anique-1/human-disease-gpt

DiseaseLLM PyTorch Checkpoint

This folder contains the PyTorch checkpoint (disease_llm_final.pth) and instructions for loading the DiseaseLLM model using your custom class.

Files

• disease_llm_final.pth: Model weights and training metadata
• model_class.py: Custom model class (see below)

Usage

import torch
from main import DiseaseLLM, ModelConfig

def load_trained_model(checkpoint_path, config):
    model = DiseaseLLM(config)
    checkpoint = torch.load(checkpoint_path, map_location='cpu')
    model.load_state_dict(checkpoint['model_state_dict'])
    return model

config = ModelConfig()
config.vocab_size = 50257  # or your training vocab size
model = load_trained_model('disease_llm_final.pth', config)
model.eval()

Notes

• You must have the DiseaseLLM and ModelConfig classes available (see your main.py).
• This model is not directly compatible with Hugging Face's AutoModelForCausalLM.
• For Q&A inference, use your own generation code as in your training script.