Spaces:
Build error
Build error
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
3 |
+
import torch
|
4 |
+
|
5 |
+
# Load the model and tokenizer
|
6 |
+
model_name = "mrcuddle/SD-Prompter-1B"
|
7 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
8 |
+
model = AutoModelForCausalLM.from_pretrained(model_name)
|
9 |
+
|
10 |
+
# Function to generate a response
|
11 |
+
def chat(message, history):
|
12 |
+
# Combine the message and history into a single input
|
13 |
+
input_text = " ".join([f"{user}: {msg}" for user, msg in history] + [f"User: {message}"])
|
14 |
+
inputs = tokenizer(input_text, return_tensors="pt")
|
15 |
+
|
16 |
+
# Generate a response
|
17 |
+
with torch.no_grad():
|
18 |
+
outputs = model.generate(inputs.input_ids, max_length=50, num_return_sequences=1)
|
19 |
+
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
20 |
+
|
21 |
+
# Extract only the new response part
|
22 |
+
response = response.replace(input_text, "").strip()
|
23 |
+
|
24 |
+
# Append the new message and response to the history
|
25 |
+
history.append(("User", message))
|
26 |
+
history.append(("Assistant", response))
|
27 |
+
|
28 |
+
return history, history
|
29 |
+
|
30 |
+
# Create the Gradio chat interface
|
31 |
+
iface = gr.ChatInterface(
|
32 |
+
fn=chat,
|
33 |
+
title="Llama3.2 1B Stable Diffusion Prompter",
|
34 |
+
description="Generate Stable Diffusion Prompt with Llama3.2"
|
35 |
+
)
|
36 |
+
|
37 |
+
# Launch the interface
|
38 |
+
iface.launch()
|