Computer Vision Benchmarks 2025

from transformers import AutoImageProcessor, AutoModelForImageClassification
from PIL import Image
import torch

# Load model (swap for EfficientNet, ConvNeXt, etc.)
processor = AutoImageProcessor.from_pretrained("google/vit-large-patch16-224")
model = AutoModelForImageClassification.from_pretrained("google/vit-large-patch16-224")

# Classify an image
image = Image.open("photo.jpg")
inputs = processor(images=image, return_tensors="pt")

with torch.no_grad():
    outputs = model(**inputs)

predicted_class_idx = outputs.logits.argmax(-1).item()
print(f"Predicted: {model.config.id2label[predicted_class_idx]}")

from ultralytics import YOLO

# Load YOLO11 (best real-time detector)
model = YOLO("yolo11x.pt")

# Run detection
results = model("image.jpg")

# Process results
for result in results:
    boxes = result.boxes  # Bounding boxes
    for box in boxes:
        cls = int(box.cls[0])
        conf = float(box.conf[0])
        xyxy = box.xyxy[0].tolist()  # [x1, y1, x2, y2]
        print(f"{model.names[cls]}: {conf:.2f} at {xyxy}")

# Or just save annotated image
results[0].save("output.jpg")

from transformers import SamModel, SamProcessor
from PIL import Image
import torch

# Load SAM 2 (best zero-shot segmenter)
processor = SamProcessor.from_pretrained("facebook/sam-vit-huge")
model = SamModel.from_pretrained("facebook/sam-vit-huge")

image = Image.open("photo.jpg")

# Segment with a point prompt (x, y coordinates)
input_points = [[[400, 300]]]  # Center of object to segment
inputs = processor(image, input_points=input_points, return_tensors="pt")

with torch.no_grad():
    outputs = model(**inputs)

masks = processor.image_processor.post_process_masks(
    outputs.pred_masks,
    inputs["original_sizes"],
    inputs["reshaped_input_sizes"]
)

# masks[0] contains the segmentation mask

import open_clip
import torch
from PIL import Image

# Load CLIP/SigLIP for image embeddings
model, _, preprocess = open_clip.create_model_and_transforms(
    'ViT-L-14', pretrained='laion2b_s32b_b82k'
)
model.eval()
tokenizer = open_clip.get_tokenizer('ViT-L-14')

# Embed an image
image = preprocess(Image.open('photo.jpg')).unsqueeze(0)
with torch.no_grad():
    image_features = model.encode_image(image)
    image_features /= image_features.norm(dim=-1, keepdim=True)

# Search with text
text = tokenizer(['a photo of a cat', 'a photo of a dog'])
with torch.no_grad():
    text_features = model.encode_text(text)
    text_features /= text_features.norm(dim=-1, keepdim=True)

# Cosine similarity
similarity = (100.0 * image_features @ text_features.T).softmax(dim=-1)
print(f"Cat: {similarity[0][0]:.1%}, Dog: {similarity[0][1]:.1%}")