Initial commit

.dockerignore

@@ -0,0 +1,20 @@
+__pycache__
+*.pyc
+*.pyo
+*.pyd
+.Python
+env
+pip-log.txt
+pip-delete-this-directory.txt
+.tox
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.log
+.pytest_cache
+.env
+.venv
+.DS_Store

.gitattributes

@@ -0,0 +1,38 @@
+*.pdf filter=lfs diff=lfs merge=lfs -text
+*.txt filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+static/ia_gen_droits_auteur.pdf filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,2 @@
+__pycache__
+.DS_Store

Dockerfile

@@ -0,0 +1,37 @@
+FROM python:3.9-slim
+
+# WORKDIR /app
+
+# COPY requirements.txt .
+# RUN pip install --no-cache-dir -r requirements.txt
+
+# COPY . .
+
+# EXPOSE 8080
+
+# CMD ["python", "app.py"]
+
+
+FROM python:3.9-slim
+
+# Set the working directory inside the container
+WORKDIR /app
+
+# Install dependencies
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy the rest of the application code
+COPY . .
+
+# Expose the port your React app runs on
+EXPOSE 7860
+
+# Add ownership to the user
+RUN chown -R 1001:1001 /app
+
+# Change to the created user
+USER 1001
+
+# Command to run the application
+CMD ["python", "app.py"]

README.md

@@ -0,0 +1,14 @@
+---
+title: GenAI Attribution Simulator
+emoji: 🎁
+colorFrom: gray
+colorTo: yellow
+sdk: docker
+pinned: false
+license: apache-2.0
+short_description: A reward simulator for training data attribution
+---
+
+# Reward Simulator
+
+This repository contains the code for the reward simulator. It is a tool that can be used to simulate the reward that training data would receive from a model's generations. 

app.py

@@ -0,0 +1,285 @@
+# docker build -t reward-simulator .docker run -p 7860:7860 -v $(pwd)/data:/app/data reward-simulator
+
+from PIL import Image
+import numpy as np
+import io
+import faiss
+
+import requests
+import torch
+
+from request import get_ft, get_topk
+from flickrapi import FlickrAPI
+
+from flask import Flask, request, render_template, jsonify, send_from_directory
+app = Flask(__name__)
+
+PRESET_IMAGES = {
+    1: "static/1.webp",
+    2: "static/2.webp",
+    3: "static/3.webp"
+}
+
+# Add Flickr configuration
+FLICKR_API_KEY = '80ef21a6f7eb0984ea613c316a89ca69'
+FLICKR_API_SECRET = '4d0e8ce6734f4b3f'
+flickr = FlickrAPI(FLICKR_API_KEY, FLICKR_API_SECRET, format='parsed-json', store_token=False)
+
+def get_photo_id(url):
+    """Extract photo ID from Flickr URL"""
+    try:
+        return url.split('/')[-1].split('_')[0]
+    except:
+        return None
+
+def get_other_info(url):
+    """Get author information from Flickr"""
+    try:
+        photo_id = get_photo_id(url)
+        if photo_id:
+            photo_info = flickr.photos.getInfo(photo_id=photo_id)
+            license = photo_info['photo']['license']
+            owner = photo_info['photo']['owner']
+            flickr_url = f"https://www.flickr.com/photos/{owner.get('nsid', '')}/{photo_id}"
+            return {
+                'username': owner.get('username', ''),
+                'realname': owner.get('realname', ''),
+                'nsid': owner.get('nsid', ''),
+                'flickr_url': flickr_url,
+                'license': license
+            }
+    except:
+        pass
+    return {
+        'username': 'Unknown',
+        'realname': 'Unknown',
+        'nsid': '',
+        'flickr_url': '',
+        'license': 'Unknown'
+    }
+
+
+def load_model():
+    """Load DINOv2 model once and cache it"""
+    torch.hub.set_dir('static')
+    model = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14')
+    model.eval()
+    model.to(torch.device('cuda' if torch.cuda.is_available() else 'cpu'))
+    return model
+
+def load_index(index_path):
+    """Load FAISS index once and cache it"""
+    return faiss.read_index(index_path)
+
+def distance_to_similarity(distances, temp=1e-4):
+    """Convert distance to similarity"""
+    for ii in range(len(distances)):
+        contribs = distances[ii].max() - distances[ii]
+        contribs = contribs / temp
+        sum_contribs = np.exp(contribs).sum()
+        distances[ii] = np.exp(contribs) / sum_contribs
+    return distances
+
+def calculate_rewards(subscription, num_generations, author_share, ro_share, num_users_k, similarities, num_authors=1800):
+    """Calculate rewards based on user inputs and similarities"""
+    num_users = num_users_k * 1000
+    
+    # Monthly revenue allocated to authors
+    authors_monthly_revenue = subscription * num_users * (author_share / 100)
+    
+    rewards = []
+    for sim in similarities[0]:
+        # Attribution bonus based on similarity score and number of neighbors
+        attribution_bonus = sim * len(similarities[0])
+        
+        # Calculate monthly rewards
+        author_month_reward = (authors_monthly_revenue / num_authors) * attribution_bonus
+        ro_month_reward = author_month_reward / (author_share / 100) * (ro_share / 100)
+        
+        rewards.append({
+            'paid_per_month': f"{subscription:.0f}€",
+            'attribution': f"{sim*100:.0f}%",
+            'author_month_reward': f"{author_month_reward:.0f}€",
+            'ro_month_reward': f"{ro_month_reward:.0f}€"
+            # 'paid_per_month': f"{subscription:.0f}€",
+            # 'paid_per_gen': f"{paid_per_gen:.2f}€",
+            # 'aro_share': f"{aro_share:.2f}c€",
+            # 'attribution': f"{sim*100:.0f}%",
+            # 'training_data_reward': f"{training_data_reward:.2f}c€",
+            # 'author_month_reward': f"{author_month_reward:.0f}€",
+            # 'ro_month_reward': f"{ro_month_reward:.0f}€"
+        })
+    return rewards
+
+# Global variables for model and index
+model = None
+index = None
+urls = None
+
+def init_model():
+    global model, index, urls
+    model = load_model()
+    index = load_index("data/openimages_index.bin")
+    with open("data/openimages_urls.txt", "r") as f:
+        urls = f.readlines()
+
+@app.route('/')
+def home():
+    return render_template('index.html')
+
+@app.route('/static/<path:filename>')
+def serve_static(filename):
+    return send_from_directory('static', filename)
+
+@app.route('/select_preset/<int:preset_id>')
+def select_preset(preset_id):
+    if preset_id not in PRESET_IMAGES:
+        return jsonify({'error': 'Invalid preset ID'}), 400
+    
+    try:
+        image_path = PRESET_IMAGES[preset_id]
+        image = Image.open(image_path).convert('RGB')
+        
+        # Use default parameters for presets
+        params = {
+            'subscription': 12,
+            'num_generations': 60,
+            'author_share': 5,
+            'ro_share': 10,
+            'num_users_k': 500,
+            'num_neighbors': 10
+        }
+        
+        # Get features and search
+        features = get_ft(model, image)
+        distances, indices = get_topk(index, features, topk=params['num_neighbors'])
+        
+        # Collect valid results first
+        valid_results = []
+        valid_similarities = []
+        for i in range(params['num_neighbors']):
+            image_url = urls[indices[0][i]].strip()
+            try:
+                response = requests.head(image_url)
+                if response.status_code == 200:
+                    valid_results.append({
+                        'index': i,
+                        'url': image_url
+                    })
+                    valid_similarities.append(distances[0][i])
+            except requests.RequestException:
+                continue
+        
+        # Renormalize similarities for valid results
+        if valid_similarities:
+            similarities = distance_to_similarity(np.array([valid_similarities]), temp=1e-5)
+            
+            # Calculate rewards with renormalized similarities
+            rewards = calculate_rewards(
+                params['subscription'],
+                params['num_generations'],
+                params['author_share'],
+                params['ro_share'],
+                params['num_users_k'],
+                similarities
+            )
+            
+            # Build final results
+            results = []
+            for i, result in enumerate(valid_results):
+                other_info = get_other_info(result['url'])
+                results.append({
+                    'image_url': result['url'],
+                    'rewards': rewards[i],
+                    'other': other_info
+                })
+            
+            return jsonify({'results': results})
+        
+    except Exception as e:
+        return jsonify({'error': str(e)}), 500
+
+DEFAULT_PARAMS = {
+    'subscription': 12,
+    'num_generations': 60,
+    'author_share': 5,
+    'ro_share': 10,
+    'num_users_k': 500,
+    'num_neighbors': 8,
+    'num_authors': 1800
+}
+
+@app.route('/process', methods=['POST'])
+def process_image():
+    if 'image' not in request.files:
+        return jsonify({'error': 'No image provided'}), 400
+    
+    try:
+        image_file = request.files['image']
+        image = Image.open(io.BytesIO(image_file.read())).convert('RGB')
+        
+        # Use default parameters if none provided
+        params = DEFAULT_PARAMS.copy()
+        if request.form:
+            params.update({
+                'subscription': float(request.form.get('subscription', params['subscription'])),
+                'num_generations': int(request.form.get('num_generations', params['num_generations'])),
+                'author_share': float(request.form.get('author_share', params['author_share'])),
+                'ro_share': float(request.form.get('ro_share', params['ro_share'])),
+                'num_users_k': int(request.form.get('num_users_k', params['num_users_k'])),
+                'num_neighbors': int(request.form.get('num_neighbors', params['num_neighbors'])),
+                'num_authors': int(request.form.get('num_authors', DEFAULT_PARAMS['num_authors'])),
+            })
+        
+        # Process image
+        features = get_ft(model, image)
+        distances, indices = get_topk(index, features, topk=params['num_neighbors'])
+        
+        # Collect valid results first
+        valid_results = []
+        valid_similarities = []
+        for i in range(params['num_neighbors']):
+            image_url = urls[indices[0][i]].strip()
+            try:
+                response = requests.head(image_url)
+                if response.status_code == 200:
+                    valid_results.append({
+                        'index': i,
+                        'url': image_url
+                    })
+                    valid_similarities.append(distances[0][i])
+            except requests.RequestException:
+                continue
+        
+        # Renormalize similarities for valid results
+        if valid_similarities:
+            similarities = distance_to_similarity(np.array([valid_similarities]), temp=1e-5)
+            
+            # Calculate rewards with renormalized similarities
+            rewards = calculate_rewards(
+                params['subscription'],
+                params['num_generations'],
+                params['author_share'],
+                params['ro_share'],
+                params['num_users_k'],
+                similarities
+            )
+            
+            # Build final results
+            results = []
+            for i, result in enumerate(valid_results):
+                other_info = get_other_info(result['url'])
+                results.append({
+                    'image_url': result['url'],
+                    'rewards': rewards[i],
+                    'other': other_info
+                })
+            
+            return jsonify({'results': results})
+        
+    except Exception as e:
+        return jsonify({'error': str(e)}), 500
+
+if __name__ == '__main__':
+    init_model()
+    app.run(host='0.0.0.0', port=7860)

count_authors.py

@@ -0,0 +1,95 @@
+import tqdm
+from multiprocessing import Pool, cpu_count
+import signal
+import sys
+import time
+
+from flickrapi import FlickrAPI
+
+# Add Flickr configuration
+FLICKR_API_KEY = '80ef21a6f7eb0984ea613c316a89ca69'
+FLICKR_API_SECRET = '4d0e8ce6734f4b3f'
+flickr = FlickrAPI(FLICKR_API_KEY, FLICKR_API_SECRET, format='parsed-json', store_token=False)
+
+def get_photo_id(url):
+    """Extract photo ID from Flickr URL"""
+    try:
+        return url.split('/')[-1].split('_')[0]
+    except:
+        return None
+
+def get_other_info(url):
+    """Get author information from Flickr"""
+    try:
+        photo_id = get_photo_id(url)
+        if photo_id:
+            # wait for 0.1 second 
+            time.sleep(0.1)
+            photo_info = flickr.photos.getInfo(photo_id=photo_id)
+            license = photo_info['photo']['license']
+            owner = photo_info['photo']['owner']
+            flickr_url = f"https://www.flickr.com/photos/{owner.get('nsid', '')}/{photo_id}"
+            return {
+                'username': owner.get('username', ''),
+                'realname': owner.get('realname', ''),
+                'nsid': owner.get('nsid', ''),
+                'flickr_url': flickr_url,
+                'license': license
+            }
+    except:
+        pass
+    return {
+        'username': 'Unknown',
+        'realname': 'Unknown',
+        'nsid': '',
+        'flickr_url': '',
+        'license': 'Unknown'
+    }
+
+def init_worker():
+    """Initialize worker process to handle signals"""
+    signal.signal(signal.SIGINT, signal.SIG_IGN)
+
+def process_url(url):
+    try:
+        return get_other_info(url)
+    except Exception as e:
+        return {
+            'username': 'Error',
+            'realname': str(e),
+            'nsid': '',
+            'flickr_url': url,
+            'license': 'Unknown'
+        }
+
+def process_urls_in_chunks(urls, chunk_size=100000):
+    authors = []
+    with Pool(cpu_count(), initializer=init_worker) as pool:
+        try:
+            # Process URLs in chunks
+            for i in range(0, len(urls), chunk_size):
+                chunk = urls[i:i + chunk_size]
+                chunk_results = list(tqdm.tqdm(
+                    pool.imap(process_url, chunk),
+                    total=len(chunk),
+                    desc=f"Processing chunk {i//chunk_size + 1}"
+                ))
+                authors.extend(chunk_results)
+        except KeyboardInterrupt:
+            pool.terminate()
+            pool.join()
+            print("\nProcessing interrupted by user")
+            sys.exit(1)
+    return authors
+
+if __name__ == "__main__":
+    urls_file = "data/openimages_urls.txt"
+    with open(urls_file) as f:
+        urls = [url.strip() for url in f.readlines()][:100000]
+
+    authors = process_urls_in_chunks(urls)
+    
+    # Count unique authors
+    unique_authors = len(set([author['username'] for author in authors]))
+    print(f"unique_authors: {unique_authors}")
+    print(f"Number of unique authors: {unique_authors}")

data/openimages_index.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:190ffe0f10bee7782f1f99a3ec340da49b41a3402988df4e5d8b3f57ed2fbacb
+size 102102985

data/openimages_urls.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:071fa57fc38809e0d50c91bff77ebba6c9b4ce3aa3e1fa803f24de0b5411c93f
+size 519686095

data_prep.py

@@ -0,0 +1,39 @@
+import os
+import tarfile
+import shutil
+
+# Define the base directory (update this to your folder location)
+base_dir = "..."
+
+# Define the output directory
+output_dir = os.path.join(base_dir, "all_extracted")
+
+# Create the output directory if it doesn't exist
+os.makedirs(output_dir, exist_ok=True)
+
+# Iterate over the folders and files in the base directory
+for root, dirs, files in os.walk(base_dir):
+    for file in files:
+        if file.endswith(".tar"):
+            tar_path = os.path.join(root, file)
+            tar_name = os.path.basename(file)[:-4]
+            print(f"Extracting: {tar_path}")
+            
+            # Open the .tar file and extract its contents
+            with tarfile.open(tar_path) as tar:
+                # Extract to a temporary location
+                temp_dir = os.path.join(base_dir, "temp_extract")
+                os.makedirs(temp_dir, exist_ok=True)
+                tar.extractall(temp_dir)
+                
+                
+                # Move the extracted files to the output directory
+                for extracted_file in os.listdir(temp_dir):
+                    source_path = os.path.join(temp_dir, extracted_file)
+                    target_path = os.path.join(output_dir, f"{tar_name}_{extracted_file}")
+                    shutil.move(source_path, target_path)
+                
+                # Clean up the temporary directory
+                shutil.rmtree(temp_dir)
+
+print(f"All files extracted to: {output_dir}")

request.py

@@ -0,0 +1,56 @@
+from PIL import Image
+import faiss
+import numpy as np
+import torch
+from torchvision import transforms
+
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+transform = transforms.Compose([
+    transforms.ToTensor(),
+    transforms.Resize(256),
+    transforms.CenterCrop(224),
+    transforms.Normalize(
+        mean=(0.485, 0.456, 0.406), 
+        std=(0.229, 0.224, 0.225)
+    )
+])
+
+def get_ft(
+    extractor: torch.nn.Module,
+    img: Image.Image
+) -> np.ndarray:
+    img = transform(img)
+    ft = extractor(img.unsqueeze(0).to(device))
+    return ft.detach().cpu().numpy()
+
+def get_topk(
+    index: faiss.Index,
+    ft: np.ndarray,
+    topk: int = 10
+) -> tuple[np.ndarray, np.ndarray]:
+    """
+    Get top-k nearest neighbors from the index
+    Args:
+        index: Faiss index
+        ft: Input feature
+        topk: Number of nearest neighbors to return
+    Returns:
+        Tuple of (distances, indices) for top-k matches
+    """
+    # Search index for nearest neighbors
+    distances, indices = index.search(ft, topk)
+    return distances, indices
+ 
+
+# EXAMPLE:
+
+# image = Image.open('path/to/your/image.jpg')
+# image = transform(image)
+
+# extractor = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14')
+# extractor.eval()
+# extractor.to(device)
+
+# ft = get_ft(...)
+# indices, distances = ...

requirements.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e9cfce4c1d93ce8a9f05048c712cc8123d2961a2f8cab06f17c76aa15faa5ad
+size 74

static/RO_Summary.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab4cb5a27b5cf7ad13089c6655efd2d9dbb3786348485c400ef712c3f48e8de1
+size 4906779

static/Reg_Summary.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7943ae6d56647d14de48ef9b8d288adcd522fb7a7ba8baf8c9de214a071e9d29
+size 42656

static/checkpoints/dinov2_vits14_pretrain.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b938bf1bc15cd2ec0feacfe3a1bb553fe8ea9ca46a7e1d8d00217f29aef60cd9
+size 88283115

static/facebookresearch_dinov2_main/.github/workflows/lint.yaml

@@ -0,0 +1,38 @@
+name: Lint
+
+on:
+  push:
+    branches:
+      - main
+  pull_request:
+    branches:
+      - main
+
+jobs:
+  run-linters:
+    name: Run linters
+    runs-on: ubuntu-20.04
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: 3.9
+          cache: 'pip'
+          cache-dependency-path: '**/requirements*.txt'
+      - name: Install Python (development) dependencies
+        run: |
+          pip install -r requirements-dev.txt
+      - name: Run flake8
+        run: |
+          flake8
+      - name: Run black
+        if: always()
+        run: |
+          black --check dinov2
+      - name: Run pylint
+        if: always()
+        run: |
+          pylint --exit-zero dinov2

static/facebookresearch_dinov2_main/.gitignore

@@ -0,0 +1,11 @@
+build/
+dist/
+*.egg-info/
+**/__pycache__/
+
+**/.ipynb_checkpoints
+**/.ipynb_checkpoints/**
+
+*.swp
+
+.vscode/

static/facebookresearch_dinov2_main/CODE_OF_CONDUCT.md

@@ -0,0 +1,80 @@
+# Code of Conduct
+
+## Our Pledge
+
+In the interest of fostering an open and welcoming environment, we as
+contributors and maintainers pledge to make participation in our project and
+our community a harassment-free experience for everyone, regardless of age, body
+size, disability, ethnicity, sex characteristics, gender identity and expression,
+level of experience, education, socio-economic status, nationality, personal
+appearance, race, religion, or sexual identity and orientation.
+
+## Our Standards
+
+Examples of behavior that contributes to creating a positive environment
+include:
+
+* Using welcoming and inclusive language
+* Being respectful of differing viewpoints and experiences
+* Gracefully accepting constructive criticism
+* Focusing on what is best for the community
+* Showing empathy towards other community members
+
+Examples of unacceptable behavior by participants include:
+
+* The use of sexualized language or imagery and unwelcome sexual attention or
+advances
+* Trolling, insulting/derogatory comments, and personal or political attacks
+* Public or private harassment
+* Publishing others' private information, such as a physical or electronic
+address, without explicit permission
+* Other conduct which could reasonably be considered inappropriate in a
+professional setting
+
+## Our Responsibilities
+
+Project maintainers are responsible for clarifying the standards of acceptable
+behavior and are expected to take appropriate and fair corrective action in
+response to any instances of unacceptable behavior.
+
+Project maintainers have the right and responsibility to remove, edit, or
+reject comments, commits, code, wiki edits, issues, and other contributions
+that are not aligned to this Code of Conduct, or to ban temporarily or
+permanently any contributor for other behaviors that they deem inappropriate,
+threatening, offensive, or harmful.
+
+## Scope
+
+This Code of Conduct applies within all project spaces, and it also applies when
+an individual is representing the project or its community in public spaces.
+Examples of representing a project or community include using an official
+project e-mail address, posting via an official social media account, or acting
+as an appointed representative at an online or offline event. Representation of
+a project may be further defined and clarified by project maintainers.
+
+This Code of Conduct also applies outside the project spaces when there is a
+reasonable belief that an individual's behavior may have a negative impact on
+the project or its community.
+
+## Enforcement
+
+Instances of abusive, harassing, or otherwise unacceptable behavior may be
+reported by contacting the project team at <[email protected]>. All
+complaints will be reviewed and investigated and will result in a response that
+is deemed necessary and appropriate to the circumstances. The project team is
+obligated to maintain confidentiality with regard to the reporter of an incident.
+Further details of specific enforcement policies may be posted separately.
+
+Project maintainers who do not follow or enforce the Code of Conduct in good
+faith may face temporary or permanent repercussions as determined by other
+members of the project's leadership.
+
+## Attribution
+
+This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4,
+available at https://www.contributor-covenant.org/version/1/4/code-of-conduct.html
+
+[homepage]: https://www.contributor-covenant.org
+
+For answers to common questions about this code of conduct, see
+https://www.contributor-covenant.org/faq

static/facebookresearch_dinov2_main/CONTRIBUTING.md

@@ -0,0 +1,31 @@
+# Contributing to DINOv2
+We want to make contributing to this project as easy and transparent as
+possible.
+
+## Pull Requests
+We actively welcome your pull requests.
+
+1. Fork the repo and create your branch from `main`.
+2. If you've added code that should be tested, add tests.
+3. If you've changed APIs, update the documentation.
+4. Ensure the test suite passes.
+5. Make sure your code lints.
+6. If you haven't already, complete the Contributor License Agreement ("CLA").
+
+## Contributor License Agreement ("CLA")
+In order to accept your pull request, we need you to submit a CLA. You only need
+to do this once to work on any of Meta's open source projects.
+
+Complete your CLA here: <https://code.facebook.com/cla>
+
+## Issues
+We use GitHub issues to track public bugs. Please ensure your description is
+clear and has sufficient instructions to be able to reproduce the issue.
+
+Meta has a [bounty program](https://www.facebook.com/whitehat/) for the safe
+disclosure of security bugs. In those cases, please go through the process
+outlined on that page and do not file a public issue.
+
+## License
+By contributing to DINOv2, you agree that your contributions will be licensed
+under the LICENSE file in the root directory of this source tree.

static/facebookresearch_dinov2_main/LICENSE

@@ -0,0 +1,203 @@
+
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

static/facebookresearch_dinov2_main/MODEL_CARD.md

@@ -0,0 +1,272 @@
+# Model Card for DINOv2-S/B/L/g
+
+These are Vision Transformer models trained following the method described in the papers:
+"DINOv2: Learning Robust Visual Features without Supervision"
+and
+"Vision Transformers Need Registers".
+
+We provide 8 models: 
+- 1 ViT-g trained from scratch with 3 ViT-S/B/L models distilled from the ViT-g, without registers.
+- 1 ViT-g trained from scratch with 3 ViT-S/B/L models distilled from the ViT-g, with registers.
+
+## Model Details
+The model takes an image as input and returns a class token and patch tokens, and optionally 4 register tokens.
+
+The embedding dimension is: 
+- 384 for ViT-S.
+- 768 for ViT-B.
+- 1024 for ViT-L.
+- 1536 for ViT-g.
+
+The models follow a Transformer architecture, with a patch size of 14. In the case of registers, we add 4 register tokens, learned during training, to the input sequence after the patch embedding.
+
+For a 224x224 image, this results in 1 class token + 256 patch tokens, and optionally 4 register tokens.
+
+The models can accept larger images provided the image shapes are multiples of the patch size (14). 
+If this condition is not verified, the model will crop to the closest smaller multiple of the patch size.
+
+### Model Description
+
+- **Developed by:** Meta AI
+- **Model type:** Vision Transformer
+- **License:** Apache License 2.0
+
+- **Repository:** https://github.com/facebookresearch/dinov2
+- **Paper:** https://arxiv.org/abs/2304.07193
+- **Demo:** https://dinov2.metademolab.com/
+
+## Uses
+
+The models are vision backbones providing multi-purpose features for downstream tasks.
+
+### Direct Use
+
+The models can be used without fine-tuning, with downstream classifiers as simple as linear layers, to obtain competitive results:
+- on depth estimation, semantic segmentation, using linear layers.
+- on image classification, using k-NN classifiers on the class token.
+- on image classification, with logistic regression classifiers applied on the class token.
+- on image classification, with a linear layer applied on the class token and the average of the patch tokens.
+- on image retrieval using nearest neighbors.
+
+### Downstream Use
+
+It is technically possible to perform fine-tuning on the models, for small gains (we measured +2% on ImageNet-1k classification). 
+We recommend keeping this as a very last step and only when necessary, as the features already provide good performance out-of-the-box.
+
+## Bias, Risks, and Limitations
+
+Despite improvements thanks to the training method not using annotations, we still observe significant biases in our models toward rich households from Western countries.
+
+### Recommendations
+
+We expect fine-tuning will increase the biases in the features produced by the model as they will be tuned to the fine-tuning labels.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+```python
+import torch
+
+# DINOv2
+dinov2_vits14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14')
+dinov2_vitb14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14')
+dinov2_vitl14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14')
+dinov2_vitg14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14')
+
+# DINOv2 with registers
+dinov2_vits14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14_reg')
+dinov2_vitb14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14_reg')
+dinov2_vitl14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14_reg')
+dinov2_vitg14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14_reg')
+```
+
+## Training Details
+
+### Training Data
+
+- **Training data:** LVD-142M (see paper)
+- **Training regime:** fp16 using PyTorch-FSDP mixed-precision.
+
+### Training Procedure 
+
+- **Training objective:**
+  - DINO self-distillation loss with multi-crop
+  - iBOT masked-image modeling loss
+  - KoLeo regularization on [CLS] tokens
+- **Architectures:**
+  - ViT-S (21M params): Patch size 14, embedding dimension 384, 6 heads, MLP FFN
+  - ViT-B (86M params): Patch size 14, embedding dimension 768, 12 heads, MLP FFN
+  - ViT-L (0.3B params): Patch size 14, embedding dimension 1024, 16 heads, MLP FFN
+  - ViT-g (1.1B params): Patch size 14, embedding dimension 1536, 24 heads, SwiGLU FFN
+- **Distillation:**
+  - Distillation follows the standard DINOv2 pretraining procedure, except the teacher is a pretrained ViT-g, frozen.
+
+## Evaluation
+
+We refer users to the associated papers for the evaluation protocols.
+
+<table>
+  <tr>
+    <th colspan="2"></th>
+    <th colspan="3">ImageNet-1k</th>
+    <th>NYU-Depth v2</th>
+    <th>SUN-RGBD</th>
+    <th>ADE20k</th>
+    <th>iNaturalist 2018</th>
+    <th>Oxford-H</th>
+  </tr>
+  <tr>
+    <th rowspan="2">model</th>
+    <th rowspan="2">with <br /> registers</th>
+    <th>classif. (acc)</th>
+    <th>classif. (acc)</th>
+    <th>classif. V2 (acc)</th>
+    <th>depth (RMSE)</th>
+    <th>depth (RMSE)</th>
+    <th>segm. (mAP)</th>
+    <th>classif. (acc)</th>
+    <th>retrieval (mAP)</th>
+  </tr>
+  <tr>
+    <!-- <th>^</th> -->
+    <th>k-NN</th>
+    <th>linear</th>
+    <th>linear</th>
+    <th>linear<br />4 layers</th>
+    <th>NYU-D transfer</th>
+    <th>multiscale</th>
+    <th>linear</th>
+    <th>nearest neighbor</th>
+  </tr>
+  <tr>
+    <td>ViT-S/14</td>
+    <td align="center">:x:</td>
+    <td align="right">79.0%</td>
+    <td align="right">81.1%</td>
+    <td align="right">70.8%</td> 
+    <td align="right">0.417</td> 
+    <td align="right">0.431</td> 
+    <td align="right">47.2</td> 
+    <td align="right">69.5%</td> 
+    <td align="right">43.2</td> 
+  </tr>
+  <tr>
+    <td>ViT-S/14</td>
+    <td align="center">:white_check_mark:</td>
+    <td align="right">79.1%</td>
+    <td align="right">80.9%</td>
+    <td align="right">71.0%</td> 
+    <td align="right">N/A</td> 
+    <td align="right">N/A</td> 
+    <td align="right">N/A</td> 
+    <td align="right">67.6%</td> 
+    <td align="right">39.5</td> 
+  </tr>
+  <tr>
+    <td>ViT-B/14</td>
+    <td align="center">:x:</td>
+    <td align="right">82.1%</td>
+    <td align="right">84.5%</td>
+    <td align="right">74.9%</td>
+    <td align="right">0.362</td> 
+    <td align="right">0.400</td> 
+    <td align="right">51.3</td> 
+    <td align="right">76.3%</td> 
+    <td align="right">49.5</td> 
+  </tr>
+    <td>ViT-B/14</td>
+    <td align="center">:white_check_mark:</td>
+    <td align="right">82.0%</td>
+    <td align="right">84.6%</td>
+    <td align="right">75.6%</td>
+    <td align="right">N/A</td> 
+    <td align="right">N/A</td> 
+    <td align="right">N/A</td> 
+    <td align="right">73.8%</td> 
+    <td align="right">51.0</td> 
+  </tr>
+  <tr>
+    <td>ViT-L/14</td>
+    <td align="center">:x:</td>
+    <td align="right">83.5%</td>
+    <td align="right">86.3%</td>
+    <td align="right">77.6%</td>
+    <td align="right">0.333</td> 
+    <td align="right">0.396</td> 
+    <td align="right">53.1</td> 
+    <td align="right">79.8%</td> 
+    <td align="right">54.0</td> 
+  </tr>
+  <tr>
+    <td>ViT-L/14</td>
+    <td align="center">:white_check_mark:</td>
+    <td align="right">83.8%</td>
+    <td align="right">86.7%</td>
+    <td align="right">78.5%</td>
+    <td align="right">N/A</td> 
+    <td align="right">N/A</td> 
+    <td align="right">N/A</td> 
+    <td align="right">80.9%</td> 
+    <td align="right">55.7</td> 
+  </tr>
+  <tr>
+    <td>ViT-g/14</td>
+    <td align="center">:x:</td>
+    <td align="right">83.5%</td>
+    <td align="right">86.5%</td>
+    <td align="right">78.4%</td>
+    <td align="right">0.298</td> 
+    <td align="right">0.362</td> 
+    <td align="right">53.0</td> 
+    <td align="right">81.6%</td> 
+    <td align="right">52.3</td> 
+  </tr>
+  <tr>
+  <tr>
+    <td>ViT-g/14</td>
+    <td align="center">:white_check_mark:</td>
+    <td align="right">83.7%</td>
+    <td align="right">87.1%</td>
+    <td align="right">78.8%</td>
+    <td align="right">N/A</td> 
+    <td align="right">N/A</td> 
+    <td align="right">N/A</td> 
+    <td align="right">81.5%</td> 
+    <td align="right">58.2</td> 
+  </tr>
+</table>
+
+## Environmental Impact
+
+- **Hardware Type:** Nvidia A100
+- **Hours used:** 22,000 for ViT-g, 4,500 for ViT-S distillation, 5,300 for ViT-B distillation, 8,000 for ViT-L distillation
+- **Cloud Provider:** Private infra
+- **Compute Region:** USA
+- **Carbon Emitted:** 7t CO2eq
+
+#### Hardware
+
+Nvidia A100 GPUs
+
+#### Software
+
+PyTorch 2.0,
+xFormers 0.0.18
+
+**BibTeX**
+
+```
+@misc{oquab2023dinov2,
+  title={DINOv2: Learning Robust Visual Features without Supervision},
+  author={Oquab, Maxime and Darcet, Timothée and Moutakanni, Theo and Vo, Huy and Szafraniec, Marc and Khalidov, Vasil and Fernandez, Pierre and Haziza, Daniel and Massa, Francisco and El-Nouby, Alaaeldin and Howes, Russell and Huang, Po-Yao and Xu, Hu and Sharma, Vasu and Li, Shang-Wen and Galuba, Wojciech and Rabbat, Mike and Assran, Mido and Ballas, Nicolas and Synnaeve, Gabriel and Misra, Ishan and Jegou, Herve and Mairal, Julien and Labatut, Patrick and Joulin, Armand and Bojanowski, Piotr},
+  journal={arXiv:2304.07193},
+  year={2023}
+}
+@misc{darcet2023vitneedreg,
+  title={Vision Transformers Need Registers},
+  author={Darcet, Timothée and Oquab, Maxime and Mairal, Julien and Bojanowski, Piotr},
+  journal={arXiv:2309.16588},
+  year={2023}
+}
+```

static/facebookresearch_dinov2_main/README.md

@@ -0,0 +1,620 @@
+:new: [2023-10-26] *Added DINOv2 backbones with registers, following [Vision Transformers Need Registers](https://arxiv.org/abs/2309.16588).*
+
+# DINOv2: Learning Robust Visual Features without Supervision
+
+**[Meta AI Research, FAIR](https://ai.facebook.com/research/)**
+
+Maxime Oquab,
+Timothée Darcet,
+Théo Moutakanni,
+Huy V. Vo,
+Marc Szafraniec,
+Vasil Khalidov,
+Patrick Labatut,
+Armand Joulin,
+Piotr Bojanowski
+
+[[`Paper #1`](https://arxiv.org/abs/2304.07193)] [`Paper #2`](https://arxiv.org/abs/2309.16588)] [[`Blog`](https://ai.facebook.com/blog/dino-v2-computer-vision-self-supervised-learning/)] [[`Demo`](https://dinov2.metademolab.com)] [[`BibTeX`](#citing-dinov2)]
+
+PyTorch implementation and pretrained models for DINOv2. For details, see the papers: **[DINOv2: Learning Robust Visual Features without Supervision](https://arxiv.org/abs/2304.07193)** and **[Vision Transformers Need Registers](https://arxiv.org/abs/2309.16588)**.
+
+DINOv2 models produce high-performance visual features that can be directly employed with classifiers as simple as linear layers on a variety of computer vision tasks; these visual features are robust and perform well across domains without any requirement for fine-tuning. The models were pretrained on a dataset of 142 M images without using any labels or annotations.
+
+https://github.com/facebookresearch/dinov2/assets/60359573/f168823e-7922-415a-b429-578badf5c356
+
+<div align="center">
+  Visualization of the three first principal components of the patch features of all frames, mapped to RGB values.
+</div>
+
+## Pretrained models
+
+<table style="margin: auto">
+  <thead>
+    <tr>
+      <th>model</th>
+      <th># of<br />params</th>
+      <th>with<br />registers</th>
+      <th>ImageNet<br />k-NN</th>
+      <th>ImageNet<br />linear</th>
+      <th>download</th>
+    </tr>
+  </thead>
+  <tbody>
+    <tr>
+      <td>ViT-S/14 distilled</td>
+      <td align="right">21 M</td>
+      <td align="center">:x:</td>
+      <td align="right">79.0%</td>
+      <td align="right">81.1%</td>
+      <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_pretrain.pth">backbone only</a></td>
+    </tr>
+    <tr>
+      <td>ViT-S/14 distilled</td>
+      <td align="right">21 M</td>
+      <td align="center">:white_check_mark:</td>
+      <td align="right">79.1%</td>
+      <td align="right">80.9%</td>
+      <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_reg4_pretrain.pth">backbone only</a></td>
+    </tr>
+    <tr>
+      <td>ViT-B/14 distilled</td>
+      <td align="right">86 M</td>
+      <td align="center">:x:</td>
+      <td align="right">82.1%</td>
+      <td align="right">84.5%</td>
+      <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_pretrain.pth">backbone only</a></td>
+    </tr>
+    <tr>
+      <td>ViT-B/14 distilled</td>
+      <td align="right">86 M</td>
+      <td align="center">:white_check_mark:</td>
+      <td align="right">82.0%</td>
+      <td align="right">84.6%</td>
+      <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_reg4_pretrain.pth">backbone only</a></td>
+    </tr>
+    <tr>
+      <td>ViT-L/14 distilled</td>
+      <td align="right">300 M</td>
+      <td align="center">:x:</td>
+      <td align="right">83.5%</td>
+      <td align="right">86.3%</td>
+      <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_pretrain.pth">backbone only</a></td>
+    </tr>
+    <tr>
+      <td>ViT-L/14 distilled</td>
+      <td align="right">300 M</td>
+      <td align="center">:white_check_mark:</td>
+      <td align="right">83.8%</td>
+      <td align="right">86.7%</td>
+      <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_reg4_pretrain.pth">backbone only</a></td>
+    </tr>
+    <tr>
+      <td>ViT-g/14</td>
+      <td align="right">1,100 M</td>
+      <td align="center">:x:</td>
+      <td align="right">83.5%</td>
+      <td align="right">86.5%</td>
+      <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_pretrain.pth">backbone only</a></td>
+    </tr>
+    <tr>
+      <td>ViT-g/14</td>
+      <td align="right">1,100 M</td>
+      <td align="center">:white_check_mark:</td>
+      <td align="right">83.7%</td>
+      <td align="right">87.1%</td>
+      <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_reg4_pretrain.pth">backbone only</a></td>
+    </tr>
+  </tbody>
+</table>
+
+### Pretrained backbones (via PyTorch Hub)
+
+Please follow the instructions [here](https://pytorch.org/get-started/locally/) to install PyTorch (the only required dependency for loading the model). Installing PyTorch with CUDA support is strongly recommended.
+
+A corresponding [model card](MODEL_CARD.md) is included in the repository.
+
+```python
+import torch
+
+# DINOv2
+dinov2_vits14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14')
+dinov2_vitb14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14')
+dinov2_vitl14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14')
+dinov2_vitg14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14')
+
+# DINOv2 with registers
+dinov2_vits14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14_reg')
+dinov2_vitb14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14_reg')
+dinov2_vitl14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14_reg')
+dinov2_vitg14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14_reg')
+```
+
+### Pretrained heads - Image classification
+
+<table style="margin: auto">
+  <thead>
+    <tr>
+      <th rowspan="2">backbone</th>
+      <th rowspan="2">with<br />registers</th>
+      <th>download</th>
+    </tr>
+    <tr>
+      <th>ImageNet</th>
+    </tr>
+  </thead>
+  <tbody>
+    <tr>
+      <td>ViT-S/14 distilled</td>
+      <td align="center">:x:</td>
+      <td>
+        linear head (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_linear4_head.pth">4 layers</a>)
+      </td>
+    </tr>
+    <tr>
+      <td>ViT-S/14 distilled</td>
+      <td align="center">:white_check_mark:</td>
+      <td>
+        linear head (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_reg4_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_reg4_linear4_head.pth">4 layers</a>)
+      </td>
+    </tr>
+    <tr>
+      <td>ViT-B/14 distilled</td>
+      <td align="center">:x:</td>
+      <td>
+        linear head (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_linear4_head.pth">4 layers</a>)
+    </tr>
+    <tr>
+      <td>ViT-B/14 distilled</td>
+      <td align="center">:white_check_mark:</td>
+      <td>
+        linear head (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_reg4_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_reg4_linear4_head.pth">4 layers</a>)
+    </tr>
+    <tr>
+      <td>ViT-L/14 distilled</td>
+      <td align="center">:x:</td>
+      <td>
+        linear head (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_linear4_head.pth">4 layers</a>)
+    </tr>
+    <tr>
+      <td>ViT-L/14 distilled</td>
+      <td align="center">:white_check_mark:</td>
+      <td>
+        linear head (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_reg4_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_reg4_linear4_head.pth">4 layers</a>)
+    </tr>
+    <tr>
+      <td>ViT-g/14</td>
+      <td align="center">:x:</td>
+      <td>
+        linear head (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_linear4_head.pth">4 layers</a>)
+    </tr>
+    <tr>
+      <td>ViT-g/14</td>
+      <td align="center">:white_check_mark:</td>
+      <td>
+        linear head (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_lreg4_inear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_reg4_linear4_head.pth">4 layers</a>)
+    </tr>
+  </tbody>
+</table>
+
+The (full) classifier models can be loaded via PyTorch Hub:
+
+```python
+import torch
+
+# DINOv2
+dinov2_vits14_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14_lc')
+dinov2_vitb14_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14_lc')
+dinov2_vitl14_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14_lc')
+dinov2_vitg14_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14_lc')
+
+# DINOv2 with registers
+dinov2_vits14_reg_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14_reg_lc')
+dinov2_vitb14_reg_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14_reg_lc')
+dinov2_vitl14_reg_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14_reg_lc')
+dinov2_vitg14_reg_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14_reg_lc')
+```
+
+### Pretrained heads - Depth estimation
+
+<table style="margin: auto">
+  <thead>
+    <tr>
+      <th rowspan="2">backbone</th>
+      <th colspan="2">download head</th>
+    </tr>
+    <tr>
+      <th>NYUd</th>
+      <th>KITTI</th>
+    </tr>
+  </thead>
+  <tbody>
+    <tr>
+      <td>ViT-S/14 distilled</td>
+      <td>
+        linear (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_nyu_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_nyu_linear4_head.pth">4 layers</a>),
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_nyu_dpt_head.pth">DPT</a>
+      </td>
+      <td>
+        linear (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_kitti_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_kitti_linear4_head.pth">4 layers</a>),
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_kitti_dpt_head.pth">DPT</a>
+      </td>
+    </tr>
+    <tr>
+      <td>ViT-B/14 distilled</td>
+      <td>
+        linear (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_nyu_linear4_head.pth">4 layers</a>),
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_nyu_dpt_head.pth">DPT</a>
+      </td>
+      <td>
+        linear (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_kitti_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_kitti_linear4_head.pth">4 layers</a>),
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_kitti_dpt_head.pth">DPT</a>
+      </td>
+    </tr>
+    <tr>
+      <td>ViT-L/14 distilled</td>
+      <td>
+        linear (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_nyu_linear4_head.pth">4 layers</a>),
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_nyu_dpt_head.pth">DPT</a>
+      </td>
+      <td>
+        linear (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_kitti_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_kitti_linear4_head.pth">4 layers</a>),
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_kitti_dpt_head.pth">DPT</a>
+      </td>
+    </tr>
+    <tr>
+      <td>ViT-g/14</td>
+      <td>
+        linear (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_nyu_linear4_head.pth">4 layers</a>),
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_nyu_dpt_head.pth">DPT</a>
+      </td>
+      <td>
+        linear (<a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_kitti_linear_head.pth">1 layer</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_kitti_linear4_head.pth">4 layers</a>),
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_kitti_dpt_head.pth">DPT</a>
+      </td>
+    </tr>
+  </tbody>
+</table>
+
+### Pretrained heads - Semantic segmentation
+
+<table style="margin: auto">
+  <thead>
+    <tr>
+      <th rowspan="2">backbone</th>
+      <th>download model</th>
+      <th colspan="2">download head</th>
+    </tr>
+    <tr>
+      <th>ADE20K</th>
+      <th>ADE20K</th>
+      <th>VOC2012</th>
+    </tr>
+  </thead>
+  <tbody>
+    <tr>
+      <td>ViT-S/14 distilled</td>
+      <td></td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_ade20k_linear_head.pth">linear</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_ade20k_ms_head.pth">multi-scale</a>
+      </td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_voc2012_linear_head.pth">linear</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_voc2012_ms_head.pth">multi-scale</a>
+      </td>
+    </tr>
+    <tr>
+      <td>ViT-B/14 distilled</td>
+      <td></td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_ade20k_linear_head.pth">linear</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_ade20k_ms_head.pth">multi-scale</a>
+      </td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_voc2012_linear_head.pth">linear</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_voc2012_ms_head.pth">multi-scale</a>
+      </td>
+    </tr>
+    <tr>
+      <td>ViT-L/14 distilled</td>
+      <td></td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_ade20k_linear_head.pth">linear</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_ade20k_ms_head.pth">multi-scale</a>
+      </td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_voc2012_linear_head.pth">linear</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_voc2012_ms_head.pth">multi-scale</a>
+      </td>
+    </tr>
+    <tr>
+      <td>ViT-g/14</td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_ade20k_m2f.pth">Mask2Former</a>
+      </td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_ade20k_linear_head.pth">linear</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_ade20k_ms_head.pth">multi-scale</a>
+      </td>
+      <td>
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_voc2012_linear_head.pth">linear</a>,
+        <a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_voc2012_ms_head.pth">multi-scale</a>
+      </td>
+    </tr>
+  </tbody>
+</table>
+
+## Installation
+
+The training and evaluation code requires PyTorch 2.0 and [xFormers](https://github.com/facebookresearch/xformers) 0.0.18 as well as a number of other 3rd party packages. Note that the code has only been tested with the specified versions and also expects a Linux environment. To setup all the required dependencies for training and evaluation, please follow the instructions below:
+
+*[conda](https://docs.conda.io/projects/conda/en/latest/user-guide/getting-started.html)* **(Recommended)** - Clone the repository and then create and activate a `dinov2` conda environment using the provided environment definition:
+
+```shell
+conda env create -f conda.yaml
+conda activate dinov2
+```
+
+*[pip](https://pip.pypa.io/en/stable/getting-started/)* - Clone the repository and then use the provided `requirements.txt` to install the dependencies:
+
+```shell
+pip install -r requirements.txt
+```
+
+For dense tasks (depth estimation and semantic segmentation), there are additional dependencies (specific versions of `mmcv` and `mmsegmentation`) which are captured in the `extras` dependency specifications:
+
+*[conda](https://docs.conda.io/projects/conda/en/latest/user-guide/getting-started.html)* **(Recommended)**:
+
+```shell
+conda env create -f conda-extras.yaml
+conda activate dinov2-extras
+```
+
+*[pip](https://pip.pypa.io/en/stable/getting-started/)*:
+
+```shell
+pip install -r requirements.txt -r requirements-extras.txt
+```
+
+## Data preparation
+
+### ImageNet-1k
+
+The root directory of the dataset should hold the following contents:
+
+- `<ROOT>/test/ILSVRC2012_test_00000001.JPEG`
+- `<ROOT>/test/[..]`
+- `<ROOT>/test/ILSVRC2012_test_00100000.JPEG`
+- `<ROOT>/train/n01440764/n01440764_10026.JPEG`
+- `<ROOT>/train/[...]`
+- `<ROOT>/train/n15075141/n15075141_9993.JPEG`
+- `<ROOT>/val/n01440764/ILSVRC2012_val_00000293.JPEG`
+- `<ROOT>/val/[...]`
+- `<ROOT>/val/n15075141/ILSVRC2012_val_00049174.JPEG`
+- `<ROOT>/labels.txt`
+
+The provided dataset implementation expects a few additional metadata files to be present under the extra directory:
+
+- `<EXTRA>/class-ids-TRAIN.npy`
+- `<EXTRA>/class-ids-VAL.npy`
+- `<EXTRA>/class-names-TRAIN.npy`
+- `<EXTRA>/class-names-VAL.npy`
+- `<EXTRA>/entries-TEST.npy`
+- `<EXTRA>/entries-TRAIN.npy`
+- `<EXTRA>/entries-VAL.npy`
+
+These metadata files can be generated (once) with the following lines of Python code:
+
+```python
+from dinov2.data.datasets import ImageNet
+
+for split in ImageNet.Split:
+    dataset = ImageNet(split=split, root="<ROOT>", extra="<EXTRA>")
+    dataset.dump_extra()
+```
+
+Note that the root and extra directories do not have to be distinct directories.
+
+### ImageNet-22k
+
+Please adapt the [dataset class](dinov2/data/datasets/image_net_22k.py) to match your local setup.
+
+<br />
+
+:warning: To execute the commands provided in the next sections for training and evaluation, the `dinov2` package should be included in the Python module search path, i.e. simply prefix the command to run with `PYTHONPATH=.`.
+
+## Training
+
+### Fast setup: training DINOv2 ViT-L/16 on ImageNet-1k
+
+Run DINOv2 training on 4 A100-80GB nodes (32 GPUs) in a SLURM cluster environment with submitit:
+
+```shell
+python dinov2/run/train/train.py \
+    --nodes 4 \
+    --config-file dinov2/configs/train/vitl16_short.yaml \
+    --output-dir <PATH/TO/OUTPUT/DIR> \
+    train.dataset_path=ImageNet:split=TRAIN:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET>
+```
+
+Training time is approximately 1 day and the resulting checkpoint should reach 81.6% on k-NN eval and 82.9% on linear eval.
+
+The training code saves the weights of the teacher in the `eval` folder every 12500 iterations for evaluation.
+
+### Long setup: training DINOv2 ViT-L/14 on ImageNet-22k
+
+Run DINOv2 training on 12 A100-80GB nodes (96 GPUs) in a SLURM cluster environment with submitit:
+
+```shell
+python dinov2/run/train/train.py \
+    --nodes 12 \
+    --config-file dinov2/configs/train/vitl14.yaml \
+    --output-dir <PATH/TO/OUTPUT/DIR> \
+    train.dataset_path=ImageNet22k:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET>
+```
+
+Training time is approximately 3.3 days and the resulting checkpoint should reach 82.0% on k-NN eval and 84.5% on linear eval.
+
+The training code saves the weights of the teacher in the `eval` folder every 12500 iterations for evaluation.
+
+
+## Evaluation
+
+The training code regularly saves the teacher weights. In order to evaluate the model, run the following evaluation on a single node:
+
+### k-NN classification on ImageNet-1k
+
+```shell
+python dinov2/run/eval/knn.py \
+    --config-file <PATH/TO/OUTPUT/DIR>/config.yaml \
+    --pretrained-weights <PATH/TO/OUTPUT/DIR>/eval/training_24999/teacher_checkpoint.pth \
+    --output-dir <PATH/TO/OUTPUT/DIR>/eval/training_24999/knn \
+    --train-dataset ImageNet:split=TRAIN:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET> \
+    --val-dataset ImageNet:split=VAL:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET>
+```
+
+### Logistic regression classification on ImageNet-1k
+
+```shell
+python dinov2/run/eval/log_regression.py \
+    --config-file <PATH/TO/OUTPUT/DIR>/config.yaml \
+    --pretrained-weights <PATH/TO/OUTPUT/DIR>/eval/training_24999/teacher_checkpoint.pth \
+    --output-dir <PATH/TO/OUTPUT/DIR>/eval/training_24999/logreg \
+    --train-dataset ImageNet:split=TRAIN:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET> \
+    --val-dataset ImageNet:split=VAL:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET>
+```
+
+### Linear classification with data augmentation on ImageNet-1k
+
+```shell
+python dinov2/run/eval/linear.py \
+    --config-file <PATH/TO/OUTPUT/DIR>/config.yaml \
+    --pretrained-weights <PATH/TO/OUTPUT/DIR>/eval/training_24999/teacher_checkpoint.pth \
+    --output-dir <PATH/TO/OUTPUT/DIR>/eval/training_24999/linear \
+    --train-dataset ImageNet:split=TRAIN:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET> \
+    --val-dataset ImageNet:split=VAL:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET>
+```
+
+We release the weights from evaluating the different models:
+
+<table style="margin: auto">
+  <tr>
+    <th>model</th>
+    <th>with<br />registers</th>
+    <th>ImageNet<br />top-1</th>
+    <th>linear evaluation</th>
+  </tr>
+  <tr>
+    <td>ViT-S/14 distilled</td>
+    <td align="center">:x:</td>
+    <td align="right">81.1%</td>
+    <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_linear_head.pth">linear head weights</a></td>
+  </tr>
+  <tr>
+    <td>ViT-S/14 distilled</td>
+    <td align="center">:white_check_mark:</td>
+    <td align="right">80.8%</td>
+    <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_reg4_linear_head.pth">linear head weights</a></td>
+  </tr>
+  <tr>
+    <td>ViT-B/14 distilled</td>
+    <td align="center">:x:</td>
+    <td align="right">84.5%</td>
+    <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_linear_head.pth">linear head weights</a></td>
+  </tr>
+  <tr>
+    <td>ViT-B/14 distilled</td>
+    <td align="center">:white_check_mark:</td>
+    <td align="right">84.4%</td>
+    <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_reg4_linear_head.pth">linear head weights</a></td>
+  </tr>
+  <tr>
+    <td>ViT-L/14 distilled</td>
+    <td align="center">:x:</td>
+    <td align="right">86.3%</td>
+    <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_linear_head.pth">linear head weights</a></td>
+  </tr>
+  <tr>
+    <td>ViT-L/14 distilled</td>
+    <td align="center">:white_check_mark:</td>
+    <td align="right">86.5%</td>
+    <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_reg4_linear_head.pth">linear head weights</a></td>
+  </tr>
+  <tr>
+    <td>ViT-g/14</td>
+    <td align="center">:x:</td>
+    <td align="right">86.5%</td>
+    <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_linear_head.pth">linear head weights</a></td>
+  </tr>
+  <tr>
+    <td>ViT-g/14</td>
+    <td align="center">:white_check_mark:</td>
+    <td align="right">87.0%</td>
+    <td><a href="https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_reg4_linear_head.pth">linear head weights</a></td>
+  </tr>
+</table>
+
+The performance of the provided pretrained model weights can be evaluated as follows on ImageNet-1k:
+
+```shell
+python dinov2/run/eval/linear.py \
+    --config-file dinov2/configs/eval/vitg14_pretrain.yaml \
+    --pretrained-weights https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_pretrain.pth \
+    --train-dataset ImageNet:split=TRAIN:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET> \
+    --val-dataset ImageNet:split=VAL:root=<PATH/TO/DATASET>:extra=<PATH/TO/DATASET>
+```
+
+## Notebooks
+
+A few notebooks are provided to help the community leverage the models and code:
+
+<ul>
+  <li><a href="https://github.com/facebookresearch/dinov2/blob/main/notebooks/depth_estimation.ipynb">Depth estimation</a> - How to load and use the depth heads in combination with a matching backbone via mmcv</li>
+  <li><a href="https://github.com/facebookresearch/dinov2/blob/main/notebooks/semantic_segmentation.ipynb">Semantic segmentation</a> - How to load and use the segmentation heads in combination with a matching backbone via mmcv, and also how to load and use the Mask2Former-based segmentation model trained on ADE20K</li>
+</ul>
+
+## License
+
+DINOv2 code and model weights are released under the Apache License 2.0. See [LICENSE](LICENSE) for additional details.
+
+## Contributing
+
+See [contributing](CONTRIBUTING.md) and the [code of conduct](CODE_OF_CONDUCT.md).
+
+## Citing DINOv2
+
+If you find this repository useful, please consider giving a star :star: and citation :t-rex::
+
+```
+@misc{oquab2023dinov2,
+  title={DINOv2: Learning Robust Visual Features without Supervision},
+  author={Oquab, Maxime and Darcet, Timothée and Moutakanni, Theo and Vo, Huy V. and Szafraniec, Marc and Khalidov, Vasil and Fernandez, Pierre and Haziza, Daniel and Massa, Francisco and El-Nouby, Alaaeldin and Howes, Russell and Huang, Po-Yao and Xu, Hu and Sharma, Vasu and Li, Shang-Wen and Galuba, Wojciech and Rabbat, Mike and Assran, Mido and Ballas, Nicolas and Synnaeve, Gabriel and Misra, Ishan and Jegou, Herve and Mairal, Julien and Labatut, Patrick and Joulin, Armand and Bojanowski, Piotr},
+  journal={arXiv:2304.07193},
+  year={2023}
+}
+```
+
+```
+@misc{darcet2023vitneedreg,
+  title={Vision Transformers Need Registers},
+  author={Darcet, Timothée and Oquab, Maxime and Mairal, Julien and Bojanowski, Piotr},
+  journal={arXiv:2309.16588},
+  year={2023}
+}
+```

static/facebookresearch_dinov2_main/conda-extras.yaml

@@ -0,0 +1,24 @@
+name: dinov2-extras
+channels:
+  - defaults
+  - pytorch
+  - nvidia
+  - xformers
+  - conda-forge
+dependencies:
+  - python=3.9
+  - pytorch::pytorch=2.0.0
+  - pytorch::pytorch-cuda=11.7.0
+  - pytorch::torchvision=0.15.0
+  - omegaconf
+  - torchmetrics=0.10.3
+  - fvcore
+  - iopath
+  - xformers::xformers=0.0.18
+  - pip
+  - pip:
+    - git+https://github.com/facebookincubator/submitit
+    - --extra-index-url https://pypi.nvidia.com
+    - cuml-cu11
+    - mmcv-full==1.5.0
+    - mmsegmentation==0.27.0

static/facebookresearch_dinov2_main/conda.yaml

@@ -0,0 +1,22 @@
+name: dinov2
+channels:
+  - defaults
+  - pytorch
+  - nvidia
+  - xformers
+  - conda-forge
+dependencies:
+  - python=3.9
+  - pytorch::pytorch=2.0.0
+  - pytorch::pytorch-cuda=11.7.0
+  - pytorch::torchvision=0.15.0
+  - omegaconf
+  - torchmetrics=0.10.3
+  - fvcore
+  - iopath
+  - xformers::xformers=0.0.18
+  - pip
+  - pip:
+    - git+https://github.com/facebookincubator/submitit
+    - --extra-index-url https://pypi.nvidia.com
+    - cuml-cu11

static/facebookresearch_dinov2_main/dinov2/__init__.py

@@ -0,0 +1,6 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+__version__ = "0.0.1"

static/facebookresearch_dinov2_main/dinov2/configs/__init__.py

@@ -0,0 +1,22 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+import pathlib
+
+from omegaconf import OmegaConf
+
+
+def load_config(config_name: str):
+    config_filename = config_name + ".yaml"
+    return OmegaConf.load(pathlib.Path(__file__).parent.resolve() / config_filename)
+
+
+dinov2_default_config = load_config("ssl_default_config")
+
+
+def load_and_merge_config(config_name: str):
+    default_config = OmegaConf.create(dinov2_default_config)
+    loaded_config = load_config(config_name)
+    return OmegaConf.merge(default_config, loaded_config)

static/facebookresearch_dinov2_main/dinov2/configs/eval/vitb14_pretrain.yaml

@@ -0,0 +1,6 @@
+student:
+  arch: vit_base
+  patch_size: 14
+crops:
+  global_crops_size: 518  # this is to set up the position embeddings properly
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/eval/vitb14_reg4_pretrain.yaml

@@ -0,0 +1,9 @@
+student:
+  arch: vit_base
+  patch_size: 14
+  num_register_tokens: 4
+  interpolate_antialias: true
+  interpolate_offset: 0.0
+crops:
+  global_crops_size: 518  # this is to set up the position embeddings properly
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/eval/vitg14_pretrain.yaml

@@ -0,0 +1,7 @@
+student:
+  arch: vit_giant2
+  patch_size: 14
+  ffn_layer: swiglufused
+crops:
+  global_crops_size: 518  # this is to set up the position embeddings properly
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/eval/vitg14_reg4_pretrain.yaml

@@ -0,0 +1,10 @@
+student:
+  arch: vit_giant2
+  patch_size: 14
+  ffn_layer: swiglufused
+  num_register_tokens: 4
+  interpolate_antialias: true
+  interpolate_offset: 0.0
+crops:
+  global_crops_size: 518  # this is to set up the position embeddings properly
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/eval/vitl14_pretrain.yaml

@@ -0,0 +1,6 @@
+student:
+  arch: vit_large
+  patch_size: 14
+crops:
+  global_crops_size: 518  # this is to set up the position embeddings properly
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/eval/vitl14_reg4_pretrain.yaml

@@ -0,0 +1,9 @@
+student:
+  arch: vit_large
+  patch_size: 14
+  num_register_tokens: 4
+  interpolate_antialias: true
+  interpolate_offset: 0.0
+crops:
+  global_crops_size: 518  # this is to set up the position embeddings properly
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/eval/vits14_pretrain.yaml

@@ -0,0 +1,6 @@
+student:
+  arch: vit_small
+  patch_size: 14
+crops:
+  global_crops_size: 518  # this is to set up the position embeddings properly
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/eval/vits14_reg4_pretrain.yaml

@@ -0,0 +1,9 @@
+student:
+  arch: vit_small
+  patch_size: 14
+  num_register_tokens: 4
+  interpolate_antialias: true
+  interpolate_offset: 0.0
+crops:
+  global_crops_size: 518  # this is to set up the position embeddings properly
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/ssl_default_config.yaml

@@ -0,0 +1,118 @@
+MODEL:
+  WEIGHTS: ''
+compute_precision:
+  grad_scaler: true
+  teacher:
+    backbone:
+      sharding_strategy: SHARD_GRAD_OP
+      mixed_precision:
+        param_dtype: fp16
+        reduce_dtype: fp16
+        buffer_dtype: fp32
+    dino_head:
+      sharding_strategy: SHARD_GRAD_OP
+      mixed_precision:
+        param_dtype: fp16
+        reduce_dtype: fp16
+        buffer_dtype: fp32
+    ibot_head:
+      sharding_strategy: SHARD_GRAD_OP
+      mixed_precision:
+        param_dtype: fp16
+        reduce_dtype: fp16
+        buffer_dtype: fp32
+  student:
+    backbone:
+      sharding_strategy: SHARD_GRAD_OP
+      mixed_precision:
+        param_dtype: fp16
+        reduce_dtype: fp16
+        buffer_dtype: fp32
+    dino_head:
+      sharding_strategy: SHARD_GRAD_OP
+      mixed_precision:
+        param_dtype: fp16
+        reduce_dtype: fp32
+        buffer_dtype: fp32
+    ibot_head:
+      sharding_strategy: SHARD_GRAD_OP
+      mixed_precision:
+        param_dtype: fp16
+        reduce_dtype: fp32
+        buffer_dtype: fp32
+dino:
+  loss_weight: 1.0
+  head_n_prototypes: 65536
+  head_bottleneck_dim: 256
+  head_nlayers: 3
+  head_hidden_dim: 2048
+  koleo_loss_weight: 0.1
+ibot:
+  loss_weight: 1.0
+  mask_sample_probability: 0.5
+  mask_ratio_min_max:
+  - 0.1
+  - 0.5
+  separate_head: false
+  head_n_prototypes: 65536
+  head_bottleneck_dim: 256
+  head_nlayers: 3
+  head_hidden_dim: 2048
+train:
+  batch_size_per_gpu: 64
+  dataset_path: ImageNet:split=TRAIN
+  output_dir: .
+  saveckp_freq: 20
+  seed: 0
+  num_workers: 10
+  OFFICIAL_EPOCH_LENGTH: 1250
+  cache_dataset: true
+  centering: "centering" # or "sinkhorn_knopp"
+student:
+  arch: vit_large
+  patch_size: 16
+  drop_path_rate: 0.3
+  layerscale: 1.0e-05
+  drop_path_uniform: true
+  pretrained_weights: ''
+  ffn_layer: "mlp"
+  block_chunks: 0
+  qkv_bias: true
+  proj_bias: true
+  ffn_bias: true
+  num_register_tokens: 0
+  interpolate_antialias: false
+  interpolate_offset: 0.1
+teacher:
+  momentum_teacher: 0.992
+  final_momentum_teacher: 1
+  warmup_teacher_temp: 0.04
+  teacher_temp: 0.07
+  warmup_teacher_temp_epochs: 30
+optim:
+  epochs: 100
+  weight_decay: 0.04
+  weight_decay_end: 0.4
+  base_lr: 0.004  # learning rate for a batch size of 1024
+  lr: 0.  # will be set after applying scaling rule
+  warmup_epochs: 10
+  min_lr: 1.0e-06
+  clip_grad: 3.0
+  freeze_last_layer_epochs: 1
+  scaling_rule: sqrt_wrt_1024
+  patch_embed_lr_mult: 0.2
+  layerwise_decay: 0.9
+  adamw_beta1: 0.9
+  adamw_beta2: 0.999
+crops:
+  global_crops_scale:
+  - 0.32
+  - 1.0
+  local_crops_number: 8
+  local_crops_scale:
+  - 0.05
+  - 0.32
+  global_crops_size: 224
+  local_crops_size: 96
+evaluation:
+  eval_period_iterations: 12500

static/facebookresearch_dinov2_main/dinov2/configs/train/vitg14.yaml

@@ -0,0 +1,26 @@
+dino:
+  head_n_prototypes: 131072
+  head_bottleneck_dim: 384
+ibot:
+  separate_head: true
+  head_n_prototypes: 131072
+train:
+  batch_size_per_gpu: 12
+  dataset_path: ImageNet22k
+  centering: sinkhorn_knopp
+student:
+  arch: vit_giant2
+  patch_size: 14
+  drop_path_rate: 0.4
+  ffn_layer: swiglufused
+  block_chunks: 4
+teacher:
+  momentum_teacher: 0.994
+optim:
+  epochs: 500
+  weight_decay_end: 0.2
+  base_lr: 2.0e-04  # learning rate for a batch size of 1024
+  warmup_epochs: 80
+  layerwise_decay: 1.0
+crops:
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/train/vitl14.yaml

@@ -0,0 +1,26 @@
+dino:
+  head_n_prototypes: 131072
+  head_bottleneck_dim: 384
+ibot:
+  separate_head: true
+  head_n_prototypes: 131072
+train:
+  batch_size_per_gpu: 32
+  dataset_path: ImageNet22k
+  centering: sinkhorn_knopp
+student:
+  arch: vit_large
+  patch_size: 14
+  drop_path_rate: 0.4
+  ffn_layer: swiglufused
+  block_chunks: 4
+teacher:
+  momentum_teacher: 0.994
+optim:
+  epochs: 500
+  weight_decay_end: 0.2
+  base_lr: 2.0e-04  # learning rate for a batch size of 1024
+  warmup_epochs: 80
+  layerwise_decay: 1.0
+crops:
+  local_crops_size: 98

static/facebookresearch_dinov2_main/dinov2/configs/train/vitl16_short.yaml

@@ -0,0 +1,6 @@
+# this corresponds to the default config
+train:
+  dataset_path: ImageNet:split=TRAIN
+  batch_size_per_gpu: 64
+student:
+  block_chunks: 4

static/facebookresearch_dinov2_main/dinov2/data/__init__.py

@@ -0,0 +1,10 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+from .adapters import DatasetWithEnumeratedTargets
+from .loaders import make_data_loader, make_dataset, SamplerType
+from .collate import collate_data_and_cast
+from .masking import MaskingGenerator
+from .augmentations import DataAugmentationDINO

static/facebookresearch_dinov2_main/dinov2/data/adapters.py

@@ -0,0 +1,28 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+from typing import Any, Tuple
+
+from torch.utils.data import Dataset
+
+
+class DatasetWithEnumeratedTargets(Dataset):
+    def __init__(self, dataset):
+        self._dataset = dataset
+
+    def get_image_data(self, index: int) -> bytes:
+        return self._dataset.get_image_data(index)
+
+    def get_target(self, index: int) -> Tuple[Any, int]:
+        target = self._dataset.get_target(index)
+        return (index, target)
+
+    def __getitem__(self, index: int) -> Tuple[Any, Tuple[Any, int]]:
+        image, target = self._dataset[index]
+        target = index if target is None else target
+        return image, (index, target)
+
+    def __len__(self) -> int:
+        return len(self._dataset)

static/facebookresearch_dinov2_main/dinov2/data/augmentations.py

@@ -0,0 +1,118 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+import logging
+
+from torchvision import transforms
+
+from .transforms import (
+    GaussianBlur,
+    make_normalize_transform,
+)
+
+
+logger = logging.getLogger("dinov2")
+
+
+class DataAugmentationDINO(object):
+    def __init__(
+        self,
+        global_crops_scale,
+        local_crops_scale,
+        local_crops_number,
+        global_crops_size=224,
+        local_crops_size=96,
+    ):
+        self.global_crops_scale = global_crops_scale
+        self.local_crops_scale = local_crops_scale
+        self.local_crops_number = local_crops_number
+        self.global_crops_size = global_crops_size
+        self.local_crops_size = local_crops_size
+
+        logger.info("###################################")
+        logger.info("Using data augmentation parameters:")
+        logger.info(f"global_crops_scale: {global_crops_scale}")
+        logger.info(f"local_crops_scale: {local_crops_scale}")
+        logger.info(f"local_crops_number: {local_crops_number}")
+        logger.info(f"global_crops_size: {global_crops_size}")
+        logger.info(f"local_crops_size: {local_crops_size}")
+        logger.info("###################################")
+
+        # random resized crop and flip
+        self.geometric_augmentation_global = transforms.Compose(
+            [
+                transforms.RandomResizedCrop(
+                    global_crops_size, scale=global_crops_scale, interpolation=transforms.InterpolationMode.BICUBIC
+                ),
+                transforms.RandomHorizontalFlip(p=0.5),
+            ]
+        )
+
+        self.geometric_augmentation_local = transforms.Compose(
+            [
+                transforms.RandomResizedCrop(
+                    local_crops_size, scale=local_crops_scale, interpolation=transforms.InterpolationMode.BICUBIC
+                ),
+                transforms.RandomHorizontalFlip(p=0.5),
+            ]
+        )
+
+        # color distorsions / blurring
+        color_jittering = transforms.Compose(
+            [
+                transforms.RandomApply(
+                    [transforms.ColorJitter(brightness=0.4, contrast=0.4, saturation=0.2, hue=0.1)],
+                    p=0.8,
+                ),
+                transforms.RandomGrayscale(p=0.2),
+            ]
+        )
+
+        global_transfo1_extra = GaussianBlur(p=1.0)
+
+        global_transfo2_extra = transforms.Compose(
+            [
+                GaussianBlur(p=0.1),
+                transforms.RandomSolarize(threshold=128, p=0.2),
+            ]
+        )
+
+        local_transfo_extra = GaussianBlur(p=0.5)
+
+        # normalization
+        self.normalize = transforms.Compose(
+            [
+                transforms.ToTensor(),
+                make_normalize_transform(),
+            ]
+        )
+
+        self.global_transfo1 = transforms.Compose([color_jittering, global_transfo1_extra, self.normalize])
+        self.global_transfo2 = transforms.Compose([color_jittering, global_transfo2_extra, self.normalize])
+        self.local_transfo = transforms.Compose([color_jittering, local_transfo_extra, self.normalize])
+
+    def __call__(self, image):
+        output = {}
+
+        # global crops:
+        im1_base = self.geometric_augmentation_global(image)
+        global_crop_1 = self.global_transfo1(im1_base)
+
+        im2_base = self.geometric_augmentation_global(image)
+        global_crop_2 = self.global_transfo2(im2_base)
+
+        output["global_crops"] = [global_crop_1, global_crop_2]
+
+        # global crops for teacher:
+        output["global_crops_teacher"] = [global_crop_1, global_crop_2]
+
+        # local crops:
+        local_crops = [
+            self.local_transfo(self.geometric_augmentation_local(image)) for _ in range(self.local_crops_number)
+        ]
+        output["local_crops"] = local_crops
+        output["offsets"] = ()
+
+        return output

static/facebookresearch_dinov2_main/dinov2/data/collate.py

@@ -0,0 +1,49 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+import torch
+import random
+
+
+def collate_data_and_cast(samples_list, mask_ratio_tuple, mask_probability, dtype, n_tokens=None, mask_generator=None):
+    # dtype = torch.half  # TODO: Remove
+
+    n_global_crops = len(samples_list[0][0]["global_crops"])
+    n_local_crops = len(samples_list[0][0]["local_crops"])
+
+    collated_global_crops = torch.stack([s[0]["global_crops"][i] for i in range(n_global_crops) for s in samples_list])
+
+    collated_local_crops = torch.stack([s[0]["local_crops"][i] for i in range(n_local_crops) for s in samples_list])
+
+    B = len(collated_global_crops)
+    N = n_tokens
+    n_samples_masked = int(B * mask_probability)
+    probs = torch.linspace(*mask_ratio_tuple, n_samples_masked + 1)
+    upperbound = 0
+    masks_list = []
+    for i in range(0, n_samples_masked):
+        prob_min = probs[i]
+        prob_max = probs[i + 1]
+        masks_list.append(torch.BoolTensor(mask_generator(int(N * random.uniform(prob_min, prob_max)))))
+        upperbound += int(N * prob_max)
+    for i in range(n_samples_masked, B):
+        masks_list.append(torch.BoolTensor(mask_generator(0)))
+
+    random.shuffle(masks_list)
+
+    collated_masks = torch.stack(masks_list).flatten(1)
+    mask_indices_list = collated_masks.flatten().nonzero().flatten()
+
+    masks_weight = (1 / collated_masks.sum(-1).clamp(min=1.0)).unsqueeze(-1).expand_as(collated_masks)[collated_masks]
+
+    return {
+        "collated_global_crops": collated_global_crops.to(dtype),
+        "collated_local_crops": collated_local_crops.to(dtype),
+        "collated_masks": collated_masks,
+        "mask_indices_list": mask_indices_list,
+        "masks_weight": masks_weight,
+        "upperbound": upperbound,
+        "n_masked_patches": torch.full((1,), fill_value=mask_indices_list.shape[0], dtype=torch.long),
+    }

static/facebookresearch_dinov2_main/dinov2/data/datasets/__init__.py

@@ -0,0 +1,7 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+from .image_net import ImageNet
+from .image_net_22k import ImageNet22k

static/facebookresearch_dinov2_main/dinov2/data/datasets/decoders.py

@@ -0,0 +1,31 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+from io import BytesIO
+from typing import Any
+
+from PIL import Image
+
+
+class Decoder:
+    def decode(self) -> Any:
+        raise NotImplementedError
+
+
+class ImageDataDecoder(Decoder):
+    def __init__(self, image_data: bytes) -> None:
+        self._image_data = image_data
+
+    def decode(self) -> Image:
+        f = BytesIO(self._image_data)
+        return Image.open(f).convert(mode="RGB")
+
+
+class TargetDecoder(Decoder):
+    def __init__(self, target: Any):
+        self._target = target
+
+    def decode(self) -> Any:
+        return self._target

static/facebookresearch_dinov2_main/dinov2/data/datasets/extended.py

@@ -0,0 +1,38 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+from typing import Any, Tuple
+
+from torchvision.datasets import VisionDataset
+
+from .decoders import TargetDecoder, ImageDataDecoder
+
+
+class ExtendedVisionDataset(VisionDataset):
+    def __init__(self, *args, **kwargs) -> None:
+        super().__init__(*args, **kwargs)  # type: ignore
+
+    def get_image_data(self, index: int) -> bytes:
+        raise NotImplementedError
+
+    def get_target(self, index: int) -> Any:
+        raise NotImplementedError
+
+    def __getitem__(self, index: int) -> Tuple[Any, Any]:
+        try:
+            image_data = self.get_image_data(index)
+            image = ImageDataDecoder(image_data).decode()
+        except Exception as e:
+            raise RuntimeError(f"can not read image for sample {index}") from e
+        target = self.get_target(index)
+        target = TargetDecoder(target).decode()
+
+        if self.transforms is not None:
+            image, target = self.transforms(image, target)
+
+        return image, target
+
+    def __len__(self) -> int:
+        raise NotImplementedError

static/facebookresearch_dinov2_main/dinov2/data/datasets/image_net.py

@@ -0,0 +1,290 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+import csv
+from enum import Enum
+import logging
+import os
+from typing import Callable, List, Optional, Tuple, Union
+
+import numpy as np
+
+from .extended import ExtendedVisionDataset
+
+
+logger = logging.getLogger("dinov2")
+_Target = int
+
+
+class _Split(Enum):
+    TRAIN = "train"
+    VAL = "val"
+    TEST = "test"  # NOTE: torchvision does not support the test split
+
+    @property
+    def length(self) -> int:
+        split_lengths = {
+            _Split.TRAIN: 1_281_167,
+            _Split.VAL: 50_000,
+            _Split.TEST: 100_000,
+        }
+        return split_lengths[self]
+
+    def get_dirname(self, class_id: Optional[str] = None) -> str:
+        return self.value if class_id is None else os.path.join(self.value, class_id)
+
+    def get_image_relpath(self, actual_index: int, class_id: Optional[str] = None) -> str:
+        dirname = self.get_dirname(class_id)
+        if self == _Split.TRAIN:
+            basename = f"{class_id}_{actual_index}"
+        else:  # self in (_Split.VAL, _Split.TEST):
+            basename = f"ILSVRC2012_{self.value}_{actual_index:08d}"
+        return os.path.join(dirname, basename + ".JPEG")
+
+    def parse_image_relpath(self, image_relpath: str) -> Tuple[str, int]:
+        assert self != _Split.TEST
+        dirname, filename = os.path.split(image_relpath)
+        class_id = os.path.split(dirname)[-1]
+        basename, _ = os.path.splitext(filename)
+        actual_index = int(basename.split("_")[-1])
+        return class_id, actual_index
+
+
+class ImageNet(ExtendedVisionDataset):
+    Target = Union[_Target]
+    Split = Union[_Split]
+
+    def __init__(
+        self,
+        *,
+        split: "ImageNet.Split",
+        root: str,
+        extra: str,
+        transforms: Optional[Callable] = None,
+        transform: Optional[Callable] = None,
+        target_transform: Optional[Callable] = None,
+    ) -> None:
+        super().__init__(root, transforms, transform, target_transform)
+        self._extra_root = extra
+        self._split = split
+
+        self._entries = None
+        self._class_ids = None
+        self._class_names = None
+
+    @property
+    def split(self) -> "ImageNet.Split":
+        return self._split
+
+    def _get_extra_full_path(self, extra_path: str) -> str:
+        return os.path.join(self._extra_root, extra_path)
+
+    def _load_extra(self, extra_path: str) -> np.ndarray:
+        extra_full_path = self._get_extra_full_path(extra_path)
+        return np.load(extra_full_path, mmap_mode="r")
+
+    def _save_extra(self, extra_array: np.ndarray, extra_path: str) -> None:
+        extra_full_path = self._get_extra_full_path(extra_path)
+        os.makedirs(self._extra_root, exist_ok=True)
+        np.save(extra_full_path, extra_array)
+
+    @property
+    def _entries_path(self) -> str:
+        return f"entries-{self._split.value.upper()}.npy"
+
+    @property
+    def _class_ids_path(self) -> str:
+        return f"class-ids-{self._split.value.upper()}.npy"
+
+    @property
+    def _class_names_path(self) -> str:
+        return f"class-names-{self._split.value.upper()}.npy"
+
+    def _get_entries(self) -> np.ndarray:
+        if self._entries is None:
+            self._entries = self._load_extra(self._entries_path)
+        assert self._entries is not None
+        return self._entries
+
+    def _get_class_ids(self) -> np.ndarray:
+        if self._split == _Split.TEST:
+            assert False, "Class IDs are not available in TEST split"
+        if self._class_ids is None:
+            self._class_ids = self._load_extra(self._class_ids_path)
+        assert self._class_ids is not None
+        return self._class_ids
+
+    def _get_class_names(self) -> np.ndarray:
+        if self._split == _Split.TEST:
+            assert False, "Class names are not available in TEST split"
+        if self._class_names is None:
+            self._class_names = self._load_extra(self._class_names_path)
+        assert self._class_names is not None
+        return self._class_names
+
+    def find_class_id(self, class_index: int) -> str:
+        class_ids = self._get_class_ids()
+        return str(class_ids[class_index])
+
+    def find_class_name(self, class_index: int) -> str:
+        class_names = self._get_class_names()
+        return str(class_names[class_index])
+
+    def get_image_data(self, index: int) -> bytes:
+        entries = self._get_entries()
+        actual_index = entries[index]["actual_index"]
+
+        class_id = self.get_class_id(index)
+
+        image_relpath = self.split.get_image_relpath(actual_index, class_id)
+        image_full_path = os.path.join(self.root, image_relpath)
+        with open(image_full_path, mode="rb") as f:
+            image_data = f.read()
+        return image_data
+
+    def get_target(self, index: int) -> Optional[Target]:
+        entries = self._get_entries()
+        class_index = entries[index]["class_index"]
+        return None if self.split == _Split.TEST else int(class_index)
+
+    def get_targets(self) -> Optional[np.ndarray]:
+        entries = self._get_entries()
+        return None if self.split == _Split.TEST else entries["class_index"]
+
+    def get_class_id(self, index: int) -> Optional[str]:
+        entries = self._get_entries()
+        class_id = entries[index]["class_id"]
+        return None if self.split == _Split.TEST else str(class_id)
+
+    def get_class_name(self, index: int) -> Optional[str]:
+        entries = self._get_entries()
+        class_name = entries[index]["class_name"]
+        return None if self.split == _Split.TEST else str(class_name)
+
+    def __len__(self) -> int:
+        entries = self._get_entries()
+        assert len(entries) == self.split.length
+        return len(entries)
+
+    def _load_labels(self, labels_path: str) -> List[Tuple[str, str]]:
+        labels_full_path = os.path.join(self.root, labels_path)
+        labels = []
+
+        try:
+            with open(labels_full_path, "r") as f:
+                reader = csv.reader(f)
+                for row in reader:
+                    class_id, class_name = row
+                    labels.append((class_id, class_name))
+        except OSError as e:
+            raise RuntimeError(f'can not read labels file "{labels_full_path}"') from e
+
+        return labels
+
+    def _dump_entries(self) -> None:
+        split = self.split
+        if split == ImageNet.Split.TEST:
+            dataset = None
+            sample_count = split.length
+            max_class_id_length, max_class_name_length = 0, 0
+        else:
+            labels_path = "labels.txt"
+            logger.info(f'loading labels from "{labels_path}"')
+            labels = self._load_labels(labels_path)
+
+            # NOTE: Using torchvision ImageFolder for consistency
+            from torchvision.datasets import ImageFolder
+
+            dataset_root = os.path.join(self.root, split.get_dirname())
+            dataset = ImageFolder(dataset_root)
+            sample_count = len(dataset)
+            max_class_id_length, max_class_name_length = -1, -1
+            for sample in dataset.samples:
+                _, class_index = sample
+                class_id, class_name = labels[class_index]
+                max_class_id_length = max(len(class_id), max_class_id_length)
+                max_class_name_length = max(len(class_name), max_class_name_length)
+
+        dtype = np.dtype(
+            [
+                ("actual_index", "<u4"),
+                ("class_index", "<u4"),
+                ("class_id", f"U{max_class_id_length}"),
+                ("class_name", f"U{max_class_name_length}"),
+            ]
+        )
+        entries_array = np.empty(sample_count, dtype=dtype)
+
+        if split == ImageNet.Split.TEST:
+            old_percent = -1
+            for index in range(sample_count):
+                percent = 100 * (index + 1) // sample_count
+                if percent > old_percent:
+                    logger.info(f"creating entries: {percent}%")
+                    old_percent = percent
+
+                actual_index = index + 1
+                class_index = np.uint32(-1)
+                class_id, class_name = "", ""
+                entries_array[index] = (actual_index, class_index, class_id, class_name)
+        else:
+            class_names = {class_id: class_name for class_id, class_name in labels}
+
+            assert dataset
+            old_percent = -1
+            for index in range(sample_count):
+                percent = 100 * (index + 1) // sample_count
+                if percent > old_percent:
+                    logger.info(f"creating entries: {percent}%")
+                    old_percent = percent
+
+                image_full_path, class_index = dataset.samples[index]
+                image_relpath = os.path.relpath(image_full_path, self.root)
+                class_id, actual_index = split.parse_image_relpath(image_relpath)
+                class_name = class_names[class_id]
+                entries_array[index] = (actual_index, class_index, class_id, class_name)
+
+        logger.info(f'saving entries to "{self._entries_path}"')
+        self._save_extra(entries_array, self._entries_path)
+
+    def _dump_class_ids_and_names(self) -> None:
+        split = self.split
+        if split == ImageNet.Split.TEST:
+            return
+
+        entries_array = self._load_extra(self._entries_path)
+
+        max_class_id_length, max_class_name_length, max_class_index = -1, -1, -1
+        for entry in entries_array:
+            class_index, class_id, class_name = (
+                entry["class_index"],
+                entry["class_id"],
+                entry["class_name"],
+            )
+            max_class_index = max(int(class_index), max_class_index)
+            max_class_id_length = max(len(str(class_id)), max_class_id_length)
+            max_class_name_length = max(len(str(class_name)), max_class_name_length)
+
+        class_count = max_class_index + 1
+        class_ids_array = np.empty(class_count, dtype=f"U{max_class_id_length}")
+        class_names_array = np.empty(class_count, dtype=f"U{max_class_name_length}")
+        for entry in entries_array:
+            class_index, class_id, class_name = (
+                entry["class_index"],
+                entry["class_id"],
+                entry["class_name"],
+            )
+            class_ids_array[class_index] = class_id
+            class_names_array[class_index] = class_name
+
+        logger.info(f'saving class IDs to "{self._class_ids_path}"')
+        self._save_extra(class_ids_array, self._class_ids_path)
+
+        logger.info(f'saving class names to "{self._class_names_path}"')
+        self._save_extra(class_names_array, self._class_names_path)
+
+    def dump_extra(self) -> None:
+        self._dump_entries()
+        self._dump_class_ids_and_names()

static/facebookresearch_dinov2_main/dinov2/data/datasets/image_net_22k.py

@@ -0,0 +1,302 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+#
+# This source code is licensed under the Apache License, Version 2.0
+# found in the LICENSE file in the root directory of this source tree.
+
+from dataclasses import dataclass
+from enum import Enum
+from functools import lru_cache
+from gzip import GzipFile
+from io import BytesIO
+from mmap import ACCESS_READ, mmap
+import os
+from typing import Any, Callable, List, Optional, Set, Tuple
+import warnings
+
+import numpy as np
+
+from .extended import ExtendedVisionDataset
+
+
+_Labels = int
+
+_DEFAULT_MMAP_CACHE_SIZE = 16  # Warning: This can exhaust file descriptors
+
+
+@dataclass
+class _ClassEntry:
+    block_offset: int
+    maybe_filename: Optional[str] = None
+
+
+@dataclass
+class _Entry:
+    class_index: int  # noqa: E701
+    start_offset: int
+    end_offset: int
+    filename: str
+
+
+class _Split(Enum):
+    TRAIN = "train"
+    VAL = "val"
+
+    @property
+    def length(self) -> int:
+        return {
+            _Split.TRAIN: 11_797_647,
+            _Split.VAL: 561_050,
+        }[self]
+
+    def entries_path(self):
+        return f"imagenet21kp_{self.value}.txt"
+
+
+def _get_tarball_path(class_id: str) -> str:
+    return f"{class_id}.tar"
+
+
+def _make_mmap_tarball(tarballs_root: str, mmap_cache_size: int):
+    @lru_cache(maxsize=mmap_cache_size)
+    def _mmap_tarball(class_id: str) -> mmap:
+        tarball_path = _get_tarball_path(class_id)
+        tarball_full_path = os.path.join(tarballs_root, tarball_path)
+        with open(tarball_full_path) as f:
+            return mmap(fileno=f.fileno(), length=0, access=ACCESS_READ)
+
+    return _mmap_tarball
+
+
+class ImageNet22k(ExtendedVisionDataset):
+    _GZIPPED_INDICES: Set[int] = {
+        841_545,
+        1_304_131,
+        2_437_921,
+        2_672_079,
+        2_795_676,
+        2_969_786,
+        6_902_965,
+        6_903_550,
+        6_903_628,
+        7_432_557,
+        7_432_589,
+        7_813_809,
+        8_329_633,
+        10_296_990,
+        10_417_652,
+        10_492_265,
+        10_598_078,
+        10_782_398,
+        10_902_612,
+        11_203_736,
+        11_342_890,
+        11_397_596,
+        11_589_762,
+        11_705_103,
+        12_936_875,
+        13_289_782,
+    }
+    Labels = _Labels
+
+    def __init__(
+        self,
+        *,
+        root: str,
+        extra: str,
+        transforms: Optional[Callable] = None,
+        transform: Optional[Callable] = None,
+        target_transform: Optional[Callable] = None,
+        mmap_cache_size: int = _DEFAULT_MMAP_CACHE_SIZE,
+    ) -> None:
+        super().__init__(root, transforms, transform, target_transform)
+        self._extra_root = extra
+
+        entries_path = self._get_entries_path(root)
+        self._entries = self._load_extra(entries_path)
+
+        class_ids_path = self._get_class_ids_path(root)
+        self._class_ids = self._load_extra(class_ids_path)
+
+        self._gzipped_indices = ImageNet22k._GZIPPED_INDICES
+        self._mmap_tarball = _make_mmap_tarball(self._tarballs_root, mmap_cache_size)
+
+    def _get_entries_path(self, root: Optional[str] = None) -> str:
+        return "entries.npy"
+
+    def _get_class_ids_path(self, root: Optional[str] = None) -> str:
+        return "class-ids.npy"
+
+    def _find_class_ids(self, path: str) -> List[str]:
+        class_ids = []
+
+        with os.scandir(path) as entries:
+            for entry in entries:
+                root, ext = os.path.splitext(entry.name)
+                if ext != ".tar":
+                    continue
+                class_ids.append(root)
+
+        return sorted(class_ids)
+
+    def _load_entries_class_ids(self, root: Optional[str] = None) -> Tuple[List[_Entry], List[str]]:
+        root = self.get_root(root)
+        entries: List[_Entry] = []
+        class_ids = self._find_class_ids(root)
+
+        for class_index, class_id in enumerate(class_ids):
+            path = os.path.join(root, "blocks", f"{class_id}.log")
+            class_entries = []
+
+            try:
+                with open(path) as f:
+                    for line in f:
+                        line = line.rstrip()
+                        block, filename = line.split(":")
+                        block_offset = int(block[6:])
+                        filename = filename[1:]
+
+                        maybe_filename = None
+                        if filename != "** Block of NULs **":
+                            maybe_filename = filename
+                            _, ext = os.path.splitext(filename)
+                            # assert ext == ".JPEG"
+
+                        class_entry = _ClassEntry(block_offset, maybe_filename)
+                        class_entries.append(class_entry)
+            except OSError as e:
+                raise RuntimeError(f'can not read blocks file "{path}"') from e
+
+            assert class_entries[-1].maybe_filename is None
+
+            for class_entry1, class_entry2 in zip(class_entries, class_entries[1:]):
+                assert class_entry1.block_offset <= class_entry2.block_offset
+                start_offset = 512 * class_entry1.block_offset
+                end_offset = 512 * class_entry2.block_offset
+                assert class_entry1.maybe_filename is not None
+                filename = class_entry1.maybe_filename
+                entry = _Entry(class_index, start_offset, end_offset, filename)
+                # Skip invalid image files (PIL throws UnidentifiedImageError)
+                if filename == "n06470073_47249.JPEG":
+                    continue
+                entries.append(entry)
+
+        return entries, class_ids
+
+    def _load_extra(self, extra_path: str) -> np.ndarray:
+        extra_root = self._extra_root
+        extra_full_path = os.path.join(extra_root, extra_path)
+        return np.load(extra_full_path, mmap_mode="r")
+
+    def _save_extra(self, extra_array: np.ndarray, extra_path: str) -> None:
+        extra_root = self._extra_root
+        extra_full_path = os.path.join(extra_root, extra_path)
+        os.makedirs(extra_root, exist_ok=True)
+        np.save(extra_full_path, extra_array)
+
+    @property
+    def _tarballs_root(self) -> str:
+        return self.root
+
+    def find_class_id(self, class_index: int) -> str:
+        return str(self._class_ids[class_index])
+
+    def get_image_data(self, index: int) -> bytes:
+        entry = self._entries[index]
+        class_id = entry["class_id"]
+        class_mmap = self._mmap_tarball(class_id)
+
+        start_offset, end_offset = entry["start_offset"], entry["end_offset"]
+        try:
+            mapped_data = class_mmap[start_offset:end_offset]
+            data = mapped_data[512:]  # Skip entry header block
+
+            if len(data) >= 2 and tuple(data[:2]) == (0x1F, 0x8B):
+                assert index in self._gzipped_indices, f"unexpected gzip header for sample {index}"
+                with GzipFile(fileobj=BytesIO(data)) as g:
+                    data = g.read()
+        except Exception as e:
+            raise RuntimeError(f"can not retrieve image data for sample {index} " f'from "{class_id}" tarball') from e
+
+        return data
+
+    def get_target(self, index: int) -> Any:
+        return int(self._entries[index]["class_index"])
+
+    def get_targets(self) -> np.ndarray:
+        return self._entries["class_index"]
+
+    def get_class_id(self, index: int) -> str:
+        return str(self._entries[index]["class_id"])
+
+    def get_class_ids(self) -> np.ndarray:
+        return self._entries["class_id"]
+
+    def __getitem__(self, index: int) -> Tuple[Any, Any]:
+        with warnings.catch_warnings():
+            warnings.simplefilter("ignore")
+            return super().__getitem__(index)
+
+    def __len__(self) -> int:
+        return len(self._entries)
+
+    def _dump_entries(self, *args, **kwargs) -> None:
+        entries, class_ids = self._load_entries_class_ids(*args, **kwargs)
+
+        max_class_id_length, max_filename_length, max_class_index = -1, -1, -1
+        for entry in entries:
+            class_id = class_ids[entry.class_index]
+            max_class_index = max(entry.class_index, max_class_index)
+            max_class_id_length = max(len(class_id), max_class_id_length)
+            max_filename_length = max(len(entry.filename), max_filename_length)
+
+        dtype = np.dtype(
+            [
+                ("class_index", "<u4"),
+                ("class_id", f"U{max_class_id_length}"),
+                ("start_offset", "<u4"),
+                ("end_offset", "<u4"),
+                ("filename", f"U{max_filename_length}"),
+            ]
+        )
+        sample_count = len(entries)
+        entries_array = np.empty(sample_count, dtype=dtype)
+        for i, entry in enumerate(entries):
+            class_index = entry.class_index
+            class_id = class_ids[class_index]
+            start_offset = entry.start_offset
+            end_offset = entry.end_offset
+            filename = entry.filename
+            entries_array[i] = (
+                class_index,
+                class_id,
+                start_offset,
+                end_offset,
+                filename,
+            )
+
+        entries_path = self._get_entries_path(*args, **kwargs)
+        self._save_extra(entries_array, entries_path)
+
+    def _dump_class_ids(self, *args, **kwargs) -> None:
+        entries_path = self._get_entries_path(*args, **kwargs)
+        entries_array = self._load_extra(entries_path)
+
+        max_class_id_length, max_class_index = -1, -1
+        for entry in entries_array:
+            class_index, class_id = entry["class_index"], entry["class_id"]
+            max_class_index = max(int(class_index), max_class_index)
+            max_class_id_length = max(len(str(class_id)), max_class_id_length)
+
+        class_ids_array = np.empty(max_class_index + 1, dtype=f"U{max_class_id_length}")
+        for entry in entries_array:
+            class_index, class_id = entry["class_index"], entry["class_id"]
+            class_ids_array[class_index] = class_id
+        class_ids_path = self._get_class_ids_path(*args, **kwargs)
+        self._save_extra(class_ids_array, class_ids_path)
+
+    def _dump_extra(self, *args, **kwargs) -> None:
+        self._dump_entries(*args, *kwargs)
+        self._dump_class_ids(*args, *kwargs)
+
+    def dump_extra(self, root: Optional[str] = None) -> None:
+        return self._dump_extra(root)