mistral_recipe1m_5500_e4

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: 1.1888

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: 3
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.3816	0.1081	20	1.3174
1.2441	0.2162	40	1.2527
1.2209	0.3243	60	1.2323
1.1979	0.4324	80	1.2186
1.2077	0.5405	100	1.2124
1.2124	0.6486	120	1.2048
1.1849	0.7568	140	1.1994
1.2012	0.8649	160	1.1927
1.2006	0.9730	180	1.1888

Framework versions

• PEFT 0.13.2
• Transformers 4.45.2
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0