buffettgpt-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: 1.2554

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: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.828	1.0	66	1.2278
1.0948	2.0	132	1.0691
0.9567	3.0	198	1.0193
0.8749	4.0	264	0.9978
0.8046	5.0	330	0.9876
0.7429	6.0	396	0.9978
0.6911	7.0	462	1.0185
0.6456	8.0	528	1.0300
0.6068	9.0	594	1.0749
0.5747	10.0	660	1.1074
0.5475	11.0	726	1.1306
0.5258	12.0	792	1.1687
0.505	13.0	858	1.2040
0.4917	14.0	924	1.2410
0.4829	15.0	990	1.2554

Framework versions

• PEFT 0.11.1
• Transformers 4.39.3
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2