shawgpt-ft

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

• Loss: 0.4850

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: 0.0002
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.6301	0.9974	285	0.4880
0.4865	1.9983	571	0.4871
0.4846	2.9991	857	0.4851
0.4835	4.0	1143	0.4841
0.4845	4.9974	1428	0.4836
0.4823	5.9983	1714	0.4834
0.4816	6.9991	2000	0.4835
0.4809	8.0	2286	0.4831
0.4816	8.9974	2571	0.4835
0.4776	9.9738	2850	0.4850

Framework versions

• PEFT 0.12.0
• Transformers 4.42.4
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1