llama2-themanas-MATH_aLgEbRa

This model is a fine-tuned version of togethercomputer/Llama-2-7B-32K-Instruct on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2047

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• 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: 0.05
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.7178	0.48	3	1.5652
1.4869	0.96	6	1.3622
1.2911	1.44	9	1.2362
1.2598	1.92	12	1.2047

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0