llama-2-13b-fp16-english-to-hinglish-translation

This model is a fine-tuned version of TheBloke/Llama-2-13B-fp16 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7183
• Rouge Scores: {'rouge1': 0.9211071676780294, 'rouge2': 0.8281733381382788, 'rougeL': 0.8654973199305598, 'rougeLsum': 0.921063419066345}
• Bleu Scores: [0.9430761623222732, 0.9291834498365306, 0.9116309260224186, 0.8930302852359673]
• Gen Len: 2048.0

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: 2
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge Scores	Bleu Scores	Gen Len
0.7634	1.0	500	0.7242	{'rouge1': 0.9226356216479343, 'rouge2': 0.8298878258164286, 'rougeL': 0.8653820285905314, 'rougeLsum': 0.9225028041599788}	[0.9438235001535462, 0.9299872780706577, 0.9124966958220146, 0.8939299935025186]	2048.0
0.5477	2.0	1000	0.7183	{'rouge1': 0.9211071676780294, 'rouge2': 0.8281733381382788, 'rougeL': 0.8654973199305598, 'rougeLsum': 0.921063419066345}	[0.9430761623222732, 0.9291834498365306, 0.9116309260224186, 0.8930302852359673]	2048.0

Framework versions

• PEFT 0.8.2
• Transformers 4.38.0.dev0
• Pytorch 2.1.0+cu118
• Datasets 2.16.2.dev0
• Tokenizers 0.15.1