mistral_instruct_generation

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

• Loss: 1.6654

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: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 2
• training_steps: 100

Training results

Training Loss	Epoch	Step	Validation Loss
1.9883	0.5882	10	1.8932
1.8095	1.1765	20	1.7633
1.6942	1.7647	30	1.7123
1.6099	2.3529	40	1.6814
1.5807	2.9412	50	1.6573
1.5032	3.5294	60	1.6472
1.4055	4.1176	70	1.6427
1.4101	4.7059	80	1.6471
1.3395	5.2941	90	1.6440
1.2556	5.8824	100	1.6654

Framework versions

• PEFT 0.11.1
• Transformers 4.42.4
• Pytorch 2.2.0
• Datasets 2.20.0
• Tokenizers 0.19.1