metadata
library_name: transformers
base_model: /scratch/gpfs/jg9904/saved_models/Mistral-7B-Instruct-v0.3
tags:
- alignment-handbook
- generated_from_trainer
datasets:
- >-
/scratch/gpfs/jg9904/unintentional-unalignment/data_files/data-mistral-7b-instruct-sppo-iter1/90_variance
model-index:
- name: mistral-dpo-lr-5.0e-7-beta-0.01
results: []
mistral-dpo-lr-5.0e-7-beta-0.01
This model is a fine-tuned version of /scratch/gpfs/jg9904/saved_models/Mistral-7B-Instruct-v0.3 on the /scratch/gpfs/jg9904/unintentional-unalignment/data_files/data-mistral-7b-instruct-sppo-iter1/90_variance dataset. It achieves the following results on the evaluation set:
- Loss: 0.4177
- Rewards/chosen: -0.6130
- Rewards/rejected: -1.5691
- Rewards/accuracies: 0.8741
- Rewards/margins: 0.9561
- Logps/rejected: -497.8502
- Logps/chosen: -361.6783
- Logits/rejected All: -2.7429
- Logits/chosen All: -2.7344
- Logits/rejected Sum: 8021.8809
- Logits/chosen Sum: 8575.8633
- Logits/rejected Avg: 21.7435
- Logits/chosen Avg: 21.2499
- Gradient/inner Product: 1124073472.0
- Gradient/nabla Chosen Logps: 37120.0
- Gradient/nabla Rejected Logps: 45824.0
- Gradient/correlation: 0.4668
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-07
- train_batch_size: 8
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- total_train_batch_size: 64
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected All | Logits/chosen All | Logits/rejected Sum | Logits/chosen Sum | Logits/rejected Avg | Logits/chosen Avg | Gradient/inner Product | Gradient/nabla Chosen Logps | Gradient/nabla Rejected Logps | Gradient/correlation |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No log | 0 | 0 | 0.6931 | 0.0 | 0.0 | 0.0 | 0.0 | -340.9357 | -300.3766 | -2.8879 | -2.8885 | 7277.5859 | 7814.5264 | 19.8154 | 19.5578 | 77594624.0 | 16256.0 | 17152.0 | 0.2236 |
| 0.6346 | 0.3788 | 100 | 0.5216 | -0.5880 | -1.1850 | 0.7396 | 0.5970 | -459.4361 | -359.1805 | -2.8931 | -2.9041 | 7889.0952 | 8463.7900 | 21.4027 | 21.0064 | 562036736.0 | 34304.0 | 41984.0 | 0.4043 |
| 0.6253 | 0.7576 | 200 | 0.4177 | -0.6130 | -1.5691 | 0.8741 | 0.9561 | -497.8502 | -361.6783 | -2.7429 | -2.7344 | 8021.8809 | 8575.8633 | 21.7435 | 21.2499 | 1124073472.0 | 37120.0 | 45824.0 | 0.4668 |
Framework versions
- Transformers 4.45.0
- Pytorch 2.5.1+cu124
- Datasets 2.14.6
- Tokenizers 0.20.4