File size: 2,374 Bytes
985b94e f8b7a6c 985b94e 93a46f5 9d94d25 93a46f5 d8856f3 93a46f5 7f05911 d8856f3 9d94d25 93a46f5 7f05911 d8856f3 f8b7a6c 71235da 93a46f5 9d94d25 fc5bec5 9d94d25 93a46f5 985b94e 93a46f5 9c7b0c8 93a46f5 985b94e |
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 |
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "" # Prevent CUDA initialization outside ZeroGPU
import spaces # Import spaces first
import gradio as gr
from peft import AutoPeftModelForCausalLM
from transformers import AutoTokenizer
# Load the model and tokenizer globally
model = AutoPeftModelForCausalLM.from_pretrained("eforse01/lora_model").to("cuda") # Move model to CUDA
tokenizer = AutoTokenizer.from_pretrained("eforse01/lora_model")
@spaces.GPU(duration=120) # Decorate the function for ZeroGPU
def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, min_p):
# Construct messages for the chat template
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})
# Tokenize the input messages
inputs = tokenizer.apply_chat_template(
messages,
tokenize=True,
add_generation_prompt=True,
return_tensors="pt", # Return tensors for PyTorch
)
# Ensure input_ids is moved to the same device as the model
input_ids = inputs.to("cuda") # Move input_ids to CUDA
print("Input IDs shape:", input_ids.shape)
# Generate response
output = model.generate(
input_ids=input_ids, # Pass tensor explicitly as input_ids
max_new_tokens=max_tokens,
use_cache=True,
temperature=temperature,
min_p=min_p,
)
# Debug output
print("Generated Output Shape:", output.shape)
print("Generated Output:", output)
# Decode and format the response
response = tokenizer.decode(output[0], skip_special_tokens=True)
# Yield the response
yield response.split("assistant")[-1]
# Gradio Interface
demo = gr.ChatInterface(
respond,
additional_inputs=[
gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
gr.Slider(minimum=1, maximum=2048, value=2048, step=1, label="Max new tokens"),
gr.Slider(minimum=0.1, maximum=4.0, value=1.5, step=0.1, label="Temperature"),
gr.Slider(minimum=0.1, maximum=1.0, value=0.99, step=0.01, label="Min-p"),
],
)
if __name__ == "__main__":
demo.launch()
|