app.py

import torch
import streamlit as st
from transformers import AutoModelForCausalLM, AutoTokenizer

MODEL_ID = "mps/blue-scrub-150M"


st.set_page_config(
    page_title="Blue Scrub 150M",
    page_icon="🩺",
    layout="wide",
)


@st.cache_resource(show_spinner="Loading model from Hugging Face Hub...")
def load_model():
    tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
    model = AutoModelForCausalLM.from_pretrained(
        MODEL_ID,
        torch_dtype=torch.float32,
        low_cpu_mem_usage=True,
    )
    model.eval()
    return tokenizer, model


def generate_text(prompt: str, max_new_tokens: int, temperature: float, top_p: float) -> str:
    tokenizer, model = load_model()
    inputs = tokenizer(prompt, return_tensors="pt")

    do_sample = temperature > 0
    generation_kwargs = {
        "max_new_tokens": max_new_tokens,
        "do_sample": do_sample,
        "pad_token_id": tokenizer.eos_token_id,
        "eos_token_id": tokenizer.eos_token_id,
    }
    if do_sample:
        generation_kwargs.update({"temperature": temperature, "top_p": top_p})

    with torch.no_grad():
        output_ids = model.generate(**inputs, **generation_kwargs)

    generated_ids = output_ids[0][inputs["input_ids"].shape[-1]:]
    return tokenizer.decode(generated_ids, skip_special_tokens=True).strip()


st.title("🩺 Blue Scrub 150M Inference")
st.caption(f"Model: `{MODEL_ID}` · Streamlit CPU Space · non-instruction-tuned base model")

st.warning(
    "This is a base/non-instruction-tuned model. It completes text and is not a chat assistant. "
    "Do not use outputs as medical advice. Free CPU inference can be slow."
)

with st.sidebar:
    st.header("Generation settings")
    max_new_tokens = st.slider("Max new tokens", min_value=8, max_value=256, value=96, step=8)
    temperature = st.slider("Temperature", min_value=0.0, max_value=2.0, value=0.7, step=0.1)
    top_p = st.slider("Top-p", min_value=0.05, max_value=1.0, value=0.9, step=0.05)
    st.markdown("---")
    st.markdown("[Open model card](https://huggingface.co/mps/blue-scrub-150M)")

prompt = st.text_area(
    "Prompt",
    value="Medical evidence suggests that",
    height=180,
    help="Use continuation-style prompts because this is a base model, not an instruction model.",
)

if st.button("Generate", type="primary", disabled=not prompt.strip()):
    try:
        with st.spinner("Generating..."):
            text = generate_text(prompt, max_new_tokens, temperature, top_p)
        st.subheader("Generated continuation")
        st.write(text or "No text generated.")
    except Exception as exc:
        st.error(f"Inference failed: {exc}")