Model save
Browse files- README.md +81 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +891 -0
README.md
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: alignment-handbook/zephyr-7b-sft-full
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- dpo
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
model-index:
|
| 9 |
+
- name: zephyr-7b-dpo-full-ultrabin-high-curriculum
|
| 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 |
+
# zephyr-7b-dpo-full-ultrabin-high-curriculum
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.5057
|
| 21 |
+
- Rewards/chosen: -1.0106
|
| 22 |
+
- Rewards/rejected: -1.9562
|
| 23 |
+
- Rewards/accuracies: 0.7617
|
| 24 |
+
- Rewards/margins: 0.9455
|
| 25 |
+
- Logps/rejected: -458.2784
|
| 26 |
+
- Logps/chosen: -363.6942
|
| 27 |
+
- Logits/rejected: 2.1024
|
| 28 |
+
- Logits/chosen: 1.3120
|
| 29 |
+
|
| 30 |
+
## Model description
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Intended uses & limitations
|
| 35 |
+
|
| 36 |
+
More information needed
|
| 37 |
+
|
| 38 |
+
## Training and evaluation data
|
| 39 |
+
|
| 40 |
+
More information needed
|
| 41 |
+
|
| 42 |
+
## Training procedure
|
| 43 |
+
|
| 44 |
+
### Training hyperparameters
|
| 45 |
+
|
| 46 |
+
The following hyperparameters were used during training:
|
| 47 |
+
- learning_rate: 5e-07
|
| 48 |
+
- train_batch_size: 8
|
| 49 |
+
- eval_batch_size: 8
|
| 50 |
+
- seed: 55
|
| 51 |
+
- distributed_type: multi-GPU
|
| 52 |
+
- num_devices: 8
|
| 53 |
+
- gradient_accumulation_steps: 2
|
| 54 |
+
- total_train_batch_size: 128
|
| 55 |
+
- total_eval_batch_size: 64
|
| 56 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 57 |
+
- lr_scheduler_type: cosine
|
| 58 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 59 |
+
- num_epochs: 1
|
| 60 |
+
|
| 61 |
+
### Training results
|
| 62 |
+
|
| 63 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 64 |
+
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 65 |
+
| 0.6557 | 0.1046 | 50 | 0.6479 | -0.0110 | -0.1284 | 0.7070 | 0.1174 | -275.5045 | -263.7303 | -2.6116 | -2.6493 |
|
| 66 |
+
| 0.5639 | 0.2092 | 100 | 0.5652 | -0.6096 | -1.2098 | 0.7383 | 0.6002 | -383.6375 | -323.5883 | -0.5618 | -0.9012 |
|
| 67 |
+
| 0.5323 | 0.3138 | 150 | 0.5405 | -0.5498 | -1.2901 | 0.7617 | 0.7402 | -391.6696 | -317.6140 | 0.4792 | -0.0900 |
|
| 68 |
+
| 0.536 | 0.4184 | 200 | 0.5354 | -0.6382 | -1.3831 | 0.7656 | 0.7449 | -400.9734 | -326.4470 | 0.6525 | -0.0195 |
|
| 69 |
+
| 0.5163 | 0.5230 | 250 | 0.5185 | -1.1124 | -1.9604 | 0.7383 | 0.8480 | -458.7008 | -373.8662 | 2.4883 | 1.7620 |
|
| 70 |
+
| 0.5018 | 0.6276 | 300 | 0.5108 | -0.9326 | -1.8124 | 0.7578 | 0.8798 | -443.9044 | -355.8924 | 2.0905 | 1.3198 |
|
| 71 |
+
| 0.4999 | 0.7322 | 350 | 0.5094 | -1.0356 | -1.9491 | 0.7461 | 0.9135 | -457.5764 | -366.1917 | 2.0403 | 1.2353 |
|
| 72 |
+
| 0.4966 | 0.8368 | 400 | 0.5066 | -0.9929 | -1.9227 | 0.7578 | 0.9298 | -454.9321 | -361.9198 | 2.0226 | 1.2642 |
|
| 73 |
+
| 0.5198 | 0.9414 | 450 | 0.5057 | -1.0106 | -1.9562 | 0.7617 | 0.9455 | -458.2784 | -363.6942 | 2.1024 | 1.3120 |
|
| 74 |
+
|
| 75 |
+
|
| 76 |
+
### Framework versions
|
| 77 |
+
|
| 78 |
+
- Transformers 4.44.0.dev0
|
| 79 |
+
- Pytorch 2.1.2
|
| 80 |
+
- Datasets 2.20.0
|
| 81 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.5417051395113,
|
| 5 |
+
"train_runtime": 12463.9447,
|
| 6 |
+
"train_samples": 61134,
|
| 7 |
+
"train_samples_per_second": 4.905,
|
| 8 |
+
"train_steps_per_second": 0.038
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.44.0.dev0"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 0.5417051395113,
|
| 5 |
+
"train_runtime": 12463.9447,
|
| 6 |
+
"train_samples": 61134,
|
| 7 |
+
"train_samples_per_second": 4.905,
|
| 8 |
+
"train_steps_per_second": 0.038
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,891 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 478,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02092050209205021,
|
| 13 |
+
"grad_norm": 9.137621785616465,
|
| 14 |
+
"learning_rate": 1.0416666666666667e-07,
|
| 15 |
+
"logits/chosen": -2.6498842239379883,
|
| 16 |
+
"logits/rejected": -2.6401047706604004,
|
| 17 |
+
"logps/chosen": -294.3343505859375,
|
| 18 |
+
"logps/rejected": -289.1372985839844,
|
| 19 |
+
"loss": 0.6931,
|
| 20 |
+
"rewards/accuracies": 0.41874998807907104,
|
| 21 |
+
"rewards/chosen": -0.0002774398017209023,
|
| 22 |
+
"rewards/margins": 0.00026476362836547196,
|
| 23 |
+
"rewards/rejected": -0.0005422033718787134,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.04184100418410042,
|
| 28 |
+
"grad_norm": 8.382944921921224,
|
| 29 |
+
"learning_rate": 2.0833333333333333e-07,
|
| 30 |
+
"logits/chosen": -2.6785335540771484,
|
| 31 |
+
"logits/rejected": -2.6339826583862305,
|
| 32 |
+
"logps/chosen": -294.88311767578125,
|
| 33 |
+
"logps/rejected": -276.8221130371094,
|
| 34 |
+
"loss": 0.6926,
|
| 35 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 36 |
+
"rewards/chosen": 0.000260740373050794,
|
| 37 |
+
"rewards/margins": 0.001288960687816143,
|
| 38 |
+
"rewards/rejected": -0.0010282204020768404,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.06276150627615062,
|
| 43 |
+
"grad_norm": 8.348013142376631,
|
| 44 |
+
"learning_rate": 3.1249999999999997e-07,
|
| 45 |
+
"logits/chosen": -2.642605781555176,
|
| 46 |
+
"logits/rejected": -2.6058638095855713,
|
| 47 |
+
"logps/chosen": -282.0052185058594,
|
| 48 |
+
"logps/rejected": -247.48208618164062,
|
| 49 |
+
"loss": 0.6882,
|
| 50 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 51 |
+
"rewards/chosen": 0.006200692616403103,
|
| 52 |
+
"rewards/margins": 0.009530355222523212,
|
| 53 |
+
"rewards/rejected": -0.0033296621404588223,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.08368200836820083,
|
| 58 |
+
"grad_norm": 8.873941277767704,
|
| 59 |
+
"learning_rate": 4.1666666666666667e-07,
|
| 60 |
+
"logits/chosen": -2.6004221439361572,
|
| 61 |
+
"logits/rejected": -2.5648019313812256,
|
| 62 |
+
"logps/chosen": -263.68304443359375,
|
| 63 |
+
"logps/rejected": -269.95159912109375,
|
| 64 |
+
"loss": 0.6776,
|
| 65 |
+
"rewards/accuracies": 0.65625,
|
| 66 |
+
"rewards/chosen": 0.0284257885068655,
|
| 67 |
+
"rewards/margins": 0.026932990178465843,
|
| 68 |
+
"rewards/rejected": 0.001492797746323049,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.10460251046025104,
|
| 73 |
+
"grad_norm": 9.18858486699177,
|
| 74 |
+
"learning_rate": 4.999733114418725e-07,
|
| 75 |
+
"logits/chosen": -2.5885472297668457,
|
| 76 |
+
"logits/rejected": -2.5707509517669678,
|
| 77 |
+
"logps/chosen": -269.8135070800781,
|
| 78 |
+
"logps/rejected": -265.47064208984375,
|
| 79 |
+
"loss": 0.6557,
|
| 80 |
+
"rewards/accuracies": 0.706250011920929,
|
| 81 |
+
"rewards/chosen": 0.018491629511117935,
|
| 82 |
+
"rewards/margins": 0.0841543972492218,
|
| 83 |
+
"rewards/rejected": -0.06566276401281357,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.10460251046025104,
|
| 88 |
+
"eval_logits/chosen": -2.649272918701172,
|
| 89 |
+
"eval_logits/rejected": -2.611584424972534,
|
| 90 |
+
"eval_logps/chosen": -263.7303466796875,
|
| 91 |
+
"eval_logps/rejected": -275.50445556640625,
|
| 92 |
+
"eval_loss": 0.6479206681251526,
|
| 93 |
+
"eval_rewards/accuracies": 0.70703125,
|
| 94 |
+
"eval_rewards/chosen": -0.011004311963915825,
|
| 95 |
+
"eval_rewards/margins": 0.11741933971643448,
|
| 96 |
+
"eval_rewards/rejected": -0.12842364609241486,
|
| 97 |
+
"eval_runtime": 102.963,
|
| 98 |
+
"eval_samples_per_second": 19.424,
|
| 99 |
+
"eval_steps_per_second": 0.311,
|
| 100 |
+
"step": 50
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.12552301255230125,
|
| 104 |
+
"grad_norm": 12.415021202754499,
|
| 105 |
+
"learning_rate": 4.990398100856366e-07,
|
| 106 |
+
"logits/chosen": -2.608935832977295,
|
| 107 |
+
"logits/rejected": -2.5622076988220215,
|
| 108 |
+
"logps/chosen": -316.8729553222656,
|
| 109 |
+
"logps/rejected": -300.7294616699219,
|
| 110 |
+
"loss": 0.6376,
|
| 111 |
+
"rewards/accuracies": 0.65625,
|
| 112 |
+
"rewards/chosen": -0.07160066068172455,
|
| 113 |
+
"rewards/margins": 0.12479270994663239,
|
| 114 |
+
"rewards/rejected": -0.19639338552951813,
|
| 115 |
+
"step": 60
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.14644351464435146,
|
| 119 |
+
"grad_norm": 16.95023916970366,
|
| 120 |
+
"learning_rate": 4.967775735898179e-07,
|
| 121 |
+
"logits/chosen": -2.5255544185638428,
|
| 122 |
+
"logits/rejected": -2.4770090579986572,
|
| 123 |
+
"logps/chosen": -310.1564636230469,
|
| 124 |
+
"logps/rejected": -286.7000427246094,
|
| 125 |
+
"loss": 0.6174,
|
| 126 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 127 |
+
"rewards/chosen": -0.1686697006225586,
|
| 128 |
+
"rewards/margins": 0.18098047375679016,
|
| 129 |
+
"rewards/rejected": -0.34965020418167114,
|
| 130 |
+
"step": 70
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.16736401673640167,
|
| 134 |
+
"grad_norm": 14.651689743135485,
|
| 135 |
+
"learning_rate": 4.931986719649298e-07,
|
| 136 |
+
"logits/chosen": -2.3215534687042236,
|
| 137 |
+
"logits/rejected": -2.257579803466797,
|
| 138 |
+
"logps/chosen": -285.29833984375,
|
| 139 |
+
"logps/rejected": -311.03729248046875,
|
| 140 |
+
"loss": 0.6044,
|
| 141 |
+
"rewards/accuracies": 0.675000011920929,
|
| 142 |
+
"rewards/chosen": -0.07682778686285019,
|
| 143 |
+
"rewards/margins": 0.2939576804637909,
|
| 144 |
+
"rewards/rejected": -0.3707854449748993,
|
| 145 |
+
"step": 80
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.18828451882845187,
|
| 149 |
+
"grad_norm": 19.539798910302995,
|
| 150 |
+
"learning_rate": 4.883222001996351e-07,
|
| 151 |
+
"logits/chosen": -1.3052003383636475,
|
| 152 |
+
"logits/rejected": -1.1468216180801392,
|
| 153 |
+
"logps/chosen": -294.7815246582031,
|
| 154 |
+
"logps/rejected": -303.13427734375,
|
| 155 |
+
"loss": 0.5866,
|
| 156 |
+
"rewards/accuracies": 0.737500011920929,
|
| 157 |
+
"rewards/chosen": -0.328985333442688,
|
| 158 |
+
"rewards/margins": 0.45616284012794495,
|
| 159 |
+
"rewards/rejected": -0.7851482629776001,
|
| 160 |
+
"step": 90
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.20920502092050208,
|
| 164 |
+
"grad_norm": 25.641340302712948,
|
| 165 |
+
"learning_rate": 4.821741763807186e-07,
|
| 166 |
+
"logits/chosen": -0.7993453145027161,
|
| 167 |
+
"logits/rejected": -0.40382108092308044,
|
| 168 |
+
"logps/chosen": -333.15521240234375,
|
| 169 |
+
"logps/rejected": -387.1316833496094,
|
| 170 |
+
"loss": 0.5639,
|
| 171 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 172 |
+
"rewards/chosen": -0.5871611833572388,
|
| 173 |
+
"rewards/margins": 0.6322518587112427,
|
| 174 |
+
"rewards/rejected": -1.2194130420684814,
|
| 175 |
+
"step": 100
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.20920502092050208,
|
| 179 |
+
"eval_logits/chosen": -0.9011980295181274,
|
| 180 |
+
"eval_logits/rejected": -0.5617944598197937,
|
| 181 |
+
"eval_logps/chosen": -323.5882873535156,
|
| 182 |
+
"eval_logps/rejected": -383.63751220703125,
|
| 183 |
+
"eval_loss": 0.5651677846908569,
|
| 184 |
+
"eval_rewards/accuracies": 0.73828125,
|
| 185 |
+
"eval_rewards/chosen": -0.6095837354660034,
|
| 186 |
+
"eval_rewards/margins": 0.6001702547073364,
|
| 187 |
+
"eval_rewards/rejected": -1.2097539901733398,
|
| 188 |
+
"eval_runtime": 101.4152,
|
| 189 |
+
"eval_samples_per_second": 19.721,
|
| 190 |
+
"eval_steps_per_second": 0.316,
|
| 191 |
+
"step": 100
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.2301255230125523,
|
| 195 |
+
"grad_norm": 21.689227912073658,
|
| 196 |
+
"learning_rate": 4.747874028753375e-07,
|
| 197 |
+
"logits/chosen": -0.8007869720458984,
|
| 198 |
+
"logits/rejected": -0.4384717047214508,
|
| 199 |
+
"logps/chosen": -323.02410888671875,
|
| 200 |
+
"logps/rejected": -343.52508544921875,
|
| 201 |
+
"loss": 0.5585,
|
| 202 |
+
"rewards/accuracies": 0.65625,
|
| 203 |
+
"rewards/chosen": -0.5073341727256775,
|
| 204 |
+
"rewards/margins": 0.517856240272522,
|
| 205 |
+
"rewards/rejected": -1.0251905918121338,
|
| 206 |
+
"step": 110
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.2510460251046025,
|
| 210 |
+
"grad_norm": 40.51442433382909,
|
| 211 |
+
"learning_rate": 4.662012913161997e-07,
|
| 212 |
+
"logits/chosen": 0.0625094398856163,
|
| 213 |
+
"logits/rejected": 0.6715419292449951,
|
| 214 |
+
"logps/chosen": -367.49664306640625,
|
| 215 |
+
"logps/rejected": -383.600830078125,
|
| 216 |
+
"loss": 0.5396,
|
| 217 |
+
"rewards/accuracies": 0.768750011920929,
|
| 218 |
+
"rewards/chosen": -0.6050666570663452,
|
| 219 |
+
"rewards/margins": 0.6428467631340027,
|
| 220 |
+
"rewards/rejected": -1.2479135990142822,
|
| 221 |
+
"step": 120
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.2719665271966527,
|
| 225 |
+
"grad_norm": 31.04938830573247,
|
| 226 |
+
"learning_rate": 4.5646165232345103e-07,
|
| 227 |
+
"logits/chosen": 0.35010597109794617,
|
| 228 |
+
"logits/rejected": 1.0627130270004272,
|
| 229 |
+
"logps/chosen": -369.3751220703125,
|
| 230 |
+
"logps/rejected": -391.6081237792969,
|
| 231 |
+
"loss": 0.5421,
|
| 232 |
+
"rewards/accuracies": 0.737500011920929,
|
| 233 |
+
"rewards/chosen": -0.6456856727600098,
|
| 234 |
+
"rewards/margins": 0.6666911840438843,
|
| 235 |
+
"rewards/rejected": -1.312376856803894,
|
| 236 |
+
"step": 130
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.2928870292887029,
|
| 240 |
+
"grad_norm": 23.124000981376877,
|
| 241 |
+
"learning_rate": 4.456204510851956e-07,
|
| 242 |
+
"logits/chosen": 1.3312932252883911,
|
| 243 |
+
"logits/rejected": 1.6921491622924805,
|
| 244 |
+
"logps/chosen": -328.2577209472656,
|
| 245 |
+
"logps/rejected": -374.11846923828125,
|
| 246 |
+
"loss": 0.5318,
|
| 247 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 248 |
+
"rewards/chosen": -0.7673874497413635,
|
| 249 |
+
"rewards/margins": 0.5369923710823059,
|
| 250 |
+
"rewards/rejected": -1.3043798208236694,
|
| 251 |
+
"step": 140
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.3138075313807531,
|
| 255 |
+
"grad_norm": 28.62578465081787,
|
| 256 |
+
"learning_rate": 4.337355301007335e-07,
|
| 257 |
+
"logits/chosen": 0.7725589871406555,
|
| 258 |
+
"logits/rejected": 1.4066698551177979,
|
| 259 |
+
"logps/chosen": -374.64288330078125,
|
| 260 |
+
"logps/rejected": -426.3128356933594,
|
| 261 |
+
"loss": 0.5323,
|
| 262 |
+
"rewards/accuracies": 0.737500011920929,
|
| 263 |
+
"rewards/chosen": -0.7278563380241394,
|
| 264 |
+
"rewards/margins": 0.7114464044570923,
|
| 265 |
+
"rewards/rejected": -1.439302682876587,
|
| 266 |
+
"step": 150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.3138075313807531,
|
| 270 |
+
"eval_logits/chosen": -0.09001250565052032,
|
| 271 |
+
"eval_logits/rejected": 0.4792196750640869,
|
| 272 |
+
"eval_logps/chosen": -317.6140441894531,
|
| 273 |
+
"eval_logps/rejected": -391.6695556640625,
|
| 274 |
+
"eval_loss": 0.5405200123786926,
|
| 275 |
+
"eval_rewards/accuracies": 0.76171875,
|
| 276 |
+
"eval_rewards/chosen": -0.5498415231704712,
|
| 277 |
+
"eval_rewards/margins": 0.7402328252792358,
|
| 278 |
+
"eval_rewards/rejected": -1.290074348449707,
|
| 279 |
+
"eval_runtime": 102.6926,
|
| 280 |
+
"eval_samples_per_second": 19.476,
|
| 281 |
+
"eval_steps_per_second": 0.312,
|
| 282 |
+
"step": 150
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.33472803347280333,
|
| 286 |
+
"grad_norm": 26.58643912067179,
|
| 287 |
+
"learning_rate": 4.2087030056579986e-07,
|
| 288 |
+
"logits/chosen": -0.5741851925849915,
|
| 289 |
+
"logits/rejected": 0.09496372938156128,
|
| 290 |
+
"logps/chosen": -313.5487976074219,
|
| 291 |
+
"logps/rejected": -365.291015625,
|
| 292 |
+
"loss": 0.5399,
|
| 293 |
+
"rewards/accuracies": 0.75,
|
| 294 |
+
"rewards/chosen": -0.5437325835227966,
|
| 295 |
+
"rewards/margins": 0.6792270541191101,
|
| 296 |
+
"rewards/rejected": -1.2229596376419067,
|
| 297 |
+
"step": 160
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.35564853556485354,
|
| 301 |
+
"grad_norm": 23.464961776068243,
|
| 302 |
+
"learning_rate": 4.070934040463998e-07,
|
| 303 |
+
"logits/chosen": -0.5381526947021484,
|
| 304 |
+
"logits/rejected": 0.14733538031578064,
|
| 305 |
+
"logps/chosen": -354.4422912597656,
|
| 306 |
+
"logps/rejected": -430.18682861328125,
|
| 307 |
+
"loss": 0.5326,
|
| 308 |
+
"rewards/accuracies": 0.75,
|
| 309 |
+
"rewards/chosen": -0.7926353216171265,
|
| 310 |
+
"rewards/margins": 0.8019927740097046,
|
| 311 |
+
"rewards/rejected": -1.594628095626831,
|
| 312 |
+
"step": 170
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.37656903765690375,
|
| 316 |
+
"grad_norm": 22.467354382291056,
|
| 317 |
+
"learning_rate": 3.9247834624635404e-07,
|
| 318 |
+
"logits/chosen": -0.20919005572795868,
|
| 319 |
+
"logits/rejected": 0.49986928701400757,
|
| 320 |
+
"logps/chosen": -371.064697265625,
|
| 321 |
+
"logps/rejected": -408.84149169921875,
|
| 322 |
+
"loss": 0.5358,
|
| 323 |
+
"rewards/accuracies": 0.706250011920929,
|
| 324 |
+
"rewards/chosen": -0.9278782606124878,
|
| 325 |
+
"rewards/margins": 0.5680198669433594,
|
| 326 |
+
"rewards/rejected": -1.4958980083465576,
|
| 327 |
+
"step": 180
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.39748953974895396,
|
| 331 |
+
"grad_norm": 29.254730102983988,
|
| 332 |
+
"learning_rate": 3.7710310482256523e-07,
|
| 333 |
+
"logits/chosen": -0.3338702619075775,
|
| 334 |
+
"logits/rejected": 0.38479360938072205,
|
| 335 |
+
"logps/chosen": -319.56805419921875,
|
| 336 |
+
"logps/rejected": -378.23590087890625,
|
| 337 |
+
"loss": 0.5412,
|
| 338 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 339 |
+
"rewards/chosen": -0.709225058555603,
|
| 340 |
+
"rewards/margins": 0.857193648815155,
|
| 341 |
+
"rewards/rejected": -1.5664187669754028,
|
| 342 |
+
"step": 190
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.41841004184100417,
|
| 346 |
+
"grad_norm": 21.764790734639657,
|
| 347 |
+
"learning_rate": 3.610497133404795e-07,
|
| 348 |
+
"logits/chosen": -0.4434884190559387,
|
| 349 |
+
"logits/rejected": 0.06593348830938339,
|
| 350 |
+
"logps/chosen": -343.6837463378906,
|
| 351 |
+
"logps/rejected": -397.1659240722656,
|
| 352 |
+
"loss": 0.536,
|
| 353 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 354 |
+
"rewards/chosen": -0.6647871136665344,
|
| 355 |
+
"rewards/margins": 0.6501897573471069,
|
| 356 |
+
"rewards/rejected": -1.314976692199707,
|
| 357 |
+
"step": 200
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.41841004184100417,
|
| 361 |
+
"eval_logits/chosen": -0.019472889602184296,
|
| 362 |
+
"eval_logits/rejected": 0.65250563621521,
|
| 363 |
+
"eval_logps/chosen": -326.44696044921875,
|
| 364 |
+
"eval_logps/rejected": -400.973388671875,
|
| 365 |
+
"eval_loss": 0.5353578329086304,
|
| 366 |
+
"eval_rewards/accuracies": 0.765625,
|
| 367 |
+
"eval_rewards/chosen": -0.6381703019142151,
|
| 368 |
+
"eval_rewards/margins": 0.7449426054954529,
|
| 369 |
+
"eval_rewards/rejected": -1.3831130266189575,
|
| 370 |
+
"eval_runtime": 102.2662,
|
| 371 |
+
"eval_samples_per_second": 19.557,
|
| 372 |
+
"eval_steps_per_second": 0.313,
|
| 373 |
+
"step": 200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.4393305439330544,
|
| 377 |
+
"grad_norm": 27.745510528591602,
|
| 378 |
+
"learning_rate": 3.4440382358952115e-07,
|
| 379 |
+
"logits/chosen": 1.1720086336135864,
|
| 380 |
+
"logits/rejected": 1.8830058574676514,
|
| 381 |
+
"logps/chosen": -403.1050109863281,
|
| 382 |
+
"logps/rejected": -435.21942138671875,
|
| 383 |
+
"loss": 0.5366,
|
| 384 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 385 |
+
"rewards/chosen": -0.988387405872345,
|
| 386 |
+
"rewards/margins": 0.5626603960990906,
|
| 387 |
+
"rewards/rejected": -1.5510478019714355,
|
| 388 |
+
"step": 210
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.4602510460251046,
|
| 392 |
+
"grad_norm": 22.068359980003752,
|
| 393 |
+
"learning_rate": 3.272542485937368e-07,
|
| 394 |
+
"logits/chosen": 1.7870858907699585,
|
| 395 |
+
"logits/rejected": 2.4022154808044434,
|
| 396 |
+
"logps/chosen": -351.79974365234375,
|
| 397 |
+
"logps/rejected": -414.64129638671875,
|
| 398 |
+
"loss": 0.5367,
|
| 399 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 400 |
+
"rewards/chosen": -0.945723831653595,
|
| 401 |
+
"rewards/margins": 0.6799818277359009,
|
| 402 |
+
"rewards/rejected": -1.6257057189941406,
|
| 403 |
+
"step": 220
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.4811715481171548,
|
| 407 |
+
"grad_norm": 24.112210218904135,
|
| 408 |
+
"learning_rate": 3.096924887558854e-07,
|
| 409 |
+
"logits/chosen": 1.5326052904129028,
|
| 410 |
+
"logits/rejected": 1.9995750188827515,
|
| 411 |
+
"logps/chosen": -340.270263671875,
|
| 412 |
+
"logps/rejected": -413.39385986328125,
|
| 413 |
+
"loss": 0.494,
|
| 414 |
+
"rewards/accuracies": 0.768750011920929,
|
| 415 |
+
"rewards/chosen": -0.8984026908874512,
|
| 416 |
+
"rewards/margins": 0.7524574995040894,
|
| 417 |
+
"rewards/rejected": -1.650860071182251,
|
| 418 |
+
"step": 230
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.502092050209205,
|
| 422 |
+
"grad_norm": 27.75263813297474,
|
| 423 |
+
"learning_rate": 2.9181224366319943e-07,
|
| 424 |
+
"logits/chosen": 0.5852566361427307,
|
| 425 |
+
"logits/rejected": 1.34577214717865,
|
| 426 |
+
"logps/chosen": -346.68267822265625,
|
| 427 |
+
"logps/rejected": -395.8317565917969,
|
| 428 |
+
"loss": 0.5068,
|
| 429 |
+
"rewards/accuracies": 0.731249988079071,
|
| 430 |
+
"rewards/chosen": -0.6660369634628296,
|
| 431 |
+
"rewards/margins": 0.7894813418388367,
|
| 432 |
+
"rewards/rejected": -1.4555184841156006,
|
| 433 |
+
"step": 240
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.5230125523012552,
|
| 437 |
+
"grad_norm": 34.032720640859,
|
| 438 |
+
"learning_rate": 2.7370891215954565e-07,
|
| 439 |
+
"logits/chosen": 1.4959585666656494,
|
| 440 |
+
"logits/rejected": 2.515637159347534,
|
| 441 |
+
"logps/chosen": -362.95111083984375,
|
| 442 |
+
"logps/rejected": -426.1146545410156,
|
| 443 |
+
"loss": 0.5163,
|
| 444 |
+
"rewards/accuracies": 0.71875,
|
| 445 |
+
"rewards/chosen": -1.0264902114868164,
|
| 446 |
+
"rewards/margins": 0.8269758224487305,
|
| 447 |
+
"rewards/rejected": -1.8534657955169678,
|
| 448 |
+
"step": 250
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.5230125523012552,
|
| 452 |
+
"eval_logits/chosen": 1.7619643211364746,
|
| 453 |
+
"eval_logits/rejected": 2.488286018371582,
|
| 454 |
+
"eval_logps/chosen": -373.8662414550781,
|
| 455 |
+
"eval_logps/rejected": -458.7007751464844,
|
| 456 |
+
"eval_loss": 0.5184882879257202,
|
| 457 |
+
"eval_rewards/accuracies": 0.73828125,
|
| 458 |
+
"eval_rewards/chosen": -1.112363576889038,
|
| 459 |
+
"eval_rewards/margins": 0.8480234742164612,
|
| 460 |
+
"eval_rewards/rejected": -1.9603869915008545,
|
| 461 |
+
"eval_runtime": 103.2706,
|
| 462 |
+
"eval_samples_per_second": 19.367,
|
| 463 |
+
"eval_steps_per_second": 0.31,
|
| 464 |
+
"step": 250
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.5439330543933054,
|
| 468 |
+
"grad_norm": 24.412359940634452,
|
| 469 |
+
"learning_rate": 2.55479083351317e-07,
|
| 470 |
+
"logits/chosen": 1.4174010753631592,
|
| 471 |
+
"logits/rejected": 2.5714869499206543,
|
| 472 |
+
"logps/chosen": -396.62353515625,
|
| 473 |
+
"logps/rejected": -448.77813720703125,
|
| 474 |
+
"loss": 0.5172,
|
| 475 |
+
"rewards/accuracies": 0.6875,
|
| 476 |
+
"rewards/chosen": -1.168769121170044,
|
| 477 |
+
"rewards/margins": 0.7762032747268677,
|
| 478 |
+
"rewards/rejected": -1.9449723958969116,
|
| 479 |
+
"step": 260
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.5648535564853556,
|
| 483 |
+
"grad_norm": 27.094533941945745,
|
| 484 |
+
"learning_rate": 2.3722002126275822e-07,
|
| 485 |
+
"logits/chosen": 0.924826979637146,
|
| 486 |
+
"logits/rejected": 2.0330166816711426,
|
| 487 |
+
"logps/chosen": -410.7693786621094,
|
| 488 |
+
"logps/rejected": -429.46466064453125,
|
| 489 |
+
"loss": 0.5165,
|
| 490 |
+
"rewards/accuracies": 0.731249988079071,
|
| 491 |
+
"rewards/chosen": -0.9795888662338257,
|
| 492 |
+
"rewards/margins": 0.6443194150924683,
|
| 493 |
+
"rewards/rejected": -1.623908281326294,
|
| 494 |
+
"step": 270
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.5857740585774058,
|
| 498 |
+
"grad_norm": 25.27079656796927,
|
| 499 |
+
"learning_rate": 2.19029145890313e-07,
|
| 500 |
+
"logits/chosen": 1.045302391052246,
|
| 501 |
+
"logits/rejected": 1.8793681859970093,
|
| 502 |
+
"logps/chosen": -354.340576171875,
|
| 503 |
+
"logps/rejected": -408.4857177734375,
|
| 504 |
+
"loss": 0.5031,
|
| 505 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 506 |
+
"rewards/chosen": -0.960013210773468,
|
| 507 |
+
"rewards/margins": 0.848355770111084,
|
| 508 |
+
"rewards/rejected": -1.8083690404891968,
|
| 509 |
+
"step": 280
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.606694560669456,
|
| 513 |
+
"grad_norm": 33.50688260105116,
|
| 514 |
+
"learning_rate": 2.0100351342479216e-07,
|
| 515 |
+
"logits/chosen": 0.9456893801689148,
|
| 516 |
+
"logits/rejected": 2.003875255584717,
|
| 517 |
+
"logps/chosen": -404.9028015136719,
|
| 518 |
+
"logps/rejected": -436.6299743652344,
|
| 519 |
+
"loss": 0.496,
|
| 520 |
+
"rewards/accuracies": 0.800000011920929,
|
| 521 |
+
"rewards/chosen": -0.9352100491523743,
|
| 522 |
+
"rewards/margins": 0.8176320791244507,
|
| 523 |
+
"rewards/rejected": -1.7528421878814697,
|
| 524 |
+
"step": 290
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.6276150627615062,
|
| 528 |
+
"grad_norm": 31.055623020499947,
|
| 529 |
+
"learning_rate": 1.8323929841460178e-07,
|
| 530 |
+
"logits/chosen": 1.6724220514297485,
|
| 531 |
+
"logits/rejected": 2.5003108978271484,
|
| 532 |
+
"logps/chosen": -367.1915283203125,
|
| 533 |
+
"logps/rejected": -454.2134704589844,
|
| 534 |
+
"loss": 0.5018,
|
| 535 |
+
"rewards/accuracies": 0.6875,
|
| 536 |
+
"rewards/chosen": -1.1671972274780273,
|
| 537 |
+
"rewards/margins": 0.8203157186508179,
|
| 538 |
+
"rewards/rejected": -1.9875129461288452,
|
| 539 |
+
"step": 300
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.6276150627615062,
|
| 543 |
+
"eval_logits/chosen": 1.3198010921478271,
|
| 544 |
+
"eval_logits/rejected": 2.090496778488159,
|
| 545 |
+
"eval_logps/chosen": -355.89239501953125,
|
| 546 |
+
"eval_logps/rejected": -443.9043884277344,
|
| 547 |
+
"eval_loss": 0.5108399987220764,
|
| 548 |
+
"eval_rewards/accuracies": 0.7578125,
|
| 549 |
+
"eval_rewards/chosen": -0.9326243996620178,
|
| 550 |
+
"eval_rewards/margins": 0.879798412322998,
|
| 551 |
+
"eval_rewards/rejected": -1.8124228715896606,
|
| 552 |
+
"eval_runtime": 103.5061,
|
| 553 |
+
"eval_samples_per_second": 19.323,
|
| 554 |
+
"eval_steps_per_second": 0.309,
|
| 555 |
+
"step": 300
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.6485355648535565,
|
| 559 |
+
"grad_norm": 26.914602226083172,
|
| 560 |
+
"learning_rate": 1.6583128063291573e-07,
|
| 561 |
+
"logits/chosen": 1.2696157693862915,
|
| 562 |
+
"logits/rejected": 2.2371857166290283,
|
| 563 |
+
"logps/chosen": -391.40081787109375,
|
| 564 |
+
"logps/rejected": -433.94384765625,
|
| 565 |
+
"loss": 0.5165,
|
| 566 |
+
"rewards/accuracies": 0.762499988079071,
|
| 567 |
+
"rewards/chosen": -0.8740479350090027,
|
| 568 |
+
"rewards/margins": 0.8631827235221863,
|
| 569 |
+
"rewards/rejected": -1.737230658531189,
|
| 570 |
+
"step": 310
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.6694560669456067,
|
| 574 |
+
"grad_norm": 23.39436395049965,
|
| 575 |
+
"learning_rate": 1.488723393865766e-07,
|
| 576 |
+
"logits/chosen": 0.8202553987503052,
|
| 577 |
+
"logits/rejected": 1.8485848903656006,
|
| 578 |
+
"logps/chosen": -368.58477783203125,
|
| 579 |
+
"logps/rejected": -417.188720703125,
|
| 580 |
+
"loss": 0.5097,
|
| 581 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 582 |
+
"rewards/chosen": -0.8171142339706421,
|
| 583 |
+
"rewards/margins": 0.8511916995048523,
|
| 584 |
+
"rewards/rejected": -1.6683059930801392,
|
| 585 |
+
"step": 320
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.6903765690376569,
|
| 589 |
+
"grad_norm": 24.464796036697045,
|
| 590 |
+
"learning_rate": 1.3245295796480788e-07,
|
| 591 |
+
"logits/chosen": 0.7323199510574341,
|
| 592 |
+
"logits/rejected": 1.601609230041504,
|
| 593 |
+
"logps/chosen": -404.5626525878906,
|
| 594 |
+
"logps/rejected": -442.4979553222656,
|
| 595 |
+
"loss": 0.5021,
|
| 596 |
+
"rewards/accuracies": 0.737500011920929,
|
| 597 |
+
"rewards/chosen": -1.0547947883605957,
|
| 598 |
+
"rewards/margins": 0.6202768087387085,
|
| 599 |
+
"rewards/rejected": -1.6750714778900146,
|
| 600 |
+
"step": 330
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.7112970711297071,
|
| 604 |
+
"grad_norm": 28.602197351940898,
|
| 605 |
+
"learning_rate": 1.1666074087171627e-07,
|
| 606 |
+
"logits/chosen": 1.1269073486328125,
|
| 607 |
+
"logits/rejected": 2.291718006134033,
|
| 608 |
+
"logps/chosen": -386.19696044921875,
|
| 609 |
+
"logps/rejected": -446.26080322265625,
|
| 610 |
+
"loss": 0.51,
|
| 611 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 612 |
+
"rewards/chosen": -1.0688059329986572,
|
| 613 |
+
"rewards/margins": 0.8116031885147095,
|
| 614 |
+
"rewards/rejected": -1.8804088830947876,
|
| 615 |
+
"step": 340
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.7322175732217573,
|
| 619 |
+
"grad_norm": 27.032175742818502,
|
| 620 |
+
"learning_rate": 1.0157994641835734e-07,
|
| 621 |
+
"logits/chosen": 1.3312422037124634,
|
| 622 |
+
"logits/rejected": 2.5307185649871826,
|
| 623 |
+
"logps/chosen": -407.82183837890625,
|
| 624 |
+
"logps/rejected": -435.91339111328125,
|
| 625 |
+
"loss": 0.4999,
|
| 626 |
+
"rewards/accuracies": 0.768750011920929,
|
| 627 |
+
"rewards/chosen": -1.1246039867401123,
|
| 628 |
+
"rewards/margins": 0.9464033246040344,
|
| 629 |
+
"rewards/rejected": -2.071007490158081,
|
| 630 |
+
"step": 350
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 0.7322175732217573,
|
| 634 |
+
"eval_logits/chosen": 1.2352690696716309,
|
| 635 |
+
"eval_logits/rejected": 2.0403130054473877,
|
| 636 |
+
"eval_logps/chosen": -366.191650390625,
|
| 637 |
+
"eval_logps/rejected": -457.5764465332031,
|
| 638 |
+
"eval_loss": 0.5094150900840759,
|
| 639 |
+
"eval_rewards/accuracies": 0.74609375,
|
| 640 |
+
"eval_rewards/chosen": -1.035617470741272,
|
| 641 |
+
"eval_rewards/margins": 0.9135259389877319,
|
| 642 |
+
"eval_rewards/rejected": -1.949143409729004,
|
| 643 |
+
"eval_runtime": 101.7074,
|
| 644 |
+
"eval_samples_per_second": 19.664,
|
| 645 |
+
"eval_steps_per_second": 0.315,
|
| 646 |
+
"step": 350
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.7531380753138075,
|
| 650 |
+
"grad_norm": 27.108638768821233,
|
| 651 |
+
"learning_rate": 8.729103716819111e-08,
|
| 652 |
+
"logits/chosen": 1.3057953119277954,
|
| 653 |
+
"logits/rejected": 2.3061721324920654,
|
| 654 |
+
"logps/chosen": -390.41973876953125,
|
| 655 |
+
"logps/rejected": -483.9664001464844,
|
| 656 |
+
"loss": 0.4746,
|
| 657 |
+
"rewards/accuracies": 0.75,
|
| 658 |
+
"rewards/chosen": -1.1125240325927734,
|
| 659 |
+
"rewards/margins": 0.8978655934333801,
|
| 660 |
+
"rewards/rejected": -2.0103893280029297,
|
| 661 |
+
"step": 360
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.7740585774058577,
|
| 665 |
+
"grad_norm": 23.348973436681856,
|
| 666 |
+
"learning_rate": 7.387025063449081e-08,
|
| 667 |
+
"logits/chosen": 1.0233042240142822,
|
| 668 |
+
"logits/rejected": 1.8939619064331055,
|
| 669 |
+
"logps/chosen": -426.7222595214844,
|
| 670 |
+
"logps/rejected": -498.1934509277344,
|
| 671 |
+
"loss": 0.5237,
|
| 672 |
+
"rewards/accuracies": 0.793749988079071,
|
| 673 |
+
"rewards/chosen": -0.9773656725883484,
|
| 674 |
+
"rewards/margins": 1.0110007524490356,
|
| 675 |
+
"rewards/rejected": -1.9883663654327393,
|
| 676 |
+
"step": 370
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.7949790794979079,
|
| 680 |
+
"grad_norm": 23.77690961130278,
|
| 681 |
+
"learning_rate": 6.138919252022435e-08,
|
| 682 |
+
"logits/chosen": 0.7093733549118042,
|
| 683 |
+
"logits/rejected": 1.795745849609375,
|
| 684 |
+
"logps/chosen": -405.4716796875,
|
| 685 |
+
"logps/rejected": -439.33441162109375,
|
| 686 |
+
"loss": 0.5086,
|
| 687 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 688 |
+
"rewards/chosen": -0.994636058807373,
|
| 689 |
+
"rewards/margins": 0.8113524317741394,
|
| 690 |
+
"rewards/rejected": -1.8059885501861572,
|
| 691 |
+
"step": 380
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.8158995815899581,
|
| 695 |
+
"grad_norm": 24.44927335880225,
|
| 696 |
+
"learning_rate": 4.991445467064689e-08,
|
| 697 |
+
"logits/chosen": 1.5607655048370361,
|
| 698 |
+
"logits/rejected": 1.7856521606445312,
|
| 699 |
+
"logps/chosen": -352.96942138671875,
|
| 700 |
+
"logps/rejected": -442.06964111328125,
|
| 701 |
+
"loss": 0.5034,
|
| 702 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 703 |
+
"rewards/chosen": -1.0203418731689453,
|
| 704 |
+
"rewards/margins": 0.8515297770500183,
|
| 705 |
+
"rewards/rejected": -1.8718715906143188,
|
| 706 |
+
"step": 390
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.8368200836820083,
|
| 710 |
+
"grad_norm": 22.777855221169585,
|
| 711 |
+
"learning_rate": 3.9507259776993954e-08,
|
| 712 |
+
"logits/chosen": 0.8675743341445923,
|
| 713 |
+
"logits/rejected": 2.180140972137451,
|
| 714 |
+
"logps/chosen": -412.081787109375,
|
| 715 |
+
"logps/rejected": -439.1665954589844,
|
| 716 |
+
"loss": 0.4966,
|
| 717 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 718 |
+
"rewards/chosen": -0.9012048840522766,
|
| 719 |
+
"rewards/margins": 0.8984438180923462,
|
| 720 |
+
"rewards/rejected": -1.799648642539978,
|
| 721 |
+
"step": 400
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.8368200836820083,
|
| 725 |
+
"eval_logits/chosen": 1.2642048597335815,
|
| 726 |
+
"eval_logits/rejected": 2.022581100463867,
|
| 727 |
+
"eval_logps/chosen": -361.9197998046875,
|
| 728 |
+
"eval_logps/rejected": -454.93212890625,
|
| 729 |
+
"eval_loss": 0.5065781474113464,
|
| 730 |
+
"eval_rewards/accuracies": 0.7578125,
|
| 731 |
+
"eval_rewards/chosen": -0.9928989410400391,
|
| 732 |
+
"eval_rewards/margins": 0.9298015236854553,
|
| 733 |
+
"eval_rewards/rejected": -1.92270028591156,
|
| 734 |
+
"eval_runtime": 103.4115,
|
| 735 |
+
"eval_samples_per_second": 19.34,
|
| 736 |
+
"eval_steps_per_second": 0.309,
|
| 737 |
+
"step": 400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.8577405857740585,
|
| 741 |
+
"grad_norm": 41.05467932326275,
|
| 742 |
+
"learning_rate": 3.022313472693447e-08,
|
| 743 |
+
"logits/chosen": 1.1478136777877808,
|
| 744 |
+
"logits/rejected": 2.6675543785095215,
|
| 745 |
+
"logps/chosen": -409.5799865722656,
|
| 746 |
+
"logps/rejected": -468.4205627441406,
|
| 747 |
+
"loss": 0.4807,
|
| 748 |
+
"rewards/accuracies": 0.768750011920929,
|
| 749 |
+
"rewards/chosen": -0.9972062110900879,
|
| 750 |
+
"rewards/margins": 1.0250569581985474,
|
| 751 |
+
"rewards/rejected": -2.0222630500793457,
|
| 752 |
+
"step": 410
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.8786610878661087,
|
| 756 |
+
"grad_norm": 29.37147689145298,
|
| 757 |
+
"learning_rate": 2.2111614344599684e-08,
|
| 758 |
+
"logits/chosen": 1.628156304359436,
|
| 759 |
+
"logits/rejected": 2.2244324684143066,
|
| 760 |
+
"logps/chosen": -386.01971435546875,
|
| 761 |
+
"logps/rejected": -460.0570373535156,
|
| 762 |
+
"loss": 0.5114,
|
| 763 |
+
"rewards/accuracies": 0.737500011920929,
|
| 764 |
+
"rewards/chosen": -1.121294617652893,
|
| 765 |
+
"rewards/margins": 0.8227565884590149,
|
| 766 |
+
"rewards/rejected": -1.9440511465072632,
|
| 767 |
+
"step": 420
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.899581589958159,
|
| 771 |
+
"grad_norm": 36.49027990223903,
|
| 772 |
+
"learning_rate": 1.521597710086439e-08,
|
| 773 |
+
"logits/chosen": 1.1463916301727295,
|
| 774 |
+
"logits/rejected": 2.26253342628479,
|
| 775 |
+
"logps/chosen": -402.85162353515625,
|
| 776 |
+
"logps/rejected": -490.9013671875,
|
| 777 |
+
"loss": 0.4991,
|
| 778 |
+
"rewards/accuracies": 0.8125,
|
| 779 |
+
"rewards/chosen": -1.1392345428466797,
|
| 780 |
+
"rewards/margins": 1.0718799829483032,
|
| 781 |
+
"rewards/rejected": -2.2111144065856934,
|
| 782 |
+
"step": 430
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.9205020920502092,
|
| 786 |
+
"grad_norm": 25.14386947872744,
|
| 787 |
+
"learning_rate": 9.57301420397924e-09,
|
| 788 |
+
"logits/chosen": 1.6257444620132446,
|
| 789 |
+
"logits/rejected": 2.745250701904297,
|
| 790 |
+
"logps/chosen": -375.26214599609375,
|
| 791 |
+
"logps/rejected": -446.528076171875,
|
| 792 |
+
"loss": 0.4996,
|
| 793 |
+
"rewards/accuracies": 0.762499988079071,
|
| 794 |
+
"rewards/chosen": -1.0888123512268066,
|
| 795 |
+
"rewards/margins": 0.9304329752922058,
|
| 796 |
+
"rewards/rejected": -2.019245147705078,
|
| 797 |
+
"step": 440
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.9414225941422594,
|
| 801 |
+
"grad_norm": 25.676457242831784,
|
| 802 |
+
"learning_rate": 5.212833302556258e-09,
|
| 803 |
+
"logits/chosen": 1.4634435176849365,
|
| 804 |
+
"logits/rejected": 1.8727718591690063,
|
| 805 |
+
"logps/chosen": -356.43170166015625,
|
| 806 |
+
"logps/rejected": -456.1715393066406,
|
| 807 |
+
"loss": 0.5198,
|
| 808 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 809 |
+
"rewards/chosen": -1.0714185237884521,
|
| 810 |
+
"rewards/margins": 0.8603706359863281,
|
| 811 |
+
"rewards/rejected": -1.9317893981933594,
|
| 812 |
+
"step": 450
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.9414225941422594,
|
| 816 |
+
"eval_logits/chosen": 1.3119983673095703,
|
| 817 |
+
"eval_logits/rejected": 2.1024117469787598,
|
| 818 |
+
"eval_logps/chosen": -363.6942138671875,
|
| 819 |
+
"eval_logps/rejected": -458.2783508300781,
|
| 820 |
+
"eval_loss": 0.505741536617279,
|
| 821 |
+
"eval_rewards/accuracies": 0.76171875,
|
| 822 |
+
"eval_rewards/chosen": -1.010642647743225,
|
| 823 |
+
"eval_rewards/margins": 0.9455199241638184,
|
| 824 |
+
"eval_rewards/rejected": -1.9561628103256226,
|
| 825 |
+
"eval_runtime": 101.6569,
|
| 826 |
+
"eval_samples_per_second": 19.674,
|
| 827 |
+
"eval_steps_per_second": 0.315,
|
| 828 |
+
"step": 450
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.9623430962343096,
|
| 832 |
+
"grad_norm": 26.320704397497966,
|
| 833 |
+
"learning_rate": 2.158697848236607e-09,
|
| 834 |
+
"logits/chosen": 1.4335782527923584,
|
| 835 |
+
"logits/rejected": 2.41444730758667,
|
| 836 |
+
"logps/chosen": -418.5895080566406,
|
| 837 |
+
"logps/rejected": -466.94525146484375,
|
| 838 |
+
"loss": 0.4754,
|
| 839 |
+
"rewards/accuracies": 0.768750011920929,
|
| 840 |
+
"rewards/chosen": -1.0504125356674194,
|
| 841 |
+
"rewards/margins": 0.8714386820793152,
|
| 842 |
+
"rewards/rejected": -1.921851396560669,
|
| 843 |
+
"step": 460
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.9832635983263598,
|
| 847 |
+
"grad_norm": 24.390801418724564,
|
| 848 |
+
"learning_rate": 4.269029751107489e-10,
|
| 849 |
+
"logits/chosen": 1.4916856288909912,
|
| 850 |
+
"logits/rejected": 2.8449251651763916,
|
| 851 |
+
"logps/chosen": -409.98480224609375,
|
| 852 |
+
"logps/rejected": -462.52374267578125,
|
| 853 |
+
"loss": 0.4863,
|
| 854 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 855 |
+
"rewards/chosen": -1.0590018033981323,
|
| 856 |
+
"rewards/margins": 0.990155816078186,
|
| 857 |
+
"rewards/rejected": -2.0491576194763184,
|
| 858 |
+
"step": 470
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 1.0,
|
| 862 |
+
"step": 478,
|
| 863 |
+
"total_flos": 0.0,
|
| 864 |
+
"train_loss": 0.5417051395113,
|
| 865 |
+
"train_runtime": 12463.9447,
|
| 866 |
+
"train_samples_per_second": 4.905,
|
| 867 |
+
"train_steps_per_second": 0.038
|
| 868 |
+
}
|
| 869 |
+
],
|
| 870 |
+
"logging_steps": 10,
|
| 871 |
+
"max_steps": 478,
|
| 872 |
+
"num_input_tokens_seen": 0,
|
| 873 |
+
"num_train_epochs": 1,
|
| 874 |
+
"save_steps": 100,
|
| 875 |
+
"stateful_callbacks": {
|
| 876 |
+
"TrainerControl": {
|
| 877 |
+
"args": {
|
| 878 |
+
"should_epoch_stop": false,
|
| 879 |
+
"should_evaluate": false,
|
| 880 |
+
"should_log": false,
|
| 881 |
+
"should_save": true,
|
| 882 |
+
"should_training_stop": true
|
| 883 |
+
},
|
| 884 |
+
"attributes": {}
|
| 885 |
+
}
|
| 886 |
+
},
|
| 887 |
+
"total_flos": 0.0,
|
| 888 |
+
"train_batch_size": 8,
|
| 889 |
+
"trial_name": null,
|
| 890 |
+
"trial_params": null
|
| 891 |
+
}
|