llama2-7b-chat-hf-mental-health

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

• Loss: 2.1739
• Rouge Scores: {'rouge1': 0.9084397097988796, 'rouge2': 0.8277762468999725, 'rougeL': 0.5895858205528516, 'rougeLsum': 0.9084397097988796}
• Bleu Scores: [0.8523376774091347, 0.8462340921367671, 0.8321568176562063, 0.8159562219906903]
• 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: 0.0002
• train_batch_size: 6
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge Scores	Bleu Scores	Gen Len
2.1953	1.0	1	2.2184	{'rouge1': 0.9199491740787802, 'rouge2': 0.8318809005083514, 'rougeL': 0.5854056997640226, 'rougeLsum': 0.9199491740787802}	[0.8715254358896543, 0.8659211417516524, 0.8510857228959539, 0.833791158779687]	2048.0
2.0231	2.0	2	2.1936	{'rouge1': 0.9096744612563045, 'rouge2': 0.8254608823259653, 'rougeL': 0.5872535534158643, 'rougeLsum': 0.9096744612563045}	[0.8550666362510422, 0.8491213000075178, 0.834566911154976, 0.8177171754641129]	2048.0
1.932	3.0	3	2.1790	{'rouge1': 0.9083394562821455, 'rouge2': 0.8271173984934778, 'rougeL': 0.5857825128581924, 'rougeLsum': 0.9083394562821455}	[0.8523376774091347, 0.8462340921367671, 0.8321568176562063, 0.8159562219906903]	2048.0
1.8681	4.0	4	2.1739	{'rouge1': 0.9084397097988796, 'rouge2': 0.8277762468999725, 'rougeL': 0.5895858205528516, 'rougeLsum': 0.9084397097988796}	[0.8523376774091347, 0.8462340921367671, 0.8321568176562063, 0.8159562219906903]	2048.0

Framework versions

• PEFT 0.9.1.dev0
• Transformers 4.39.0.dev0
• Pytorch 2.1.0+cu121
• Datasets 2.18.1.dev0
• Tokenizers 0.15.2