How can I load the model after pre-training?

[MouadGhouti]

I know this might seem a bit intuitive to some of you guys, but this is genuinely my first time to ever mess with LLMs. I'm currently training a model and I was wondering how can I use the model for inference after training? I have PyTorch checkpoint files, however I have no idea how can I load the model when it is done. Also, how can I upload the model to huggingface to make it available for other people?

[huiqiao]

GPTConfig, GPT are defined in train_gpt2.py, I put them in a separate file gpt_class.py

import torch
from torch.nn import functional as F
from gpt_class import GPTConfig, GPT

device = "cpu"
if torch.cuda.is_available():
device = "cuda"

state_dict = torch.load('log/model_19072.pt')
config = state_dict['config']
model = GPT(config)
model.to(device)
model.load_state_dict(state_dict['model'])

model.eval()
torch.manual_seed(42)
torch.cuda.manual_seed(42)