jdqqjr/Mistral-7b-uncensored-sft-lora

Mistral-7b_5000_8epoch_lora

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.2 on the advbench_sft_malicious dataset. It achieves the following results on the evaluation set:

• Loss: 0.8734

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: 0.0001
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• total_eval_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 8.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.3101	2.9542	500	0.4940
0.0918	5.9084	1000	0.7312

Framework versions

• PEFT 0.11.1
• Transformers 4.41.2
• Pytorch 2.3.1+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1