Mistral-7B_task-1_120-samples_config-2_full

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the GaetanMichelet/chat-60_ft_task-1 and the GaetanMichelet/chat-120_ft_task-1 datasets. It achieves the following results on the evaluation set:

• Loss: 0.9301

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 16
• total_train_batch_size: 16
• 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
2.0764	0.9091	5	1.9517
1.7328	2.0	11	1.5258
1.2472	2.9091	16	1.1678
0.979	4.0	22	0.9953
0.909	4.9091	27	0.9437
0.7945	6.0	33	0.9338
0.6672	6.9091	38	0.9301
0.5517	8.0	44	0.9630
0.411	8.9091	49	1.0645
0.296	10.0	55	1.1561
0.2302	10.9091	60	1.3206
0.1752	12.0	66	1.3892
0.1498	12.9091	71	1.3749
0.1138	14.0	77	1.4865

Framework versions

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