from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base_model = "unsloth/Meta-Llama-3.1-8B"
lora_repo = "youth-ai-initiative/Med-instructor-Llama-by-Group-1"

tokenizer = AutoTokenizer.from_pretrained(base_model)

model = AutoModelForCausalLM.from_pretrained(
    base_model,
    device_map="auto",
    torch_dtype="auto",
)

model = PeftModel.from_pretrained(
    model,
    lora_repo,
    subfolder="model_weights/lora_weights"
)

prompt = "What are the common symptoms and risk factors of high blood pressure?"

inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(max_new_tokens=200, **inputs)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Training Hyperparameters

batch_size = 8
num_epochs = 2
learning_rate = 2e-4
LoRA Config:
- r = 32
- alpha = 64
- dropout = 0.05

Training Results

Epoch	Training Loss	Validation Loss
1	0.963223	0.927974
2	0.906526	0.920803