This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the GaetanMichelet/chat-60_ft_task-2 and the GaetanMichelet/chat-120_ft_task-2 datasets. It achieves the following results on the evaluation set:

Loss: 0.7006

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
1.1488	0.9091	5	1.0947
1.0165	2.0	11	0.9407
0.8171	2.9091	16	0.8598
0.7886	4.0	22	0.7892
0.7586	4.9091	27	0.7540
0.6893	6.0	33	0.7249
0.6344	6.9091	38	0.7066
0.5774	8.0	44	0.7006
0.5214	8.9091	49	0.7153
0.4418	10.0	55	0.7321
0.3485	10.9091	60	0.8033
0.2374	12.0	66	0.8848
0.1445	12.9091	71	1.0025
0.075	14.0	77	1.3091
0.0347	14.9091	82	1.3889

Framework versions

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

GaetanMichelet/Llama-31-8B_task-2_120-samples_config-2

Llama-31-8B_task-2_120-samples_config-2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

GaetanMichelet/
Llama-31-8B_task-2_120-samples_config-2