LoLlama-3.2-1B-lora

This model is a fine-tuned version of meta-llama/Llama-3.2-1B on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6076

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: 2e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.8355	1.0	1093	0.8073
0.763	2.0	2186	0.7528
0.7265	3.0	3279	0.7224
0.7004	4.0	4372	0.6996
0.6773	5.0	5465	0.6827
0.661	6.0	6558	0.6691
0.6466	7.0	7651	0.6580
0.6339	8.0	8744	0.6482
0.625	9.0	9837	0.6406
0.6239	10.0	10930	0.6336
0.6087	11.0	12023	0.6279
0.6038	12.0	13116	0.6233
0.5987	13.0	14209	0.6197
0.5936	14.0	15302	0.6163
0.5933	15.0	16395	0.6135
0.5905	16.0	17488	0.6114
0.5834	17.0	18581	0.6098
0.5839	18.0	19674	0.6086
0.5831	19.0	20767	0.6077
0.5806	20.0	21860	0.6076

Framework versions

• PEFT 0.14.0
• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0