shawgpt-ft

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.5866

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
4.586	0.9231	3	4.0840
4.5013	1.8462	6	3.6804
3.4295	2.7692	9	2.6498
1.7662	4.0	13	1.7049
1.6274	4.9231	16	1.4525
1.3983	5.8462	19	1.3957
1.2807	6.7692	22	1.3583
0.9365	8.0	26	1.3325
1.1829	8.9231	29	1.3339
1.0945	9.8462	32	1.3330
1.0448	10.7692	35	1.3496
0.7014	12.0	39	1.3455
0.8887	12.9231	42	1.3879
0.8145	13.8462	45	1.4225
0.7586	14.7692	48	1.4449
0.5382	16.0	52	1.5154
0.6563	16.9231	55	1.4966
0.6154	17.8462	58	1.5442
0.4083	18.4615	60	1.5866

Framework versions

• PEFT 0.13.2
• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1