File size: 2,482 Bytes
892d0c8 1f4ce0a d297ae6 5e12f9b d297ae6 ec1a268 67054df 892d0c8 d297ae6 241af22 723250f a5fe88d 7e693ff d297ae6 7e693ff 241af22 d297ae6 241af22 bb8bac0 c3cad0a 892d0c8 d297ae6 c3cad0a d297ae6 c3cad0a d297ae6 c3cad0a d297ae6 c3cad0a d297ae6 c3cad0a d297ae6 c3cad0a d297ae6 c3cad0a d297ae6 c3cad0a d297ae6 c3cad0a 354d8fc a5a67ad 354d8fc c3cad0a d297ae6 |
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 |
import os
import shutil
import subprocess
# Clone and install dependencies
print("Cloning the repository...")
subprocess.run(["git", "clone", "https://huggingface.co/irotem98/edge_vlm"])
print("Installing dependencies...")
subprocess.run(["pip", "install", "-r", "edge_vlm/requirements.txt"])
subprocess.run(["pip", "install", "sentencepiece"])
# Copy all files from edge_vlm to current directory
print("Copying files...")
source_dir = "edge_vlm"
destination_dir = "."
import torch
import gradio as gr
for item in os.listdir(source_dir):
source_item = os.path.join(source_dir, item)
destination_item = os.path.join(destination_dir, item)
if os.path.isdir(source_item):
if os.path.exists(destination_item):
shutil.rmtree(destination_item)
shutil.copytree(source_item, destination_item)
else:
shutil.copy(source_item, destination_item)
print("Files copied successfully.")
# Now import the model from the copied files
from model import MoondreamModel
# Load the model and tokenizer
print("Loading model...")
model = MoondreamModel.load_model()
print("Model loaded.")
print("Loading tokenizer...")
tokenizer = MoondreamModel.load_tokenizer()
print("Tokenizer loaded.")
# Define the default question
default_question = "Describe the image."
# Function to handle image and return generated caption
def generate_caption_with_default(image):
print("Preprocessing image...")
preprocessed_image = MoondreamModel.preprocess_image(image)
print("Image preprocessed.")
print("Generating caption...")
caption = MoondreamModel.generate_caption(model, preprocessed_image, tokenizer)
print("Caption generated.")
return caption
# Create Gradio interface
print("Setting up Gradio interface...")
interface = gr.Interface(
fn=generate_caption_with_default,
inputs=gr.Image(type="pil", label="Upload an Image"),
outputs="text",
title="Image Caption Generator",
description=(
f"The default question is: '{default_question}'.\n\n"
"Please note that the inference may take up to 200 seconds due to long captions and CPU limitations.\n\n"
"[](https://huggingface.co/irotem98/edge_vlm) "
"[](https://github.com/rotem154154/edge_vlm)"
)
)
# Launch the interface
print("Launching interface...")
interface.launch()
|