Mistral-7B_task-1_60-samples_config_2

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the GaetanMichelet/chat_60_ft_t1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.8341

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: 1
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss
1.3017	0.5714	1	1.2585
1.2501	1.7143	3	1.1216
1.1981	2.8571	5	1.0063
1.0426	4.0	7	0.9179
0.8584	4.5714	8	0.8992
0.719	5.7143	10	0.8421
0.5789	6.8571	12	0.8341
0.5789	12.5714	13	0.8661
0.4488	13.7143	15	0.9318
0.3191	14.8571	17	0.9741
0.221	16.0	19	1.0262
0.1566	16.5714	20	1.0657
0.101	17.7143	22	1.1823
0.0647	18.8571	24	1.2348

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.1.2+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1