File size: 5,061 Bytes
998b0a8 b6e525e 998b0a8 192eae5 9213095 3d5b038 9213095 192eae5 3d5b038 ef4866e b6e525e 3d5b038 64cfbfa b6e525e ef4866e 998b0a8 3d5b038 9213095 3d5b038 9213095 3d5b038 9213095 3d5b038 9213095 699de73 3d5b038 9213095 998b0a8 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 |
import gradio as gr
from huggingface_hub import InferenceClient
"""
Copied from inference in colab notebook
"""
from transformers import AutoTokenizer , AutoModelForCausalLM , TextIteratorStreamer
import torch
from threading import Thread
# Load model and tokenizer globally to avoid reloading for every request
model_path = "Heit39/llama_lora_model_1"
# Load tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=True, legacy=False)
# Load the base model (e.g., LLaMA)
base_model = AutoModelForCausalLM.from_pretrained("unsloth/Llama-3.2-3B-Instruct")
# Load LoRA adapter
from peft import PeftModel
model = PeftModel.from_pretrained(base_model, model_path)
# Define the response function
# def respond(
# message: str,
# history: list[tuple[str, str]],
# system_message: str,
# max_tokens: int,
# temperature: float,
# top_p: float,
# ):
# # Combine system message and history into a single prompt
# messages = [{"role": "system", "content": system_message}]
# for val in history:
# if val[0]:
# messages.append({"role": "user", "content": val[0]})
# if val[1]:
# messages.append({"role": "assistant", "content": val[1]})
# messages.append({"role": "user", "content": message})
# # Create a single text prompt from the messages
# prompt = ""
# for msg in messages:
# if msg["role"] == "system":
# prompt += f"[System]: {msg['content']}\n\n"
# elif msg["role"] == "user":
# prompt += f"[User]: {msg['content']}\n\n"
# elif msg["role"] == "assistant":
# prompt += f"[Assistant]: {msg['content']}\n\n"
# # Tokenize the prompt
# inputs = tokenizer(prompt, return_tensors="pt", truncation=True)
# input_ids = inputs.input_ids.to("cpu") # Ensure input is on the CPU
# # Generate response
# output_ids = model.generate(
# input_ids,
# max_length=input_ids.shape[1] + max_tokens,
# temperature=temperature,
# top_p=top_p,
# do_sample=True,
# )
# # Decode the generated text
# generated_text = tokenizer.decode(output_ids[0], skip_special_tokens=True)
# # Extract the assistant's response from the generated text
# assistant_response = generated_text[len(prompt):].strip()
# # Yield responses incrementally (simulate streaming)
# response = ""
# for token in assistant_response.split(): # Split tokens by whitespace
# response += token + " "
# yield response.strip()
def respond(
message: str,
history: list[tuple[str, str]],
system_message: str,
max_tokens: int,
temperature: float,
top_p: float,
):
# Combine system message and history into a single prompt
messages = [{"role": "system", "content": system_message}]
for val in history:
if val[0]:
messages.append({"role": "user", "content": val[0]})
if val[1]:
messages.append({"role": "assistant", "content": val[1]})
messages.append({"role": "user", "content": message})
# Create a single text prompt from the messages
prompt = ""
for msg in messages:
if msg["role"] == "system":
prompt += f"[System]: {msg['content']}\n\n"
elif msg["role"] == "user":
prompt += f"[User]: {msg['content']}\n\n"
elif msg["role"] == "assistant":
prompt += f"[Assistant]: {msg['content']}\n\n"
# Tokenize the prompt
inputs = tokenizer(prompt, return_tensors="pt", truncation=True)
input_ids = inputs.input_ids.to("cpu") # Ensure input is on the CPU
# Generate tokens incrementally
streamer = TextIteratorStreamer(tokenizer, skip_special_tokens=True)
generation_kwargs = {
"input_ids": input_ids,
"max_new_tokens": max_tokens,
"temperature": temperature,
"top_p": top_p,
"do_sample": True,
"streamer": streamer,
}
thread = Thread(target=model.generate, kwargs=generation_kwargs)
thread.start()
# Yield responses as they are generated
response = ""
for token in streamer:
response += token
if "[Chatbot]:" in response: # Only stream the part after "[Chatbot]:"
assistant_response = response.split("[Chatbot]:", 1)[1].strip()
yield assistant_response
"""
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
"""
demo = gr.ChatInterface(
respond,
additional_inputs=[
gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
gr.Slider(
minimum=0.1,
maximum=1.0,
value=0.95,
step=0.05,
label="Top-p (nucleus sampling)",
),
],
)
if __name__ == "__main__":
demo.launch()
|