Upload 6 files

.gitignore

@@ -0,0 +1,163 @@
+workspace.code-workspace
+flagged/
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/#use-with-ide
+.pdm.toml
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/

README.md

@@ -1,13 +1,13 @@
 ---
-title: CLIP CamGrad
-emoji: 🔥
-colorFrom: indigo
-colorTo: indigo
+title: Gradio OpenAI CLIP Grad-CAM
+emoji: 🔭
+colorFrom: yellow
+colorTo: blue
 sdk: gradio
-sdk_version: 3.47.1
-app_file: app.py
+sdk_version: 2.9.4
+app_file: gradcam/app.py
 pinned: false
 license: mit
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces#reference

gradcam/app.py

@@ -0,0 +1,58 @@
+import gradio as gr
+import clip
+import torch
+
+import utils
+
+#clip_model = "RN50x4"
+clip_model = "RN50x64"
+device = "cuda" if torch.cuda.is_available() else "cpu"
+model, preprocess = clip.load(clip_model, device=device, jit=False)
+model.eval()
+
+
+def grad_cam_fn(text, img, saliency_layer):
+    resize = model.visual.input_resolution
+    img = img.resize((resize, resize))
+
+    text_input = clip.tokenize([text]).to(device)
+    text_feature = model.encode_text(text_input).float()
+    image_input = preprocess(img).unsqueeze(0).to(device)
+
+    attn_map = utils.gradCAM(
+        model.visual,
+        image_input,
+        text_feature,
+        getattr(model.visual, saliency_layer)
+    )
+    attn_map = attn_map.squeeze().detach().cpu().numpy()
+    attn_map = utils.getAttMap(img, attn_map)
+
+    return attn_map
+
+
+interface = gr.Interface(
+    fn=grad_cam_fn,
+    inputs=[
+        gr.inputs.Textbox(
+            label="Target Text",
+            lines=1),
+        gr.inputs.Image(
+            label='Input Image',
+            image_mode="RGB",
+            type='pil',
+            shape=(512, 512)),
+        gr.inputs.Dropdown(
+            ["layer4", "layer3", "layer2", "layer1"],
+            default="layer4",
+            label="Saliency Layer")
+    ],
+    outputs=gr.outputs.Image(
+        type="pil", 
+        label="Attention Map"),
+    examples=[
+        ['a cat lying on the floor', 'assets/cat_dog.jpg', 'layer4'],
+        ['a dog sitting', 'assets/cat_dog.jpg', 'layer4']
+    ],
+    description="OpenAI CLIP Grad CAM")
+interface.launch()

gradcam/utils.py

@@ -0,0 +1,100 @@
+import numpy as np
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import matplotlib.cm
+from PIL import Image
+
+# Adapted from: https://colab.research.google.com/github/kevinzakka/clip_playground/blob/main/CLIP_GradCAM_Visualization.ipynb
+class Hook:
+    """Attaches to a module and records its activations and gradients."""
+
+    def __init__(self, module: nn.Module):
+        self.data = None
+        self.hook = module.register_forward_hook(self.save_grad)
+        
+    def save_grad(self, module, input, output):
+        self.data = output
+        output.requires_grad_(True)
+        output.retain_grad()
+        
+    def __enter__(self):
+        return self
+    
+    def __exit__(self, exc_type, exc_value, exc_traceback):
+        self.hook.remove()
+        
+    @property
+    def activation(self) -> torch.Tensor:
+        return self.data
+    
+    @property
+    def gradient(self) -> torch.Tensor:
+        return self.data.grad
+
+
+# Reference: https://arxiv.org/abs/1610.02391
+def gradCAM(
+    model: nn.Module,
+    input: torch.Tensor,
+    target: torch.Tensor,
+    layer: nn.Module
+) -> torch.Tensor:
+    # Zero out any gradients at the input.
+    if input.grad is not None:
+        input.grad.data.zero_()
+        
+    # Disable gradient settings.
+    requires_grad = {}
+    for name, param in model.named_parameters():
+        requires_grad[name] = param.requires_grad
+        param.requires_grad_(False)
+        
+    # Attach a hook to the model at the desired layer.
+    assert isinstance(layer, nn.Module)
+    with Hook(layer) as hook:        
+        # Do a forward and backward pass.
+        output = model(input)
+        output.backward(target)
+
+        grad = hook.gradient.float()
+        act = hook.activation.float()
+    
+        # Global average pool gradient across spatial dimension
+        # to obtain importance weights.
+        alpha = grad.mean(dim=(2, 3), keepdim=True)
+        # Weighted combination of activation maps over channel
+        # dimension.
+        gradcam = torch.sum(act * alpha, dim=1, keepdim=True)
+        # We only want neurons with positive influence so we
+        # clamp any negative ones.
+        gradcam = torch.clamp(gradcam, min=0)
+
+    # Resize gradcam to input resolution.
+    gradcam = F.interpolate(
+        gradcam,
+        input.shape[2:],
+        mode='bicubic',
+        align_corners=False)
+    
+    # Restore gradient settings.
+    for name, param in model.named_parameters():
+        param.requires_grad_(requires_grad[name])
+        
+    return gradcam
+
+
+# Modified from: https://github.com/salesforce/ALBEF/blob/main/visualization.ipynb
+def getAttMap(img, attn_map):
+    # Normalize attention map
+    attn_map = attn_map - attn_map.min()
+    if attn_map.max() > 0:
+        attn_map = attn_map / attn_map.max()
+
+    H = matplotlib.cm.jet(attn_map)
+    H = (H * 255).astype(np.uint8)[:, :, :3]
+    img_heatmap = Image.fromarray(H)
+    img_heatmap = img_heatmap.resize((256, 256))
+    
+    return Image.blend(
+        img.resize((256, 256)), img_heatmap, 0.4)

requirements.txt

@@ -0,0 +1,6 @@
+gradio>=2.9.0,<2.10.0
+torch>=1.10.0,<1.11.0
+git+https://github.com/openai/CLIP.git
+Pillow
+matplotlib
+numpy