shawgpt-ft6

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: 4.1405

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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
2.1354	1.0	1	4.2318
2.1268	2.0	2	4.2295
2.1189	3.0	3	4.2188
2.1264	4.0	4	4.2077
2.1174	5.0	5	4.1971
2.1101	6.0	6	4.1873
2.1162	7.0	7	4.1781
2.0936	8.0	8	4.1703
2.0862	9.0	9	4.1631
2.113	10.0	10	4.1567
2.0867	11.0	11	4.1517
2.0956	12.0	12	4.1470
2.0801	13.0	13	4.1438
2.092	14.0	14	4.1417
2.0841	15.0	15	4.1405

Framework versions

• PEFT 0.13.2
• Transformers 4.44.2
• Pytorch 2.5.0+cu124
• Datasets 3.0.2
• Tokenizers 0.19.1