bpp.py

import spaces
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline, TextIteratorStreamer

import os

# PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
os.environ['PYTORCH_CUDA_ALLOC_CONF'] = 'expandable_segments:True'

torch.random.manual_seed(0)

model = AutoModelForCausalLM.from_pretrained(
    "NyxKrage/Microsoft_Phi-4", 
    device_map="cuda", 
    torch_dtype="auto", 
    trust_remote_code=True, 
)
tokenizer = AutoTokenizer.from_pretrained("NyxKrage/Microsoft_Phi-4")

messages = [
    {"role": "system", "content": "You are a helpful AI assistant."},
    {"role": "user", "content": "Can you provide ways to eat combinations of bananas and dragonfruits?"},
    {"role": "assistant", "content": "Sure! Here are some ways to eat bananas and dragonfruits together: 1. Banana and dragonfruit smoothie: Blend bananas and dragonfruits together with some milk and honey. 2. Banana and dragonfruit salad: Mix sliced bananas and dragonfruits together with some lemon juice and honey."},
    {"role": "user", "content": "What about solving an 2x + 3 = 7 equation?"},
]

pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
)

streamer = TextIteratorStreamer(tokenizer)

generation_args = {
    "max_new_tokens": 500,
    "return_full_text": False,
    "temperature": 0.0,
    "do_sample": False,
    "streamer": streamer,
}

@spaces.GPU
def tuili():
    model.generate(messages, **generation_args)
    
tuili()
for new_text in streamer:
    print(new_text)