mistral_instruct_generation

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.1 on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 0.5057

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 0.03
• training_steps: 100

Training results

Training Loss	Epoch	Step	Validation Loss
0.771	0.0164	10	0.6439
0.5601	0.0328	20	0.5503
0.5247	0.0492	30	0.5304
0.5286	0.0656	40	0.5220
0.483	0.0820	50	0.5168
0.4794	0.0984	60	0.5133
0.4826	0.1148	70	0.5104
0.4665	0.1311	80	0.5084
0.4703	0.1475	90	0.5064
0.4899	0.1639	100	0.5057

Framework versions

• PEFT 0.10.1.dev0
• Transformers 4.41.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1