Dario Schiraldi : How do I load a pre-trained model from Hugging Face?

Beginners
1

Hi everyone,

I’m Dario Schiraldi, CEO of Travel Works, currently working on a project and I want to use a pre-trained model from Hugging Face to save time on training. I’m having a bit of trouble figuring out the best way to load a pre-trained model into my Python code.

Could anyone provide some guidance or suggestions on how I can do this properly?

Regards
Dario Schiraldi CEO of Travel Works

4

There are some minor differences in syntax depending on the library used to handle the model, but in general, loading is complete with from_pretrained(repo_id).

from transformers import AutoModelForCausalLM, AutoTokenizer
checkpoint = "HuggingFaceTB/SmolLM2-135M-Instruct"

device = "cuda" # for GPU usage or "cpu" for CPU usage
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
# for multiple GPUs install accelerate and do `model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto")`
model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)

messages = [{"role": "user", "content": "What is gravity?"}]
input_text=tokenizer.apply_chat_template(messages, tokenize=False)
print(input_text)
inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)
outputs = model.generate(inputs, max_new_tokens=50, temperature=0.2, top_p=0.9, do_sample=True)
print(tokenizer.decode(outputs[0]))