#duplicated from https://huggingface.co/spaces/chheplo/DeepSeek-R1-Distill-Llama-8B
import gradio as gr
import os
import spaces
from transformers import GemmaTokenizer, AutoModelForCausalLM
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
from threading import Thread
# Set an environment variable
HF_TOKEN = os.environ.get("HF_TOKEN", None)
DESCRIPTION = '''
deepseek-ai/DeepSeek-R1-Distill-Llama-8B
'''
LICENSE = """
---
"""
PLACEHOLDER = """
DeepSeek-R1-Distill-Llama-8B
Ask me anything...
"""
css = """
h1 {
text-align: center;
display: block;
}
#duplicate-button {
margin: auto;
color: white;
background: #1565c0;
border-radius: 100vh;
}
"""
model_id = "AXCXEPT/phi-4-deepseek-R1K-RL-EZO"
model_id = "AXCXEPT/phi-4-open-R1-Distill-EZOv1"
# Load the tokenizer and model
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id) # to("cuda:0")
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids("<|eot_id|>")
]
@spaces.GPU(duration=120)
def chat_llama3_8b(message: str,
history: list,
temperature: float,
max_new_tokens: int
) -> str:
"""
Generate a streaming response using the llama3-8b model.
Args:
message (str): The input message.
history (list): The conversation history used by ChatInterface.
temperature (float): The temperature for generating the response.
max_new_tokens (int): The maximum number of new tokens to generate.
Returns:
str: The generated response.
"""
conversation = []
for user, assistant in history:
conversation.extend([{"role": "user", "content": user}, {"role": "assistant", "content": assistant}])
conversation.append({"role": "user", "content": message})
input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt").to(model.device)
streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
generate_kwargs = dict(
input_ids= input_ids,
streamer=streamer,
max_new_tokens=max_new_tokens,
do_sample=True,
temperature=temperature,
eos_token_id=terminators,
)
# This will enforce greedy generation (do_sample=False) when the temperature is passed 0, avoiding the crash.
if temperature == 0:
generate_kwargs['do_sample'] = False
t = Thread(target=model.generate, kwargs=generate_kwargs)
t.start()
outputs = []
for text in streamer:
outputs.append(text)
#print(outputs)
yield "".join(outputs)
# Gradio block
chatbot=gr.Chatbot(height=450, placeholder=PLACEHOLDER, label='Gradio ChatInterface')
with gr.Blocks(fill_height=True, css=css) as demo:
gr.Markdown(DESCRIPTION)
gr.ChatInterface(
fn=chat_llama3_8b,
chatbot=chatbot,
fill_height=True,
additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=False, render=False),
additional_inputs=[
gr.Slider(minimum=0,
maximum=1,
step=0.1,
value=0.5,
label="Temperature",
render=False),
gr.Slider(minimum=128,
maximum=4096,
step=1,
value=1024,
label="Max new tokens",
render=False ),
],
examples=[
['How to setup a human base on Mars? Give short answer.'],
['Explain theory of relativity to me like I’m 8 years old.'],
['What is 9,000 * 9,000?'],
['Write a pun-filled happy birthday message to my friend Alex.'],
['Justify why a penguin might make a good king of the jungle.']
],
cache_examples=False,
)
gr.Markdown(LICENSE)
if __name__ == "__main__":
demo.launch()
import spaces
import os
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
from transformers import TextIteratorStreamer
from threading import Thread
import gradio as gr
text_generator = None
is_hugging_face = True
model_id = "AXCXEPT/phi-4-deepseek-R1K-RL-EZO"
model_id = "AXCXEPT/phi-4-open-R1-Distill-EZOv1"
huggingface_token = os.getenv("HUGGINGFACE_TOKEN")
huggingface_token = None
device = "auto" # torch.device("cuda" if torch.cuda.is_available() else "cpu")
device = "cuda"
dtype = torch.bfloat16
dtype = torch.float16
if not huggingface_token:
pass
print("no HUGGINGFACE_TOKEN if you need set secret ")
#raise ValueError("HUGGINGFACE_TOKEN environment variable is not set")
tokenizer = AutoTokenizer.from_pretrained(model_id, token=huggingface_token)
print(model_id,device,dtype)
histories = []
#model = None
if not is_hugging_face:
model = AutoModelForCausalLM.from_pretrained(
model_id, token=huggingface_token ,torch_dtype=dtype,device_map=device
)
text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer,torch_dtype=dtype,device_map=device,stream=True ) #pipeline has not to(device)
if next(model.parameters()).is_cuda:
print("The model is on a GPU")
else:
print("The model is on a CPU")
#print(f"text_generator.device='{text_generator.device}")
if str(text_generator.device).strip() == 'cuda':
print("The pipeline is using a GPU")
else:
print("The pipeline is using a CPU")
print("initialized")
def generate_text(messages):
if is_hugging_face:#need everytime initialize for ZeroGPU
model = AutoModelForCausalLM.from_pretrained(
model_id, token=huggingface_token ,torch_dtype=dtype,device_map=device
)
model.to(device)
question = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
question = tokenizer(question, return_tensors="pt").to(device)
streamer = TextIteratorStreamer(tokenizer, skip_prompt=True)
generation_kwargs = dict(question, streamer=streamer, max_new_tokens=200)
thread = Thread(target=model.generate, kwargs=generation_kwargs)
generated_output = ""
thread.start()
for new_text in streamer:
generated_output += new_text
yield generated_output
generate_text.zerogpu = True
@spaces.GPU(duration=60)
def call_generate_text(message, history):
# history.append({"role": "user", "content": message})
#print(message)
#print(history)
messages = history+[{"role":"user","content":message}]
try:
for text in generate_text(messages):
yield text
except RuntimeError as e:
print(f"An unexpected error occurred: {e}")
yield ""
demo = gr.ChatInterface(call_generate_text,type="messages")
#if __name__ == "__main__":
demo.queue()
demo.launch()