Llama2-7B-lora-r-32-generic-step-200-labels_40.0-full-precision-augmented

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.6047

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• training_steps: 200
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
5.7141	0.0366	20	5.7137
5.602	0.0731	40	5.4908
5.1745	0.1097	60	5.1042
4.6833	0.1463	80	4.6160
4.254	0.1828	100	4.1537
3.855	0.2194	120	3.8767
3.7238	0.2559	140	3.7225
3.6245	0.2925	160	3.6417
3.5803	0.3291	180	3.6102
3.5863	0.3656	200	3.6047

Framework versions

• PEFT 0.15.2
• Transformers 4.45.2
• Pytorch 2.5.0+cu121
• Datasets 3.2.0
• Tokenizers 0.20.3