Spaces:
Running
Running
Initial commit: upload app.py, requirements.txt, and README.md
Browse files- README.md +5 -13
- app.py +41 -0
- requirements.txt +4 -0
README.md
CHANGED
@@ -1,13 +1,5 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
4 |
-
|
5 |
-
|
6 |
-
sdk: gradio
|
7 |
-
sdk_version: 5.36.2
|
8 |
-
app_file: app.py
|
9 |
-
pinned: false
|
10 |
-
license: other
|
11 |
-
---
|
12 |
-
|
13 |
-
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
|
|
1 |
+
# MiniGPT-4 Image + Text Chat (CPU-friendly)
|
2 |
+
|
3 |
+
Upload an image and provide a prompt to get a caption or description.
|
4 |
+
|
5 |
+
Powered by BLIP for CPU compatibility.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
app.py
ADDED
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# app.py
|
2 |
+
import os
|
3 |
+
import torch
|
4 |
+
import gradio as gr
|
5 |
+
from PIL import Image
|
6 |
+
from transformers import BlipProcessor, BlipForConditionalGeneration
|
7 |
+
|
8 |
+
# Placeholder for MiniGPT-4 (use real model if you have GPU)
|
9 |
+
class MiniGPT4:
|
10 |
+
def __init__(self):
|
11 |
+
self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
12 |
+
self.processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
|
13 |
+
self.model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base").to(self.device)
|
14 |
+
|
15 |
+
def chat(self, image, prompt=""):
|
16 |
+
if image is None:
|
17 |
+
return "Please upload an image."
|
18 |
+
inputs = self.processor(images=image, return_tensors="pt").to(self.device)
|
19 |
+
out = self.model.generate(**inputs, max_new_tokens=75)
|
20 |
+
caption = self.processor.decode(out[0], skip_special_tokens=True)
|
21 |
+
return f"{prompt}\n\nImage description: {caption}"
|
22 |
+
|
23 |
+
# Initialize model
|
24 |
+
minigpt4 = MiniGPT4()
|
25 |
+
|
26 |
+
def respond(image, prompt):
|
27 |
+
return minigpt4.chat(image, prompt)
|
28 |
+
|
29 |
+
demo = gr.Interface(
|
30 |
+
fn=respond,
|
31 |
+
inputs=[
|
32 |
+
gr.Image(type="pil", label="Upload an Image"),
|
33 |
+
gr.Textbox(lines=2, placeholder="Ask something about the image...", label="Your Prompt")
|
34 |
+
],
|
35 |
+
outputs=gr.Textbox(label="MiniGPT-4 Response"),
|
36 |
+
title="MiniGPT-4 Chat (Image + Text)",
|
37 |
+
description="Upload an image and ask a question. The model will respond based on image and prompt."
|
38 |
+
)
|
39 |
+
|
40 |
+
if __name__ == "__main__":
|
41 |
+
demo.launch()
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
torch
|
2 |
+
transformers
|
3 |
+
gradio
|
4 |
+
Pillow
|