Llama-31-8B_task-1_120-samples_config-3

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct 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: 1.2659

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: 1e-05
train_batch_size: 1
eval_batch_size: 1
seed: 42
distributed_type: multi-GPU
gradient_accumulation_steps: 8
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 150

Training results

Training Loss	Epoch	Step	Validation Loss
1.9625	1.0	11	2.0994
2.1365	2.0	22	2.0816
2.1371	3.0	33	2.0467
2.0536	4.0	44	1.9862
1.8317	5.0	55	1.8956
1.7607	6.0	66	1.7668
1.6452	7.0	77	1.6453
1.548	8.0	88	1.5728
1.4631	9.0	99	1.5217
1.4126	10.0	110	1.4711
1.3079	11.0	121	1.4176
1.3012	12.0	132	1.3769
1.2575	13.0	143	1.3423
1.2537	14.0	154	1.3098
1.1994	15.0	165	1.2874
1.1054	16.0	176	1.2713
1.0452	17.0	187	1.2680
1.0716	18.0	198	1.2659
0.9207	19.0	209	1.2755
0.8712	20.0	220	1.2918
0.8179	21.0	231	1.3371
0.6485	22.0	242	1.3561
0.6958	23.0	253	1.4414
0.5845	24.0	264	1.5147
0.5274	25.0	275	1.5912