Limo_llama

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on the Limo dataset. It achieves the following results on the evaluation set:

• Loss: 0.7616

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: 8e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 16
• total_train_batch_size: 64
• total_eval_batch_size: 4
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
0.9047	1.0	12	0.9276
0.8453	2.0	24	0.8575
0.7935	3.0	36	0.8197
0.7744	4.0	48	0.7953
0.7254	5.0	60	0.7805
0.7332	6.0	72	0.7704
0.7149	7.0	84	0.7655
0.7298	8.0	96	0.7627
0.7228	9.0	108	0.7619
0.7103	10.0	120	0.7616

Framework versions

• PEFT 0.15.2
• Transformers 4.52.4
• Pytorch 2.8.0+cu129
• Datasets 3.6.0
• Tokenizers 0.21.4