vm-llama3-lr4e-5-be5e-4-r8

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

• Loss: 0.1354
• Accuracy: 0.4164

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: 4e-05
• train_batch_size: 26
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 5
• total_train_batch_size: 1040
• total_eval_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.2619	5.0	500	0.1481	0.3158
0.2206	10.0	1000	0.1354	0.4080

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1