Spaces:
Running
on
Zero
Running
on
Zero
File size: 1,867 Bytes
ced4fcf |
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 |
#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
Object pointing example using Rex Omni
"""
import matplotlib.pyplot as plt
import torch
from PIL import Image
from rex_omni import RexOmniVisualize, RexOmniWrapper
def main():
# Model path - replace with your actual model path
model_path = "IDEA-Research/Rex-Omni"
print("🚀 Initializing Rex Omni model...")
# Create wrapper with custom parameters
rex_model = RexOmniWrapper(
model_path=model_path,
backend="transformers", # Choose "transformers" or "vllm"
max_tokens=2048,
temperature=0.0,
top_p=0.05,
top_k=1,
repetition_penalty=1.05,
)
# Load image
image_path = "tutorials/visual_prompting_example/test_images/pigeons.jpeg" # Replace with your image path
image = Image.open(image_path).convert("RGB")
print(f"✅ Image loaded successfully!")
print(f"📏 Image size: {image.size}")
visual_prompts = [
[644, 1210, 842, 1361],
[1180, 1066, 1227, 1160], # Box 3: bottom region
]
print("🎯 Performing object pointing...")
results = rex_model.inference(
images=image,
task="visual_prompting",
visual_prompt_boxes=visual_prompts,
)
# Process results
result = results[0]
if result["success"]:
predictions = result["extracted_predictions"]
vis_image = RexOmniVisualize(
image=image,
predictions=predictions,
font_size=30,
draw_width=10,
show_labels=True,
)
# Save visualization
output_path = (
"tutorials/visual_prompting_example/test_images/pigeons_visualize.jpg"
)
vis_image.save(output_path)
else:
print(f"❌ Inference failed: {result['error']}")
if __name__ == "__main__":
main()
|