This repo contains MedLLaMA_13B, which is LLaMA-13b finetuned with some Medical Corpus.
The model was trained with the following hyperparameters:
- Epochs: 5
- Batch size: 320
- Cutoff length: 2048
- Learning rate: 2e-5
The model can be loaded as follows:
import transformers
import torch
tokenizer = transformers.LlamaTokenizer.from_pretrained('chaoyi-wu/MedLLaMA_13B')
model = transformers.LlamaForCausalLM.from_pretrained('chaoyi-wu/MedLLaMA_13B')
sentence = 'Hello, doctor'
batch = tokenizer(
sentence,
return_tensors="pt",
add_special_tokens=False
)
with torch.no_grad():
generated = model.generate(inputs = batch["input_ids"], max_length=200, do_sample=True, top_k=50)
print('model predict: ',tokenizer.decode(generated[0]))