
| 1188 |
+
"logps/rejected": -180.44479370117188,
|
| 1189 |
+
"loss": 0.667,
|
| 1190 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1191 |
+
"rewards/chosen": 0.03480786085128784,
|
| 1192 |
+
"rewards/margins": 0.05192246288061142,
|
| 1193 |
+
"rewards/margins_max": 0.05192246288061142,
|
| 1194 |
+
"rewards/margins_min": 0.05192246288061142,
|
| 1195 |
+
"rewards/margins_std": NaN,
|
| 1196 |
+
"rewards/rejected": -0.017114607617259026,
|
| 1197 |
+
"step": 650
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.49,
|
| 1201 |
+
"grad_norm": 0.359375,
|
| 1202 |
+
"learning_rate": 3.029259680573527e-07,
|
| 1203 |
+
"logits/chosen": -2.1605982780456543,
|
| 1204 |
+
"logits/rejected": -1.3930118083953857,
|
| 1205 |
+
"logps/chosen": -225.4413299560547,
|
| 1206 |
+
"logps/rejected": -222.64828491210938,
|
| 1207 |
+
"loss": 0.6667,
|
| 1208 |
+
"rewards/accuracies": 1.0,
|
| 1209 |
+
"rewards/chosen": 0.02937215566635132,
|
| 1210 |
+
"rewards/margins": 0.053660131990909576,
|
| 1211 |
+
"rewards/margins_max": 0.053660131990909576,
|
| 1212 |
+
"rewards/margins_min": 0.053660131990909576,
|
| 1213 |
+
"rewards/margins_std": NaN,
|
| 1214 |
+
"rewards/rejected": -0.02428797259926796,
|
| 1215 |
+
"step": 660
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 0.5,
|
| 1219 |
+
"grad_norm": 0.28515625,
|
| 1220 |
+
"learning_rate": 2.9659131627729555e-07,
|
| 1221 |
+
"logits/chosen": -2.1441378593444824,
|
| 1222 |
+
"logits/rejected": -1.2556051015853882,
|
| 1223 |
+
"logps/chosen": -273.76507568359375,
|
| 1224 |
+
"logps/rejected": -173.42489624023438,
|
| 1225 |
+
"loss": 0.6632,
|
| 1226 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1227 |
+
"rewards/chosen": 0.039141856133937836,
|
| 1228 |
+
"rewards/margins": 0.05589345097541809,
|
| 1229 |
+
"rewards/margins_max": 0.05589345097541809,
|
| 1230 |
+
"rewards/margins_min": 0.05589345097541809,
|
| 1231 |
+
"rewards/margins_std": NaN,
|
| 1232 |
+
"rewards/rejected": -0.016751592978835106,
|
| 1233 |
+
"step": 670
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 0.5,
|
| 1237 |
+
"grad_norm": 0.3515625,
|
| 1238 |
+
"learning_rate": 2.9022551665451433e-07,
|
| 1239 |
+
"logits/chosen": -2.107771158218384,
|
| 1240 |
+
"logits/rejected": -1.0939693450927734,
|
| 1241 |
+
"logps/chosen": -241.0987091064453,
|
| 1242 |
+
"logps/rejected": -229.3813018798828,
|
| 1243 |
+
"loss": 0.6646,
|
| 1244 |
+
"rewards/accuracies": 1.0,
|
| 1245 |
+
"rewards/chosen": 0.043823134154081345,
|
| 1246 |
+
"rewards/margins": 0.08219417929649353,
|
| 1247 |
+
"rewards/margins_max": 0.08219417929649353,
|
| 1248 |
+
"rewards/margins_min": 0.08219417929649353,
|
| 1249 |
+
"rewards/margins_std": NaN,
|
| 1250 |
+
"rewards/rejected": -0.038371045142412186,
|
| 1251 |
+
"step": 680
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.51,
|
| 1255 |
+
"grad_norm": 0.33984375,
|
| 1256 |
+
"learning_rate": 2.838328249375328e-07,
|
| 1257 |
+
"logits/chosen": -2.0250961780548096,
|
| 1258 |
+
"logits/rejected": -1.5862263441085815,
|
| 1259 |
+
"logps/chosen": -230.3356475830078,
|
| 1260 |
+
"logps/rejected": -250.5562286376953,
|
| 1261 |
+
"loss": 0.6655,
|
| 1262 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1263 |
+
"rewards/chosen": 0.022846568375825882,
|
| 1264 |
+
"rewards/margins": 0.05219893902540207,
|
| 1265 |
+
"rewards/margins_max": 0.05219893902540207,
|
| 1266 |
+
"rewards/margins_min": 0.05219893902540207,
|
| 1267 |
+
"rewards/margins_std": NaN,
|
| 1268 |
+
"rewards/rejected": -0.029352372512221336,
|
| 1269 |
+
"step": 690
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.52,
|
| 1273 |
+
"grad_norm": 0.3046875,
|
| 1274 |
+
"learning_rate": 2.774175148531329e-07,
|
| 1275 |
+
"logits/chosen": -2.0457088947296143,
|
| 1276 |
+
"logits/rejected": -1.532765507698059,
|
| 1277 |
+
"logps/chosen": -224.957763671875,
|
| 1278 |
+
"logps/rejected": -234.0838165283203,
|
| 1279 |
+
"loss": 0.6647,
|
| 1280 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1281 |
+
"rewards/chosen": 0.02778591774404049,
|
| 1282 |
+
"rewards/margins": 0.0492682009935379,
|
| 1283 |
+
"rewards/margins_max": 0.0492682009935379,
|
| 1284 |
+
"rewards/margins_min": 0.0492682009935379,
|
| 1285 |
+
"rewards/margins_std": NaN,
|
| 1286 |
+
"rewards/rejected": -0.021482281386852264,
|
| 1287 |
+
"step": 700
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.53,
|
| 1291 |
+
"grad_norm": 0.2734375,
|
| 1292 |
+
"learning_rate": 2.7098387524922667e-07,
|
| 1293 |
+
"logits/chosen": -1.8189055919647217,
|
| 1294 |
+
"logits/rejected": -1.4590321779251099,
|
| 1295 |
+
"logps/chosen": -187.11215209960938,
|
| 1296 |
+
"logps/rejected": -179.0242156982422,
|
| 1297 |
+
"loss": 0.6658,
|
| 1298 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1299 |
+
"rewards/chosen": 0.01335279643535614,
|
| 1300 |
+
"rewards/margins": 0.04278566688299179,
|
| 1301 |
+
"rewards/margins_max": 0.04278566688299179,
|
| 1302 |
+
"rewards/margins_min": 0.04278566688299179,
|
| 1303 |
+
"rewards/margins_std": NaN,
|
| 1304 |
+
"rewards/rejected": -0.0294328723102808,
|
| 1305 |
+
"step": 710
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.53,
|
| 1309 |
+
"grad_norm": 0.314453125,
|
| 1310 |
+
"learning_rate": 2.6453620722761895e-07,
|
| 1311 |
+
"logits/chosen": -2.1215291023254395,
|
| 1312 |
+
"logits/rejected": -1.4057071208953857,
|
| 1313 |
+
"logps/chosen": -323.5494689941406,
|
| 1314 |
+
"logps/rejected": -305.4034118652344,
|
| 1315 |
+
"loss": 0.6632,
|
| 1316 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1317 |
+
"rewards/chosen": 0.030355731025338173,
|
| 1318 |
+
"rewards/margins": 0.0672043040394783,
|
| 1319 |
+
"rewards/margins_max": 0.0672043040394783,
|
| 1320 |
+
"rewards/margins_min": 0.0672043040394783,
|
| 1321 |
+
"rewards/margins_std": NaN,
|
| 1322 |
+
"rewards/rejected": -0.036848582327365875,
|
| 1323 |
+
"step": 720
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 0.54,
|
| 1327 |
+
"grad_norm": 0.32421875,
|
| 1328 |
+
"learning_rate": 2.580788212685777e-07,
|
| 1329 |
+
"logits/chosen": -2.018144369125366,
|
| 1330 |
+
"logits/rejected": -1.5041452646255493,
|
| 1331 |
+
"logps/chosen": -206.16226196289062,
|
| 1332 |
+
"logps/rejected": -224.09854125976562,
|
| 1333 |
+
"loss": 0.6658,
|
| 1334 |
+
"rewards/accuracies": 1.0,
|
| 1335 |
+
"rewards/chosen": 0.019517984241247177,
|
| 1336 |
+
"rewards/margins": 0.047711052000522614,
|
| 1337 |
+
"rewards/margins_max": 0.047711052000522614,
|
| 1338 |
+
"rewards/margins_min": 0.047711052000522614,
|
| 1339 |
+
"rewards/margins_std": NaN,
|
| 1340 |
+
"rewards/rejected": -0.028193065896630287,
|
| 1341 |
+
"step": 730
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 0.55,
|
| 1345 |
+
"grad_norm": 0.30078125,
|
| 1346 |
+
"learning_rate": 2.516160343491348e-07,
|
| 1347 |
+
"logits/chosen": -2.0774478912353516,
|
| 1348 |
+
"logits/rejected": -1.3242664337158203,
|
| 1349 |
+
"logps/chosen": -261.04791259765625,
|
| 1350 |
+
"logps/rejected": -214.69058227539062,
|
| 1351 |
+
"loss": 0.6654,
|
| 1352 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1353 |
+
"rewards/chosen": 0.03108522854745388,
|
| 1354 |
+
"rewards/margins": 0.07123453170061111,
|
| 1355 |
+
"rewards/margins_max": 0.07123453170061111,
|
| 1356 |
+
"rewards/margins_min": 0.07123453170061111,
|
| 1357 |
+
"rewards/margins_std": NaN,
|
| 1358 |
+
"rewards/rejected": -0.040149301290512085,
|
| 1359 |
+
"step": 740
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 0.56,
|
| 1363 |
+
"grad_norm": 0.337890625,
|
| 1364 |
+
"learning_rate": 2.4515216705704393e-07,
|
| 1365 |
+
"logits/chosen": -2.14136004447937,
|
| 1366 |
+
"logits/rejected": -1.2130528688430786,
|
| 1367 |
+
"logps/chosen": -302.6998291015625,
|
| 1368 |
+
"logps/rejected": -223.78067016601562,
|
| 1369 |
+
"loss": 0.6606,
|
| 1370 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1371 |
+
"rewards/chosen": 0.032542888075113297,
|
| 1372 |
+
"rewards/margins": 0.0818304568529129,
|
| 1373 |
+
"rewards/margins_max": 0.0818304568529129,
|
| 1374 |
+
"rewards/margins_min": 0.0818304568529129,
|
| 1375 |
+
"rewards/margins_std": NaN,
|
| 1376 |
+
"rewards/rejected": -0.04928756505250931,
|
| 1377 |
+
"step": 750
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 0.56,
|
| 1381 |
+
"grad_norm": 0.328125,
|
| 1382 |
+
"learning_rate": 2.386915407023234e-07,
|
| 1383 |
+
"logits/chosen": -2.1050310134887695,
|
| 1384 |
+
"logits/rejected": -1.4062013626098633,
|
| 1385 |
+
"logps/chosen": -249.6138458251953,
|
| 1386 |
+
"logps/rejected": -318.61773681640625,
|
| 1387 |
+
"loss": 0.6605,
|
| 1388 |
+
"rewards/accuracies": 1.0,
|
| 1389 |
+
"rewards/chosen": 0.03660450130701065,
|
| 1390 |
+
"rewards/margins": 0.07939177751541138,
|
| 1391 |
+
"rewards/margins_max": 0.07939177751541138,
|
| 1392 |
+
"rewards/margins_min": 0.07939177751541138,
|
| 1393 |
+
"rewards/margins_std": NaN,
|
| 1394 |
+
"rewards/rejected": -0.042787276208400726,
|
| 1395 |
+
"step": 760
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.57,
|
| 1399 |
+
"grad_norm": 0.296875,
|
| 1400 |
+
"learning_rate": 2.3223847442831753e-07,
|
| 1401 |
+
"logits/chosen": -2.0049426555633545,
|
| 1402 |
+
"logits/rejected": -1.5466521978378296,
|
| 1403 |
+
"logps/chosen": -226.28701782226562,
|
| 1404 |
+
"logps/rejected": -283.0534973144531,
|
| 1405 |
+
"loss": 0.6647,
|
| 1406 |
+
"rewards/accuracies": 1.0,
|
| 1407 |
+
"rewards/chosen": 0.01987474039196968,
|
| 1408 |
+
"rewards/margins": 0.07332300394773483,
|
| 1409 |
+
"rewards/margins_max": 0.07332300394773483,
|
| 1410 |
+
"rewards/margins_min": 0.07332300394773483,
|
| 1411 |
+
"rewards/margins_std": NaN,
|
| 1412 |
+
"rewards/rejected": -0.05344826728105545,
|
| 1413 |
+
"step": 770
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.58,
|
| 1417 |
+
"grad_norm": 0.328125,
|
| 1418 |
+
"learning_rate": 2.2579728232420523e-07,
|
| 1419 |
+
"logits/chosen": -2.082897663116455,
|
| 1420 |
+
"logits/rejected": -1.373595118522644,
|
| 1421 |
+
"logps/chosen": -310.93695068359375,
|
| 1422 |
+
"logps/rejected": -240.2857208251953,
|
| 1423 |
+
"loss": 0.6632,
|
| 1424 |
+
"rewards/accuracies": 1.0,
|
| 1425 |
+
"rewards/chosen": 0.0339638814330101,
|
| 1426 |
+
"rewards/margins": 0.06445649266242981,
|
| 1427 |
+
"rewards/margins_max": 0.06445649266242981,
|
| 1428 |
+
"rewards/margins_min": 0.06445649266242981,
|
| 1429 |
+
"rewards/margins_std": NaN,
|
| 1430 |
+
"rewards/rejected": -0.030492613092064857,
|
| 1431 |
+
"step": 780
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.58,
|
| 1435 |
+
"grad_norm": 0.380859375,
|
| 1436 |
+
"learning_rate": 2.1937227054088868e-07,
|
| 1437 |
+
"logits/chosen": -2.041259288787842,
|
| 1438 |
+
"logits/rejected": -1.4105521440505981,
|
| 1439 |
+
"logps/chosen": -291.784423828125,
|
| 1440 |
+
"logps/rejected": -246.40859985351562,
|
| 1441 |
+
"loss": 0.6643,
|
| 1442 |
+
"rewards/accuracies": 1.0,
|
| 1443 |
+
"rewards/chosen": 0.024001669138669968,
|
| 1444 |
+
"rewards/margins": 0.05560433119535446,
|
| 1445 |
+
"rewards/margins_max": 0.05560433119535446,
|
| 1446 |
+
"rewards/margins_min": 0.05560433119535446,
|
| 1447 |
+
"rewards/margins_std": NaN,
|
| 1448 |
+
"rewards/rejected": -0.031602658331394196,
|
| 1449 |
+
"step": 790
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 0.59,
|
| 1453 |
+
"grad_norm": 0.318359375,
|
| 1454 |
+
"learning_rate": 2.1296773441218785e-07,
|
| 1455 |
+
"logits/chosen": -2.0651352405548096,
|
| 1456 |
+
"logits/rejected": -1.4306925535202026,
|
| 1457 |
+
"logps/chosen": -239.18588256835938,
|
| 1458 |
+
"logps/rejected": -210.4315948486328,
|
| 1459 |
+
"loss": 0.664,
|
| 1460 |
+
"rewards/accuracies": 1.0,
|
| 1461 |
+
"rewards/chosen": 0.026126444339752197,
|
| 1462 |
+
"rewards/margins": 0.06849897652864456,
|
| 1463 |
+
"rewards/margins_max": 0.06849897652864456,
|
| 1464 |
+
"rewards/margins_min": 0.06849897652864456,
|
| 1465 |
+
"rewards/margins_std": NaN,
|
| 1466 |
+
"rewards/rejected": -0.04237252101302147,
|
| 1467 |
+
"step": 800
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 0.6,
|
| 1471 |
+
"grad_norm": 0.298828125,
|
| 1472 |
+
"learning_rate": 2.065879555832674e-07,
|
| 1473 |
+
"logits/chosen": -1.9278161525726318,
|
| 1474 |
+
"logits/rejected": -1.6418416500091553,
|
| 1475 |
+
"logps/chosen": -215.7371063232422,
|
| 1476 |
+
"logps/rejected": -255.38583374023438,
|
| 1477 |
+
"loss": 0.6652,
|
| 1478 |
+
"rewards/accuracies": 1.0,
|
| 1479 |
+
"rewards/chosen": 0.019038062542676926,
|
| 1480 |
+
"rewards/margins": 0.051104627549648285,
|
| 1481 |
+
"rewards/margins_max": 0.051104627549648285,
|
| 1482 |
+
"rewards/margins_min": 0.051104627549648285,
|
| 1483 |
+
"rewards/margins_std": NaN,
|
| 1484 |
+
"rewards/rejected": -0.03206656873226166,
|
| 1485 |
+
"step": 810
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 0.61,
|
| 1489 |
+
"grad_norm": 0.34765625,
|
| 1490 |
+
"learning_rate": 2.0023719914821489e-07,
|
| 1491 |
+
"logits/chosen": -2.149594783782959,
|
| 1492 |
+
"logits/rejected": -1.4589303731918335,
|
| 1493 |
+
"logps/chosen": -219.695556640625,
|
| 1494 |
+
"logps/rejected": -241.76828002929688,
|
| 1495 |
+
"loss": 0.6628,
|
| 1496 |
+
"rewards/accuracies": 1.0,
|
| 1497 |
+
"rewards/chosen": 0.03597152605652809,
|
| 1498 |
+
"rewards/margins": 0.06724558025598526,
|
| 1499 |
+
"rewards/margins_max": 0.06724558025598526,
|
| 1500 |
+
"rewards/margins_min": 0.06724558025598526,
|
| 1501 |
+
"rewards/margins_std": NaN,
|
| 1502 |
+
"rewards/rejected": -0.03127405419945717,
|
| 1503 |
+
"step": 820
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 0.61,
|
| 1507 |
+
"grad_norm": 0.310546875,
|
| 1508 |
+
"learning_rate": 1.9391971079868313e-07,
|
| 1509 |
+
"logits/chosen": -2.0435116291046143,
|
| 1510 |
+
"logits/rejected": -1.4584076404571533,
|
| 1511 |
+
"logps/chosen": -240.8225555419922,
|
| 1512 |
+
"logps/rejected": -303.6410827636719,
|
| 1513 |
+
"loss": 0.6637,
|
| 1514 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1515 |
+
"rewards/chosen": 0.029604297131299973,
|
| 1516 |
+
"rewards/margins": 0.06337974220514297,
|
| 1517 |
+
"rewards/margins_max": 0.06337974220514297,
|
| 1518 |
+
"rewards/margins_min": 0.06337974220514297,
|
| 1519 |
+
"rewards/margins_std": NaN,
|
| 1520 |
+
"rewards/rejected": -0.033775445073843,
|
| 1521 |
+
"step": 830
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.62,
|
| 1525 |
+
"grad_norm": 0.314453125,
|
| 1526 |
+
"learning_rate": 1.8763971398550467e-07,
|
| 1527 |
+
"logits/chosen": -2.0277044773101807,
|
| 1528 |
+
"logits/rejected": -1.4923416376113892,
|
| 1529 |
+
"logps/chosen": -250.7859344482422,
|
| 1530 |
+
"logps/rejected": -254.98202514648438,
|
| 1531 |
+
"loss": 0.6673,
|
| 1532 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1533 |
+
"rewards/chosen": 0.028646240010857582,
|
| 1534 |
+
"rewards/margins": 0.054024677723646164,
|
| 1535 |
+
"rewards/margins_max": 0.054024677723646164,
|
| 1536 |
+
"rewards/margins_min": 0.054024677723646164,
|
| 1537 |
+
"rewards/margins_std": NaN,
|
| 1538 |
+
"rewards/rejected": -0.025378435850143433,
|
| 1539 |
+
"step": 840
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 0.63,
|
| 1543 |
+
"grad_norm": 0.333984375,
|
| 1544 |
+
"learning_rate": 1.8140140709517465e-07,
|
| 1545 |
+
"logits/chosen": -2.047708034515381,
|
| 1546 |
+
"logits/rejected": -1.4980854988098145,
|
| 1547 |
+
"logps/chosen": -210.7559051513672,
|
| 1548 |
+
"logps/rejected": -207.94351196289062,
|
| 1549 |
+
"loss": 0.6643,
|
| 1550 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1551 |
+
"rewards/chosen": 0.01899396814405918,
|
| 1552 |
+
"rewards/margins": 0.048039522022008896,
|
| 1553 |
+
"rewards/margins_max": 0.048039522022008896,
|
| 1554 |
+
"rewards/margins_min": 0.048039522022008896,
|
| 1555 |
+
"rewards/margins_std": NaN,
|
| 1556 |
+
"rewards/rejected": -0.029045552015304565,
|
| 1557 |
+
"step": 850
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.64,
|
| 1561 |
+
"grad_norm": 0.353515625,
|
| 1562 |
+
"learning_rate": 1.7520896064308963e-07,
|
| 1563 |
+
"logits/chosen": -1.8993380069732666,
|
| 1564 |
+
"logits/rejected": -1.4513304233551025,
|
| 1565 |
+
"logps/chosen": -223.9241943359375,
|
| 1566 |
+
"logps/rejected": -297.637939453125,
|
| 1567 |
+
"loss": 0.6628,
|
| 1568 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1569 |
+
"rewards/chosen": 0.018290331587195396,
|
| 1570 |
+
"rewards/margins": 0.0594097264111042,
|
| 1571 |
+
"rewards/margins_max": 0.0594097264111042,
|
| 1572 |
+
"rewards/margins_min": 0.0594097264111042,
|
| 1573 |
+
"rewards/margins_std": NaN,
|
| 1574 |
+
"rewards/rejected": -0.04111938923597336,
|
| 1575 |
+
"step": 860
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 0.64,
|
| 1579 |
+
"grad_norm": 0.34765625,
|
| 1580 |
+
"learning_rate": 1.6906651448541976e-07,
|
| 1581 |
+
"logits/chosen": -2.0621097087860107,
|
| 1582 |
+
"logits/rejected": -1.237048625946045,
|
| 1583 |
+
"logps/chosen": -278.96832275390625,
|
| 1584 |
+
"logps/rejected": -267.96844482421875,
|
| 1585 |
+
"loss": 0.6608,
|
| 1586 |
+
"rewards/accuracies": 1.0,
|
| 1587 |
+
"rewards/chosen": 0.03597429022192955,
|
| 1588 |
+
"rewards/margins": 0.08543314039707184,
|
| 1589 |
+
"rewards/margins_max": 0.08543314039707184,
|
| 1590 |
+
"rewards/margins_min": 0.08543314039707184,
|
| 1591 |
+
"rewards/margins_std": NaN,
|
| 1592 |
+
"rewards/rejected": -0.04945885390043259,
|
| 1593 |
+
"step": 870
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.65,
|
| 1597 |
+
"grad_norm": 0.298828125,
|
| 1598 |
+
"learning_rate": 1.6297817505147739e-07,
|
| 1599 |
+
"logits/chosen": -1.8318344354629517,
|
| 1600 |
+
"logits/rejected": -1.3975905179977417,
|
| 1601 |
+
"logps/chosen": -238.3777618408203,
|
| 1602 |
+
"logps/rejected": -247.81350708007812,
|
| 1603 |
+
"loss": 0.6627,
|
| 1604 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1605 |
+
"rewards/chosen": 0.02713797055184841,
|
| 1606 |
+
"rewards/margins": 0.06845160573720932,
|
| 1607 |
+
"rewards/margins_max": 0.06845160573720932,
|
| 1608 |
+
"rewards/margins_min": 0.06845160573720932,
|
| 1609 |
+
"rewards/margins_std": NaN,
|
| 1610 |
+
"rewards/rejected": -0.04131363332271576,
|
| 1611 |
+
"step": 880
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 0.66,
|
| 1615 |
+
"grad_norm": 0.328125,
|
| 1616 |
+
"learning_rate": 1.5694801259843254e-07,
|
| 1617 |
+
"logits/chosen": -2.059532880783081,
|
| 1618 |
+
"logits/rejected": -1.5730485916137695,
|
| 1619 |
+
"logps/chosen": -219.37161254882812,
|
| 1620 |
+
"logps/rejected": -227.13394165039062,
|
| 1621 |
+
"loss": 0.6632,
|
| 1622 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1623 |
+
"rewards/chosen": 0.010322242975234985,
|
| 1624 |
+
"rewards/margins": 0.04914015904068947,
|
| 1625 |
+
"rewards/margins_max": 0.04914015904068947,
|
| 1626 |
+
"rewards/margins_min": 0.04914015904068947,
|
| 1627 |
+
"rewards/margins_std": NaN,
|
| 1628 |
+
"rewards/rejected": -0.03881791979074478,
|
| 1629 |
+
"step": 890
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.67,
|
| 1633 |
+
"grad_norm": 0.357421875,
|
| 1634 |
+
"learning_rate": 1.5098005849021078e-07,
|
| 1635 |
+
"logits/chosen": -2.163090467453003,
|
| 1636 |
+
"logits/rejected": -1.4286366701126099,
|
| 1637 |
+
"logps/chosen": -239.53402709960938,
|
| 1638 |
+
"logps/rejected": -268.863037109375,
|
| 1639 |
+
"loss": 0.6635,
|
| 1640 |
+
"rewards/accuracies": 1.0,
|
| 1641 |
+
"rewards/chosen": 0.042647577822208405,
|
| 1642 |
+
"rewards/margins": 0.0607173927128315,
|
| 1643 |
+
"rewards/margins_max": 0.0607173927128315,
|
| 1644 |
+
"rewards/margins_min": 0.0607173927128315,
|
| 1645 |
+
"rewards/margins_std": NaN,
|
| 1646 |
+
"rewards/rejected": -0.018069813027977943,
|
| 1647 |
+
"step": 900
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.67,
|
| 1651 |
+
"grad_norm": 0.4140625,
|
| 1652 |
+
"learning_rate": 1.4507830250239272e-07,
|
| 1653 |
+
"logits/chosen": -2.114851474761963,
|
| 1654 |
+
"logits/rejected": -1.3711202144622803,
|
| 1655 |
+
"logps/chosen": -214.02481079101562,
|
| 1656 |
+
"logps/rejected": -197.78500366210938,
|
| 1657 |
+
"loss": 0.665,
|
| 1658 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1659 |
+
"rewards/chosen": 0.03332282230257988,
|
| 1660 |
+
"rewards/margins": 0.057616226375103,
|
| 1661 |
+
"rewards/margins_max": 0.057616226375103,
|
| 1662 |
+
"rewards/margins_min": 0.057616226375103,
|
| 1663 |
+
"rewards/margins_std": NaN,
|
| 1664 |
+
"rewards/rejected": -0.02429339662194252,
|
| 1665 |
+
"step": 910
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 0.68,
|
| 1669 |
+
"grad_norm": 0.32421875,
|
| 1670 |
+
"learning_rate": 1.3924669015491628e-07,
|
| 1671 |
+
"logits/chosen": -1.891593337059021,
|
| 1672 |
+
"logits/rejected": -1.4508168697357178,
|
| 1673 |
+
"logps/chosen": -190.78921508789062,
|
| 1674 |
+
"logps/rejected": -210.9984588623047,
|
| 1675 |
+
"loss": 0.6632,
|
| 1676 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1677 |
+
"rewards/chosen": 0.037054695188999176,
|
| 1678 |
+
"rewards/margins": 0.06240003556013107,
|
| 1679 |
+
"rewards/margins_max": 0.06240003556013107,
|
| 1680 |
+
"rewards/margins_min": 0.06240003556013107,
|
| 1681 |
+
"rewards/margins_std": NaN,
|
| 1682 |
+
"rewards/rejected": -0.025345345959067345,
|
| 1683 |
+
"step": 920
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 0.69,
|
| 1687 |
+
"grad_norm": 0.322265625,
|
| 1688 |
+
"learning_rate": 1.3348912007436536e-07,
|
| 1689 |
+
"logits/chosen": -1.919196367263794,
|
| 1690 |
+
"logits/rejected": -1.4774754047393799,
|
| 1691 |
+
"logps/chosen": -243.7454376220703,
|
| 1692 |
+
"logps/rejected": -358.8896484375,
|
| 1693 |
+
"loss": 0.6585,
|
| 1694 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1695 |
+
"rewards/chosen": 0.021201713010668755,
|
| 1696 |
+
"rewards/margins": 0.0581810362637043,
|
| 1697 |
+
"rewards/margins_max": 0.0581810362637043,
|
| 1698 |
+
"rewards/margins_min": 0.0581810362637043,
|
| 1699 |
+
"rewards/margins_std": NaN,
|
| 1700 |
+
"rewards/rejected": -0.036979321390390396,
|
| 1701 |
+
"step": 930
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 0.7,
|
| 1705 |
+
"grad_norm": 0.296875,
|
| 1706 |
+
"learning_rate": 1.2780944138760896e-07,
|
| 1707 |
+
"logits/chosen": -2.0747437477111816,
|
| 1708 |
+
"logits/rejected": -1.4782272577285767,
|
| 1709 |
+
"logps/chosen": -193.60910034179688,
|
| 1710 |
+
"logps/rejected": -173.97982788085938,
|
| 1711 |
+
"loss": 0.6677,
|
| 1712 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1713 |
+
"rewards/chosen": 0.020243193954229355,
|
| 1714 |
+
"rewards/margins": 0.04255754500627518,
|
| 1715 |
+
"rewards/margins_max": 0.04255754500627518,
|
| 1716 |
+
"rewards/margins_min": 0.04255754500627518,
|
| 1717 |
+
"rewards/margins_std": NaN,
|
| 1718 |
+
"rewards/rejected": -0.022314347326755524,
|
| 1719 |
+
"step": 940
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 0.7,
|
| 1723 |
+
"grad_norm": 0.375,
|
| 1724 |
+
"learning_rate": 1.222114511485317e-07,
|
| 1725 |
+
"logits/chosen": -2.0575144290924072,
|
| 1726 |
+
"logits/rejected": -1.5969568490982056,
|
| 1727 |
+
"logps/chosen": -213.1886749267578,
|
| 1728 |
+
"logps/rejected": -241.73458862304688,
|
| 1729 |
+
"loss": 0.6621,
|
| 1730 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1731 |
+
"rewards/chosen": 0.037582967430353165,
|
| 1732 |
+
"rewards/margins": 0.061435867100954056,
|
| 1733 |
+
"rewards/margins_max": 0.061435867100954056,
|
| 1734 |
+
"rewards/margins_min": 0.061435867100954056,
|
| 1735 |
+
"rewards/margins_std": NaN,
|
| 1736 |
+
"rewards/rejected": -0.02385290339589119,
|
| 1737 |
+
"step": 950
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 0.71,
|
| 1741 |
+
"grad_norm": 0.32421875,
|
| 1742 |
+
"learning_rate": 1.1669889179957723e-07,
|
| 1743 |
+
"logits/chosen": -1.9735885858535767,
|
| 1744 |
+
"logits/rejected": -1.662937879562378,
|
| 1745 |
+
"logps/chosen": -208.2716064453125,
|
| 1746 |
+
"logps/rejected": -174.0637664794922,
|
| 1747 |
+
"loss": 0.6654,
|
| 1748 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1749 |
+
"rewards/chosen": 0.009742580354213715,
|
| 1750 |
+
"rewards/margins": 0.03067430481314659,
|
| 1751 |
+
"rewards/margins_max": 0.03067430481314659,
|
| 1752 |
+
"rewards/margins_min": 0.03067430481314659,
|
| 1753 |
+
"rewards/margins_std": NaN,
|
| 1754 |
+
"rewards/rejected": -0.020931728184223175,
|
| 1755 |
+
"step": 960
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.72,
|
| 1759 |
+
"grad_norm": 0.318359375,
|
| 1760 |
+
"learning_rate": 1.112754486698008e-07,
|
| 1761 |
+
"logits/chosen": -2.019521713256836,
|
| 1762 |
+
"logits/rejected": -1.5791001319885254,
|
| 1763 |
+
"logps/chosen": -187.46524047851562,
|
| 1764 |
+
"logps/rejected": -178.125732421875,
|
| 1765 |
+
"loss": 0.6615,
|
| 1766 |
+
"rewards/accuracies": 1.0,
|
| 1767 |
+
"rewards/chosen": 0.01672300137579441,
|
| 1768 |
+
"rewards/margins": 0.04681010916829109,
|
| 1769 |
+
"rewards/margins_max": 0.04681010916829109,
|
| 1770 |
+
"rewards/margins_min": 0.04681010916829109,
|
| 1771 |
+
"rewards/margins_std": NaN,
|
| 1772 |
+
"rewards/rejected": -0.03008710779249668,
|
| 1773 |
+
"step": 970
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.73,
|
| 1777 |
+
"grad_norm": 0.373046875,
|
| 1778 |
+
"learning_rate": 1.059447475111049e-07,
|
| 1779 |
+
"logits/chosen": -1.9721561670303345,
|
| 1780 |
+
"logits/rejected": -1.4720313549041748,
|
| 1781 |
+
"logps/chosen": -247.0120849609375,
|
| 1782 |
+
"logps/rejected": -213.7181396484375,
|
| 1783 |
+
"loss": 0.6647,
|
| 1784 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1785 |
+
"rewards/chosen": 0.035904593765735626,
|
| 1786 |
+
"rewards/margins": 0.05502895638346672,
|
| 1787 |
+
"rewards/margins_max": 0.05502895638346672,
|
| 1788 |
+
"rewards/margins_min": 0.05502895638346672,
|
| 1789 |
+
"rewards/margins_std": NaN,
|
| 1790 |
+
"rewards/rejected": -0.019124364480376244,
|
| 1791 |
+
"step": 980
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 0.73,
|
| 1795 |
+
"grad_norm": 0.326171875,
|
| 1796 |
+
"learning_rate": 1.007103520743035e-07,
|
| 1797 |
+
"logits/chosen": -2.0138068199157715,
|
| 1798 |
+
"logits/rejected": -1.4093111753463745,
|
| 1799 |
+
"logps/chosen": -254.36618041992188,
|
| 1800 |
+
"logps/rejected": -262.58319091796875,
|
| 1801 |
+
"loss": 0.6628,
|
| 1802 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1803 |
+
"rewards/chosen": 0.03567090258002281,
|
| 1804 |
+
"rewards/margins": 0.05842595174908638,
|
| 1805 |
+
"rewards/margins_max": 0.05842595174908638,
|
| 1806 |
+
"rewards/margins_min": 0.05842595174908638,
|
| 1807 |
+
"rewards/margins_std": NaN,
|
| 1808 |
+
"rewards/rejected": -0.022755051031708717,
|
| 1809 |
+
"step": 990
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 0.74,
|
| 1813 |
+
"grad_norm": 0.30078125,
|
| 1814 |
+
"learning_rate": 9.557576172663575e-08,
|
| 1815 |
+
"logits/chosen": -1.9736032485961914,
|
| 1816 |
+
"logits/rejected": -1.531928300857544,
|
| 1817 |
+
"logps/chosen": -213.33682250976562,
|
| 1818 |
+
"logps/rejected": -193.63340759277344,
|
| 1819 |
+
"loss": 0.6648,
|
| 1820 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1821 |
+
"rewards/chosen": 0.012441306374967098,
|
| 1822 |
+
"rewards/margins": 0.04004935175180435,
|
| 1823 |
+
"rewards/margins_max": 0.04004935175180435,
|
| 1824 |
+
"rewards/margins_min": 0.04004935175180435,
|
| 1825 |
+
"rewards/margins_std": NaN,
|
| 1826 |
+
"rewards/rejected": -0.027608046308159828,
|
| 1827 |
+
"step": 1000
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 0.75,
|
| 1831 |
+
"grad_norm": 0.34765625,
|
| 1832 |
+
"learning_rate": 9.054440911232347e-08,
|
| 1833 |
+
"logits/chosen": -1.9753551483154297,
|
| 1834 |
+
"logits/rejected": -1.4374115467071533,
|
| 1835 |
+
"logps/chosen": -257.1168518066406,
|
| 1836 |
+
"logps/rejected": -271.955078125,
|
| 1837 |
+
"loss": 0.6641,
|
| 1838 |
+
"rewards/accuracies": 1.0,
|
| 1839 |
+
"rewards/chosen": 0.018898367881774902,
|
| 1840 |
+
"rewards/margins": 0.054756682366132736,
|
| 1841 |
+
"rewards/margins_max": 0.054756682366132736,
|
| 1842 |
+
"rewards/margins_min": 0.054756682366132736,
|
| 1843 |
+
"rewards/margins_std": NaN,
|
| 1844 |
+
"rewards/rejected": -0.035858314484357834,
|
| 1845 |
+
"step": 1010
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.76,
|
| 1849 |
+
"grad_norm": 0.30078125,
|
| 1850 |
+
"learning_rate": 8.561965785773412e-08,
|
| 1851 |
+
"logits/chosen": -2.133695125579834,
|
| 1852 |
+
"logits/rejected": -1.565525770187378,
|
| 1853 |
+
"logps/chosen": -212.60366821289062,
|
| 1854 |
+
"logps/rejected": -182.28314208984375,
|
| 1855 |
+
"loss": 0.6679,
|
| 1856 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1857 |
+
"rewards/chosen": 0.030873173847794533,
|
| 1858 |
+
"rewards/margins": 0.05696907639503479,
|
| 1859 |
+
"rewards/margins_max": 0.05696907639503479,
|
| 1860 |
+
"rewards/margins_min": 0.05696907639503479,
|
| 1861 |
+
"rewards/margins_std": NaN,
|
| 1862 |
+
"rewards/rejected": -0.026095900684595108,
|
| 1863 |
+
"step": 1020
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 0.76,
|
| 1867 |
+
"grad_norm": 0.337890625,
|
| 1868 |
+
"learning_rate": 8.080480032268452e-08,
|
| 1869 |
+
"logits/chosen": -1.9859001636505127,
|
| 1870 |
+
"logits/rejected": -1.4256551265716553,
|
| 1871 |
+
"logps/chosen": -243.8011474609375,
|
| 1872 |
+
"logps/rejected": -223.94677734375,
|
| 1873 |
+
"loss": 0.6639,
|
| 1874 |
+
"rewards/accuracies": 1.0,
|
| 1875 |
+
"rewards/chosen": 0.02358115091919899,
|
| 1876 |
+
"rewards/margins": 0.051641564816236496,
|
| 1877 |
+
"rewards/margins_max": 0.051641564816236496,
|
| 1878 |
+
"rewards/margins_min": 0.051641564816236496,
|
| 1879 |
+
"rewards/margins_std": NaN,
|
| 1880 |
+
"rewards/rejected": -0.028060410171747208,
|
| 1881 |
+
"step": 1030
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.77,
|
| 1885 |
+
"grad_norm": 0.322265625,
|
| 1886 |
+
"learning_rate": 7.61030553993895e-08,
|
| 1887 |
+
"logits/chosen": -1.9076287746429443,
|
| 1888 |
+
"logits/rejected": -1.4924620389938354,
|
| 1889 |
+
"logps/chosen": -216.83840942382812,
|
| 1890 |
+
"logps/rejected": -196.377685546875,
|
| 1891 |
+
"loss": 0.6633,
|
| 1892 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1893 |
+
"rewards/chosen": 0.013671179302036762,
|
| 1894 |
+
"rewards/margins": 0.049784958362579346,
|
| 1895 |
+
"rewards/margins_max": 0.049784958362579346,
|
| 1896 |
+
"rewards/margins_min": 0.049784958362579346,
|
| 1897 |
+
"rewards/margins_std": NaN,
|
| 1898 |
+
"rewards/rejected": -0.03611377626657486,
|
| 1899 |
+
"step": 1040
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.78,
|
| 1903 |
+
"grad_norm": 0.314453125,
|
| 1904 |
+
"learning_rate": 7.151756636052527e-08,
|
| 1905 |
+
"logits/chosen": -2.130204916000366,
|
| 1906 |
+
"logits/rejected": -1.5611693859100342,
|
| 1907 |
+
"logps/chosen": -222.9783477783203,
|
| 1908 |
+
"logps/rejected": -241.9842987060547,
|
| 1909 |
+
"loss": 0.6591,
|
| 1910 |
+
"rewards/accuracies": 1.0,
|
| 1911 |
+
"rewards/chosen": 0.03606210649013519,
|
| 1912 |
+
"rewards/margins": 0.059976816177368164,
|
| 1913 |
+
"rewards/margins_max": 0.059976816177368164,
|
| 1914 |
+
"rewards/margins_min": 0.059976816177368164,
|
| 1915 |
+
"rewards/margins_std": NaN,
|
| 1916 |
+
"rewards/rejected": -0.02391471341252327,
|
| 1917 |
+
"step": 1050
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 0.78,
|
| 1921 |
+
"grad_norm": 0.33984375,
|
| 1922 |
+
"learning_rate": 6.705139875784699e-08,
|
| 1923 |
+
"logits/chosen": -2.0021703243255615,
|
| 1924 |
+
"logits/rejected": -1.3691127300262451,
|
| 1925 |
+
"logps/chosen": -229.1208038330078,
|
| 1926 |
+
"logps/rejected": -196.5934295654297,
|
| 1927 |
+
"loss": 0.664,
|
| 1928 |
+
"rewards/accuracies": 1.0,
|
| 1929 |
+
"rewards/chosen": 0.016182031482458115,
|
| 1930 |
+
"rewards/margins": 0.04835697263479233,
|
| 1931 |
+
"rewards/margins_max": 0.04835697263479233,
|
| 1932 |
+
"rewards/margins_min": 0.04835697263479233,
|
| 1933 |
+
"rewards/margins_std": NaN,
|
| 1934 |
+
"rewards/rejected": -0.03217494487762451,
|
| 1935 |
+
"step": 1060
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 0.79,
|
| 1939 |
+
"grad_norm": 0.357421875,
|
| 1940 |
+
"learning_rate": 6.270753837276599e-08,
|
| 1941 |
+
"logits/chosen": -1.9585247039794922,
|
| 1942 |
+
"logits/rejected": -1.6326916217803955,
|
| 1943 |
+
"logps/chosen": -208.2965850830078,
|
| 1944 |
+
"logps/rejected": -220.41714477539062,
|
| 1945 |
+
"loss": 0.6634,
|
| 1946 |
+
"rewards/accuracies": 1.0,
|
| 1947 |
+
"rewards/chosen": 0.02728954516351223,
|
| 1948 |
+
"rewards/margins": 0.05682883784174919,
|
| 1949 |
+
"rewards/margins_max": 0.05682883784174919,
|
| 1950 |
+
"rewards/margins_min": 0.05682883784174919,
|
| 1951 |
+
"rewards/margins_std": NaN,
|
| 1952 |
+
"rewards/rejected": -0.02953929826617241,
|
| 1953 |
+
"step": 1070
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 0.8,
|
| 1957 |
+
"grad_norm": 0.322265625,
|
| 1958 |
+
"learning_rate": 5.848888922025552e-08,
|
| 1959 |
+
"logits/chosen": -2.0108110904693604,
|
| 1960 |
+
"logits/rejected": -1.5322282314300537,
|
| 1961 |
+
"logps/chosen": -193.75543212890625,
|
| 1962 |
+
"logps/rejected": -214.4615020751953,
|
| 1963 |
+
"loss": 0.6642,
|
| 1964 |
+
"rewards/accuracies": 1.0,
|
| 1965 |
+
"rewards/chosen": 0.02274712175130844,
|
| 1966 |
+
"rewards/margins": 0.062083650380373,
|
| 1967 |
+
"rewards/margins_max": 0.062083650380373,
|
| 1968 |
+
"rewards/margins_min": 0.062083650380373,
|
| 1969 |
+
"rewards/margins_std": NaN,
|
| 1970 |
+
"rewards/rejected": -0.03933652862906456,
|
| 1971 |
+
"step": 1080
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.81,
|
| 1975 |
+
"grad_norm": 0.3125,
|
| 1976 |
+
"learning_rate": 5.439827160742006e-08,
|
| 1977 |
+
"logits/chosen": -1.9512097835540771,
|
| 1978 |
+
"logits/rejected": -1.6602758169174194,
|
| 1979 |
+
"logps/chosen": -189.49319458007812,
|
| 1980 |
+
"logps/rejected": -244.29592895507812,
|
| 1981 |
+
"loss": 0.6643,
|
| 1982 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1983 |
+
"rewards/chosen": 0.02437346801161766,
|
| 1984 |
+
"rewards/margins": 0.0436645932495594,
|
| 1985 |
+
"rewards/margins_max": 0.0436645932495594,
|
| 1986 |
+
"rewards/margins_min": 0.0436645932495594,
|
| 1987 |
+
"rewards/margins_std": NaN,
|
| 1988 |
+
"rewards/rejected": -0.019291121512651443,
|
| 1989 |
+
"step": 1090
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 0.81,
|
| 1993 |
+
"grad_norm": 0.35546875,
|
| 1994 |
+
"learning_rate": 5.0438420248026745e-08,
|
| 1995 |
+
"logits/chosen": -1.8881117105484009,
|
| 1996 |
+
"logits/rejected": -1.3894041776657104,
|
| 1997 |
+
"logps/chosen": -198.81190490722656,
|
| 1998 |
+
"logps/rejected": -250.2027587890625,
|
| 1999 |
+
"loss": 0.6616,
|
| 2000 |
+
"rewards/accuracies": 1.0,
|
| 2001 |
+
"rewards/chosen": 0.03594060614705086,
|
| 2002 |
+
"rewards/margins": 0.07942356169223785,
|
| 2003 |
+
"rewards/margins_max": 0.07942356169223785,
|
| 2004 |
+
"rewards/margins_min": 0.07942356169223785,
|
| 2005 |
+
"rewards/margins_std": NaN,
|
| 2006 |
+
"rewards/rejected": -0.0434829518198967,
|
| 2007 |
+
"step": 1100
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.82,
|
| 2011 |
+
"grad_norm": 0.34375,
|
| 2012 |
+
"learning_rate": 4.6611982434258124e-08,
|
| 2013 |
+
"logits/chosen": -1.934699296951294,
|
| 2014 |
+
"logits/rejected": -1.1928954124450684,
|
| 2015 |
+
"logps/chosen": -261.38677978515625,
|
| 2016 |
+
"logps/rejected": -241.5679473876953,
|
| 2017 |
+
"loss": 0.6629,
|
| 2018 |
+
"rewards/accuracies": 1.0,
|
| 2019 |
+
"rewards/chosen": 0.024552561342716217,
|
| 2020 |
+
"rewards/margins": 0.060813069343566895,
|
| 2021 |
+
"rewards/margins_max": 0.060813069343566895,
|
| 2022 |
+
"rewards/margins_min": 0.060813069343566895,
|
| 2023 |
+
"rewards/margins_std": NaN,
|
| 2024 |
+
"rewards/rejected": -0.03626050800085068,
|
| 2025 |
+
"step": 1110
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.83,
|
| 2029 |
+
"grad_norm": 0.29296875,
|
| 2030 |
+
"learning_rate": 4.292151626690924e-08,
|
| 2031 |
+
"logits/chosen": -2.027463674545288,
|
| 2032 |
+
"logits/rejected": -1.3185806274414062,
|
| 2033 |
+
"logps/chosen": -252.04483032226562,
|
| 2034 |
+
"logps/rejected": -265.124267578125,
|
| 2035 |
+
"loss": 0.667,
|
| 2036 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2037 |
+
"rewards/chosen": 0.03300962597131729,
|
| 2038 |
+
"rewards/margins": 0.0655718669295311,
|
| 2039 |
+
"rewards/margins_max": 0.0655718669295311,
|
| 2040 |
+
"rewards/margins_min": 0.0655718669295311,
|
| 2041 |
+
"rewards/margins_std": NaN,
|
| 2042 |
+
"rewards/rejected": -0.032562240958213806,
|
| 2043 |
+
"step": 1120
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 0.84,
|
| 2047 |
+
"grad_norm": 0.330078125,
|
| 2048 |
+
"learning_rate": 3.9369488945212606e-08,
|
| 2049 |
+
"logits/chosen": -2.0518107414245605,
|
| 2050 |
+
"logits/rejected": -1.5059715509414673,
|
| 2051 |
+
"logps/chosen": -211.740478515625,
|
| 2052 |
+
"logps/rejected": -182.64515686035156,
|
| 2053 |
+
"loss": 0.6653,
|
| 2054 |
+
"rewards/accuracies": 1.0,
|
| 2055 |
+
"rewards/chosen": 0.016033833846449852,
|
| 2056 |
+
"rewards/margins": 0.05374104902148247,
|
| 2057 |
+
"rewards/margins_max": 0.05374104902148247,
|
| 2058 |
+
"rewards/margins_min": 0.05374104902148247,
|
| 2059 |
+
"rewards/margins_std": NaN,
|
| 2060 |
+
"rewards/rejected": -0.037707217037677765,
|
| 2061 |
+
"step": 1130
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 0.84,
|
| 2065 |
+
"grad_norm": 0.34375,
|
| 2066 |
+
"learning_rate": 3.5958275117433404e-08,
|
| 2067 |
+
"logits/chosen": -2.0312271118164062,
|
| 2068 |
+
"logits/rejected": -1.4592535495758057,
|
| 2069 |
+
"logps/chosen": -251.87002563476562,
|
| 2070 |
+
"logps/rejected": -245.94912719726562,
|
| 2071 |
+
"loss": 0.6674,
|
| 2072 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2073 |
+
"rewards/chosen": 0.025403957813978195,
|
| 2074 |
+
"rewards/margins": 0.058066435158252716,
|
| 2075 |
+
"rewards/margins_max": 0.058066435158252716,
|
| 2076 |
+
"rewards/margins_min": 0.058066435158252716,
|
| 2077 |
+
"rewards/margins_std": NaN,
|
| 2078 |
+
"rewards/rejected": -0.03266247734427452,
|
| 2079 |
+
"step": 1140
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 0.85,
|
| 2083 |
+
"grad_norm": 0.361328125,
|
| 2084 |
+
"learning_rate": 3.269015529333805e-08,
|
| 2085 |
+
"logits/chosen": -2.043374538421631,
|
| 2086 |
+
"logits/rejected": -1.4708629846572876,
|
| 2087 |
+
"logps/chosen": -234.3373565673828,
|
| 2088 |
+
"logps/rejected": -253.6455841064453,
|
| 2089 |
+
"loss": 0.6632,
|
| 2090 |
+
"rewards/accuracies": 1.0,
|
| 2091 |
+
"rewards/chosen": 0.04083864018321037,
|
| 2092 |
+
"rewards/margins": 0.06850458681583405,
|
| 2093 |
+
"rewards/margins_max": 0.06850458681583405,
|
| 2094 |
+
"rewards/margins_min": 0.06850458681583405,
|
| 2095 |
+
"rewards/margins_std": NaN,
|
| 2096 |
+
"rewards/rejected": -0.027665939182043076,
|
| 2097 |
+
"step": 1150
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.86,
|
| 2101 |
+
"grad_norm": 0.4296875,
|
| 2102 |
+
"learning_rate": 2.9567314319598e-08,
|
| 2103 |
+
"logits/chosen": -1.8975934982299805,
|
| 2104 |
+
"logits/rejected": -1.23898184299469,
|
| 2105 |
+
"logps/chosen": -194.3213348388672,
|
| 2106 |
+
"logps/rejected": -143.7144317626953,
|
| 2107 |
+
"loss": 0.6637,
|
| 2108 |
+
"rewards/accuracies": 1.0,
|
| 2109 |
+
"rewards/chosen": 0.029597630724310875,
|
| 2110 |
+
"rewards/margins": 0.058405883610248566,
|
| 2111 |
+
"rewards/margins_max": 0.058405883610248566,
|
| 2112 |
+
"rewards/margins_min": 0.058405883610248566,
|
| 2113 |
+
"rewards/margins_std": NaN,
|
| 2114 |
+
"rewards/rejected": -0.02880825474858284,
|
| 2115 |
+
"step": 1160
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 0.87,
|
| 2119 |
+
"grad_norm": 0.30078125,
|
| 2120 |
+
"learning_rate": 2.659183991914696e-08,
|
| 2121 |
+
"logits/chosen": -2.1204023361206055,
|
| 2122 |
+
"logits/rejected": -1.3706982135772705,
|
| 2123 |
+
"logps/chosen": -244.99868774414062,
|
| 2124 |
+
"logps/rejected": -214.16195678710938,
|
| 2125 |
+
"loss": 0.6647,
|
| 2126 |
+
"rewards/accuracies": 1.0,
|
| 2127 |
+
"rewards/chosen": 0.02851070463657379,
|
| 2128 |
+
"rewards/margins": 0.07035355269908905,
|
| 2129 |
+
"rewards/margins_max": 0.07035355269908905,
|
| 2130 |
+
"rewards/margins_min": 0.07035355269908905,
|
| 2131 |
+
"rewards/margins_std": NaN,
|
| 2132 |
+
"rewards/rejected": -0.04184284433722496,
|
| 2133 |
+
"step": 1170
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.87,
|
| 2137 |
+
"grad_norm": 0.310546875,
|
| 2138 |
+
"learning_rate": 2.3765721295468666e-08,
|
| 2139 |
+
"logits/chosen": -1.8714052438735962,
|
| 2140 |
+
"logits/rejected": -1.6353321075439453,
|
| 2141 |
+
"logps/chosen": -162.21560668945312,
|
| 2142 |
+
"logps/rejected": -206.0563507080078,
|
| 2143 |
+
"loss": 0.6639,
|
| 2144 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2145 |
+
"rewards/chosen": 0.027320072054862976,
|
| 2146 |
+
"rewards/margins": 0.04566752910614014,
|
| 2147 |
+
"rewards/margins_max": 0.04566752910614014,
|
| 2148 |
+
"rewards/margins_min": 0.04566752910614014,
|
| 2149 |
+
"rewards/margins_std": NaN,
|
| 2150 |
+
"rewards/rejected": -0.01834745705127716,
|
| 2151 |
+
"step": 1180
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.88,
|
| 2155 |
+
"grad_norm": 0.35546875,
|
| 2156 |
+
"learning_rate": 2.109084780274853e-08,
|
| 2157 |
+
"logits/chosen": -1.9949709177017212,
|
| 2158 |
+
"logits/rejected": -1.4209120273590088,
|
| 2159 |
+
"logps/chosen": -259.18963623046875,
|
| 2160 |
+
"logps/rejected": -235.44589233398438,
|
| 2161 |
+
"loss": 0.6632,
|
| 2162 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2163 |
+
"rewards/chosen": 0.023489583283662796,
|
| 2164 |
+
"rewards/margins": 0.07520689070224762,
|
| 2165 |
+
"rewards/margins_max": 0.07520689070224762,
|
| 2166 |
+
"rewards/margins_min": 0.07520689070224762,
|
| 2167 |
+
"rewards/margins_std": NaN,
|
| 2168 |
+
"rewards/rejected": -0.05171731114387512,
|
| 2169 |
+
"step": 1190
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 0.89,
|
| 2173 |
+
"grad_norm": 0.357421875,
|
| 2174 |
+
"learning_rate": 1.8569007682777415e-08,
|
| 2175 |
+
"logits/chosen": -2.0659101009368896,
|
| 2176 |
+
"logits/rejected": -1.2472288608551025,
|
| 2177 |
+
"logps/chosen": -237.00924682617188,
|
| 2178 |
+
"logps/rejected": -287.98492431640625,
|
| 2179 |
+
"loss": 0.6659,
|
| 2180 |
+
"rewards/accuracies": 1.0,
|
| 2181 |
+
"rewards/chosen": 0.032205354422330856,
|
| 2182 |
+
"rewards/margins": 0.06721274554729462,
|
| 2183 |
+
"rewards/margins_max": 0.06721274554729462,
|
| 2184 |
+
"rewards/margins_min": 0.06721274554729462,
|
| 2185 |
+
"rewards/margins_std": NaN,
|
| 2186 |
+
"rewards/rejected": -0.03500739112496376,
|
| 2187 |
+
"step": 1200
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 0.9,
|
| 2191 |
+
"grad_norm": 0.32421875,
|
| 2192 |
+
"learning_rate": 1.620188686945259e-08,
|
| 2193 |
+
"logits/chosen": -2.077946186065674,
|
| 2194 |
+
"logits/rejected": -1.3815569877624512,
|
| 2195 |
+
"logps/chosen": -299.30889892578125,
|
| 2196 |
+
"logps/rejected": -249.3920135498047,
|
| 2197 |
+
"loss": 0.6653,
|
| 2198 |
+
"rewards/accuracies": 1.0,
|
| 2199 |
+
"rewards/chosen": 0.03304355591535568,
|
| 2200 |
+
"rewards/margins": 0.05800234526395798,
|
| 2201 |
+
"rewards/margins_max": 0.05800234526395798,
|
| 2202 |
+
"rewards/margins_min": 0.05800234526395798,
|
| 2203 |
+
"rewards/margins_std": NaN,
|
| 2204 |
+
"rewards/rejected": -0.024958789348602295,
|
| 2205 |
+
"step": 1210
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 0.9,
|
| 2209 |
+
"grad_norm": 0.3125,
|
| 2210 |
+
"learning_rate": 1.3991067861675004e-08,
|
| 2211 |
+
"logits/chosen": -2.1188392639160156,
|
| 2212 |
+
"logits/rejected": -1.4769757986068726,
|
| 2213 |
+
"logps/chosen": -290.83831787109375,
|
| 2214 |
+
"logps/rejected": -317.05499267578125,
|
| 2215 |
+
"loss": 0.6598,
|
| 2216 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2217 |
+
"rewards/chosen": 0.02561381086707115,
|
| 2218 |
+
"rewards/margins": 0.05808825045824051,
|
| 2219 |
+
"rewards/margins_max": 0.05808825045824051,
|
| 2220 |
+
"rewards/margins_min": 0.05808825045824051,
|
| 2221 |
+
"rewards/margins_std": NaN,
|
| 2222 |
+
"rewards/rejected": -0.03247443586587906,
|
| 2223 |
+
"step": 1220
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.91,
|
| 2227 |
+
"grad_norm": 0.302734375,
|
| 2228 |
+
"learning_rate": 1.1938028665396171e-08,
|
| 2229 |
+
"logits/chosen": -1.9317171573638916,
|
| 2230 |
+
"logits/rejected": -1.4349186420440674,
|
| 2231 |
+
"logps/chosen": -260.46673583984375,
|
| 2232 |
+
"logps/rejected": -225.98544311523438,
|
| 2233 |
+
"loss": 0.6616,
|
| 2234 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2235 |
+
"rewards/chosen": 0.025805342942476273,
|
| 2236 |
+
"rewards/margins": 0.05584897845983505,
|
| 2237 |
+
"rewards/margins_max": 0.05584897845983505,
|
| 2238 |
+
"rewards/margins_min": 0.05584897845983505,
|
| 2239 |
+
"rewards/margins_std": NaN,
|
| 2240 |
+
"rewards/rejected": -0.03004363179206848,
|
| 2241 |
+
"step": 1230
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 0.92,
|
| 2245 |
+
"grad_norm": 0.291015625,
|
| 2246 |
+
"learning_rate": 1.0044141805522016e-08,
|
| 2247 |
+
"logits/chosen": -1.959181785583496,
|
| 2248 |
+
"logits/rejected": -1.532305359840393,
|
| 2249 |
+
"logps/chosen": -225.9596710205078,
|
| 2250 |
+
"logps/rejected": -205.9132537841797,
|
| 2251 |
+
"loss": 0.6615,
|
| 2252 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2253 |
+
"rewards/chosen": 0.021490924060344696,
|
| 2254 |
+
"rewards/margins": 0.04717742279171944,
|
| 2255 |
+
"rewards/margins_max": 0.04717742279171944,
|
| 2256 |
+
"rewards/margins_min": 0.04717742279171944,
|
| 2257 |
+
"rewards/margins_std": NaN,
|
| 2258 |
+
"rewards/rejected": -0.02568650245666504,
|
| 2259 |
+
"step": 1240
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 0.93,
|
| 2263 |
+
"grad_norm": 0.30078125,
|
| 2264 |
+
"learning_rate": 8.310673408334496e-09,
|
| 2265 |
+
"logits/chosen": -2.0635626316070557,
|
| 2266 |
+
"logits/rejected": -1.3920619487762451,
|
| 2267 |
+
"logps/chosen": -252.69125366210938,
|
| 2268 |
+
"logps/rejected": -257.30242919921875,
|
| 2269 |
+
"loss": 0.6648,
|
| 2270 |
+
"rewards/accuracies": 1.0,
|
| 2271 |
+
"rewards/chosen": 0.036212801933288574,
|
| 2272 |
+
"rewards/margins": 0.061691027134656906,
|
| 2273 |
+
"rewards/margins_max": 0.061691027134656906,
|
| 2274 |
+
"rewards/margins_min": 0.061691027134656906,
|
| 2275 |
+
"rewards/margins_std": NaN,
|
| 2276 |
+
"rewards/rejected": -0.02547823265194893,
|
| 2277 |
+
"step": 1250
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.93,
|
| 2281 |
+
"grad_norm": 0.373046875,
|
| 2282 |
+
"learning_rate": 6.738782355044048e-09,
|
| 2283 |
+
"logits/chosen": -2.0646793842315674,
|
| 2284 |
+
"logits/rejected": -1.4474830627441406,
|
| 2285 |
+
"logps/chosen": -206.7698516845703,
|
| 2286 |
+
"logps/rejected": -193.35592651367188,
|
| 2287 |
+
"loss": 0.6606,
|
| 2288 |
+
"rewards/accuracies": 1.0,
|
| 2289 |
+
"rewards/chosen": 0.023445401340723038,
|
| 2290 |
+
"rewards/margins": 0.059154510498046875,
|
| 2291 |
+
"rewards/margins_max": 0.059154510498046875,
|
| 2292 |
+
"rewards/margins_min": 0.059154510498046875,
|
| 2293 |
+
"rewards/margins_std": NaN,
|
| 2294 |
+
"rewards/rejected": -0.03570910915732384,
|
| 2295 |
+
"step": 1260
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 0.94,
|
| 2299 |
+
"grad_norm": 0.349609375,
|
| 2300 |
+
"learning_rate": 5.329519507038971e-09,
|
| 2301 |
+
"logits/chosen": -1.9023844003677368,
|
| 2302 |
+
"logits/rejected": -1.439485788345337,
|
| 2303 |
+
"logps/chosen": -172.00807189941406,
|
| 2304 |
+
"logps/rejected": -252.7501983642578,
|
| 2305 |
+
"loss": 0.6673,
|
| 2306 |
+
"rewards/accuracies": 1.0,
|
| 2307 |
+
"rewards/chosen": 0.027050817385315895,
|
| 2308 |
+
"rewards/margins": 0.04781107231974602,
|
| 2309 |
+
"rewards/margins_max": 0.04781107231974602,
|
| 2310 |
+
"rewards/margins_min": 0.04781107231974602,
|
| 2311 |
+
"rewards/margins_std": NaN,
|
| 2312 |
+
"rewards/rejected": -0.02076025679707527,
|
| 2313 |
+
"step": 1270
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.95,
|
| 2317 |
+
"grad_norm": 0.322265625,
|
| 2318 |
+
"learning_rate": 4.0838270033497215e-09,
|
| 2319 |
+
"logits/chosen": -1.969686508178711,
|
| 2320 |
+
"logits/rejected": -1.2621902227401733,
|
| 2321 |
+
"logps/chosen": -262.63665771484375,
|
| 2322 |
+
"logps/rejected": -258.04046630859375,
|
| 2323 |
+
"loss": 0.6626,
|
| 2324 |
+
"rewards/accuracies": 1.0,
|
| 2325 |
+
"rewards/chosen": 0.03954679146409035,
|
| 2326 |
+
"rewards/margins": 0.06421086192131042,
|
| 2327 |
+
"rewards/margins_max": 0.06421086192131042,
|
| 2328 |
+
"rewards/margins_min": 0.06421086192131042,
|
| 2329 |
+
"rewards/margins_std": NaN,
|
| 2330 |
+
"rewards/rejected": -0.024664074182510376,
|
| 2331 |
+
"step": 1280
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 0.96,
|
| 2335 |
+
"grad_norm": 0.28515625,
|
| 2336 |
+
"learning_rate": 3.002537630797747e-09,
|
| 2337 |
+
"logits/chosen": -2.1349868774414062,
|
| 2338 |
+
"logits/rejected": -1.539178729057312,
|
| 2339 |
+
"logps/chosen": -239.86355590820312,
|
| 2340 |
+
"logps/rejected": -216.4762420654297,
|
| 2341 |
+
"loss": 0.6654,
|
| 2342 |
+
"rewards/accuracies": 1.0,
|
| 2343 |
+
"rewards/chosen": 0.02870868518948555,
|
| 2344 |
+
"rewards/margins": 0.05009246617555618,
|
| 2345 |
+
"rewards/margins_max": 0.05009246617555618,
|
| 2346 |
+
"rewards/margins_min": 0.05009246617555618,
|
| 2347 |
+
"rewards/margins_std": NaN,
|
| 2348 |
+
"rewards/rejected": -0.021383779123425484,
|
| 2349 |
+
"step": 1290
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 0.96,
|
| 2353 |
+
"grad_norm": 0.2890625,
|
| 2354 |
+
"learning_rate": 2.086374267249724e-09,
|
| 2355 |
+
"logits/chosen": -1.9572795629501343,
|
| 2356 |
+
"logits/rejected": -1.5572397708892822,
|
| 2357 |
+
"logps/chosen": -175.27236938476562,
|
| 2358 |
+
"logps/rejected": -249.5878448486328,
|
| 2359 |
+
"loss": 0.6644,
|
| 2360 |
+
"rewards/accuracies": 1.0,
|
| 2361 |
+
"rewards/chosen": 0.03202231973409653,
|
| 2362 |
+
"rewards/margins": 0.0639900267124176,
|
| 2363 |
+
"rewards/margins_max": 0.0639900267124176,
|
| 2364 |
+
"rewards/margins_min": 0.0639900267124176,
|
| 2365 |
+
"rewards/margins_std": NaN,
|
| 2366 |
+
"rewards/rejected": -0.031967706978321075,
|
| 2367 |
+
"step": 1300
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 0.97,
|
| 2371 |
+
"grad_norm": 0.31640625,
|
| 2372 |
+
"learning_rate": 1.335949398349634e-09,
|
| 2373 |
+
"logits/chosen": -1.8132435083389282,
|
| 2374 |
+
"logits/rejected": -1.541999101638794,
|
| 2375 |
+
"logps/chosen": -219.1054229736328,
|
| 2376 |
+
"logps/rejected": -244.1142120361328,
|
| 2377 |
+
"loss": 0.6644,
|
| 2378 |
+
"rewards/accuracies": 1.0,
|
| 2379 |
+
"rewards/chosen": 0.01563115045428276,
|
| 2380 |
+
"rewards/margins": 0.04497561603784561,
|
| 2381 |
+
"rewards/margins_max": 0.04497561603784561,
|
| 2382 |
+
"rewards/margins_min": 0.04497561603784561,
|
| 2383 |
+
"rewards/margins_std": NaN,
|
| 2384 |
+
"rewards/rejected": -0.02934446558356285,
|
| 2385 |
+
"step": 1310
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 0.98,
|
| 2389 |
+
"grad_norm": 0.279296875,
|
| 2390 |
+
"learning_rate": 7.51764708051994e-10,
|
| 2391 |
+
"logits/chosen": -2.0751729011535645,
|
| 2392 |
+
"logits/rejected": -1.464297890663147,
|
| 2393 |
+
"logps/chosen": -209.7403564453125,
|
| 2394 |
+
"logps/rejected": -202.22215270996094,
|
| 2395 |
+
"loss": 0.6649,
|
| 2396 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2397 |
+
"rewards/chosen": 0.03106120601296425,
|
| 2398 |
+
"rewards/margins": 0.054163604974746704,
|
| 2399 |
+
"rewards/margins_max": 0.054163604974746704,
|
| 2400 |
+
"rewards/margins_min": 0.054163604974746704,
|
| 2401 |
+
"rewards/margins_std": NaN,
|
| 2402 |
+
"rewards/rejected": -0.023102398961782455,
|
| 2403 |
+
"step": 1320
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.98,
|
| 2407 |
+
"grad_norm": 0.33203125,
|
| 2408 |
+
"learning_rate": 3.342107432292018e-10,
|
| 2409 |
+
"logits/chosen": -2.0309865474700928,
|
| 2410 |
+
"logits/rejected": -1.6503880023956299,
|
| 2411 |
+
"logps/chosen": -216.9345245361328,
|
| 2412 |
+
"logps/rejected": -258.69244384765625,
|
| 2413 |
+
"loss": 0.6634,
|
| 2414 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2415 |
+
"rewards/chosen": 0.028563106432557106,
|
| 2416 |
+
"rewards/margins": 0.058413028717041016,
|
| 2417 |
+
"rewards/margins_max": 0.058413028717041016,
|
| 2418 |
+
"rewards/margins_min": 0.058413028717041016,
|
| 2419 |
+
"rewards/margins_std": NaN,
|
| 2420 |
+
"rewards/rejected": -0.029849926009774208,
|
| 2421 |
+
"step": 1330
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 0.99,
|
| 2425 |
+
"grad_norm": 0.359375,
|
| 2426 |
+
"learning_rate": 8.356665257813755e-11,
|
| 2427 |
+
"logits/chosen": -1.9864270687103271,
|
| 2428 |
+
"logits/rejected": -1.565781831741333,
|
| 2429 |
+
"logps/chosen": -220.52267456054688,
|
| 2430 |
+
"logps/rejected": -224.6304473876953,
|
| 2431 |
+
"loss": 0.6644,
|
| 2432 |
+
"rewards/accuracies": 1.0,
|
| 2433 |
+
"rewards/chosen": 0.028376713395118713,
|
| 2434 |
+
"rewards/margins": 0.06000853702425957,
|
| 2435 |
+
"rewards/margins_max": 0.06000853702425957,
|
| 2436 |
+
"rewards/margins_min": 0.06000853702425957,
|
| 2437 |
+
"rewards/margins_std": NaN,
|
| 2438 |
+
"rewards/rejected": -0.03163183480501175,
|
| 2439 |
+
"step": 1340
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 1.0,
|
| 2443 |
+
"grad_norm": 0.337890625,
|
| 2444 |
+
"learning_rate": 0.0,
|
| 2445 |
+
"logits/chosen": -2.139542818069458,
|
| 2446 |
+
"logits/rejected": -1.3155004978179932,
|
| 2447 |
+
"logps/chosen": -288.415283203125,
|
| 2448 |
+
"logps/rejected": -208.6166229248047,
|
| 2449 |
+
"loss": 0.6643,
|
| 2450 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2451 |
+
"rewards/chosen": 0.03153330832719803,
|
| 2452 |
+
"rewards/margins": 0.05774186924099922,
|
| 2453 |
+
"rewards/margins_max": 0.05774186924099922,
|
| 2454 |
+
"rewards/margins_min": 0.05774186924099922,
|
| 2455 |
+
"rewards/margins_std": NaN,
|
| 2456 |
+
"rewards/rejected": -0.026208559051156044,
|
| 2457 |
+
"step": 1350
|
| 2458 |
+
},
|
| 2459 |
+
{
|
| 2460 |
+
"epoch": 1.0,
|
| 2461 |
+
"eval_logits/chosen": -1.1541073322296143,
|
| 2462 |
+
"eval_logits/rejected": -1.0117074251174927,
|
| 2463 |
+
"eval_logps/chosen": -321.7533264160156,
|
| 2464 |
+
"eval_logps/rejected": -313.631103515625,
|
| 2465 |
+
"eval_loss": 0.6924247145652771,
|
| 2466 |
+
"eval_rewards/accuracies": 0.5640000104904175,
|
| 2467 |
+
"eval_rewards/chosen": 0.023777291178703308,
|
| 2468 |
+
"eval_rewards/margins": 0.0020897758658975363,
|
| 2469 |
+
"eval_rewards/margins_max": 0.03932913392782211,
|
| 2470 |
+
"eval_rewards/margins_min": -0.036177705973386765,
|
| 2471 |
+
"eval_rewards/margins_std": 0.033525750041007996,
|
| 2472 |
+
"eval_rewards/rejected": 0.021687516942620277,
|
| 2473 |
+
"eval_runtime": 809.7327,
|
| 2474 |
+
"eval_samples_per_second": 9.88,
|
| 2475 |
+
"eval_steps_per_second": 0.309,
|
| 2476 |
+
"step": 1350
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 1.0,
|
| 2480 |
+
"step": 1350,
|
| 2481 |
+
"total_flos": 0.0,
|
| 2482 |
+
"train_loss": 0.6709109875007911,
|
| 2483 |
+
"train_runtime": 7657.6869,
|
| 2484 |
+
"train_samples_per_second": 2.821,
|
| 2485 |
+
"train_steps_per_second": 0.176
|
| 2486 |
+
}
|
| 2487 |
+
],
|
| 2488 |
+
"logging_steps": 10,
|
| 2489 |
+
"max_steps": 1350,
|
| 2490 |
+
"num_input_tokens_seen": 0,
|
| 2491 |
+
"num_train_epochs": 1,
|
| 2492 |
+
"save_steps": 100,
|
| 2493 |
+
"total_flos": 0.0,
|
| 2494 |
+
"train_batch_size": 1,
|
| 2495 |
+
"trial_name": null,
|
| 2496 |
+
"trial_params": null
|
| 2497 |
+
}
|