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: 1.3909

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: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.7212	0.92	3	1.5434
1.4475	1.85	6	1.4012
1.3078	2.77	9	1.3524
0.9281	4.0	13	1.3281
1.2403	4.92	16	1.3273
1.2006	5.85	19	1.3225
1.126	6.77	22	1.3246
0.8129	8.0	26	1.3236
1.052	8.92	29	1.3281
1.0153	9.85	32	1.3262
0.9963	10.77	35	1.3315
0.7424	12.0	39	1.3389
0.937	12.92	42	1.3509
0.9203	13.85	45	1.3569
0.8923	14.77	48	1.3796
0.6622	16.0	52	1.3736
0.8592	16.92	55	1.3812
0.8636	17.85	58	1.3903
0.5605	18.46	60	1.3909

Framework versions

• PEFT 0.7.1
• Transformers 4.38.2
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.2