Merge pull request #4 from oneapi-src/main-upstream

Main upstream

.gitignore

@@ -0,0 +1,141 @@
+# 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
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+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/
+
+data/
+output/
+.vscode

README.md

@@ -168,13 +168,13 @@ conda config --set solver libmamba
 
 | Packages | Version | 
 | -------- | ------- |
-| intelpython3_core | 2024.0.0 |
-| python | 3.9.18 |
-| intelpython3_core | 2024.0.0 |
-| intel-aikit-tensorflow | 2024.0 |
-| tqdm | 4.64.0 |
-| pip | 23.3 |
-| opencv-python | 4.8.0.76 |
+| intelpython3_core | 2024.1.0 |
+| python | 3.9 |
+| intelpython3_core | 2024.1.0 |
+| intel-aikit-tensorflow | 2024.1 |
+| tqdm | 4.66.2 |
+| pip | 24.0 |
+| opencv-python | 4.9.0.80 |
 
 The dependencies required to properly execute this workflow can be found in the yml file [$WORKSPACE/env/intel_env.yml](env/intel_env.yml).
 

env/intel_env.yml

@@ -3,10 +3,10 @@ channels:
   - intel
   - conda-forge
 dependencies:
-  - intelpython3_core=2024.0.0
-  - python=3.9.18
-  - intel-aikit-tensorflow=2024.0
-  - tqdm=4.64.0
-  - pip=23.3  
+  - intelpython3_core=2024.1.0
+  - python=3.9
+  - intel-aikit-tensorflow=2024.1
+  - tqdm=4.66.2
+  - pip=24.0  
   - pip:
-    - opencv-python==4.8.0.76
+    - opencv-python==4.9.0.80

src/intel_neural_compressor/utils.py

@@ -44,8 +44,8 @@
                      "vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5"
 
 cryptogen = SystemRandom()
-class_colors = [(cryptogen.randint(0, 255), cryptogen.randint(
-    0, 255), cryptogen.randint(0, 255)) for _ in range(5000)]
+class_colors = [np.array((cryptogen.randint(0, 255), cryptogen.randint(
+    0, 255), cryptogen.randint(0, 255))) for _ in range(5000)]
 
 
 def get_colored_segmentation_image(seg_arr, n_classes, colors=None):
@@ -58,9 +58,8 @@ def get_colored_segmentation_image(seg_arr, n_classes, colors=None):
     seg_img = np.zeros((output_height, output_width, 3))
 
     for c in range(n_classes):
-        seg_img[:, :, 0] += ((seg_arr[:, :] == c) * (colors[c][0])).astype('uint8')
-        seg_img[:, :, 1] += ((seg_arr[:, :] == c) * (colors[c][1])).astype('uint8')
-        seg_img[:, :, 2] += ((seg_arr[:, :] == c) * (colors[c][2])).astype('uint8')
+        mask = (seg_arr == c)
+        seg_img[mask] += colors[c].astype('uint8')
 
     return seg_img
 
@@ -142,7 +141,7 @@ def get_segmentation_array(image_input, nclasses, width, height, no_reshape=Fals
         raise Exception(f"get_segmentation_array: Can't process input type {str(type(image_input))}")
 
     img = cv2.resize(img, (width, height), interpolation=cv2.INTER_NEAREST)
-    img = img[:, :, 0]
+    img = img.mean(axis=-1)
 
     for c in range(nclasses):
         seg_labels[:, :, c] = (img == c).astype(int)
@@ -335,7 +334,9 @@ def evaluate(model=None, inp_images=None, annotations=None, inp_images_dir=None,
             fn[cl_i] += np.sum((pr != cl_i) * (gt == cl_i))
             n_pixels[cl_i] += np.sum(gt == cl_i)
 
-    cl_wise_score = tp / (tp + fp + fn + 0.000000000001)
+    cl_wise_score = tp / (tp + fp + fn)
+    cl_wise_score = np.nan_to_num(cl_wise_score, nan=0.0)
+
     n_pixels_norm = n_pixels / np.sum(n_pixels)
     frequency_weighted_iu = np.sum(cl_wise_score * n_pixels_norm)
     mean_iu = np.mean(cl_wise_score)
@@ -466,11 +467,10 @@ def predict(model=None, inp=None, out_fname=None, checkpoints_path=None, overlay
     pr = model.predict(np.array([x]))[0]
     pr = pr.reshape((output_height, output_width, n_classes)).argmax(axis=2)
 
-    seg_img = visualize_segmentation(pr, inp, n_classes=n_classes, colors=colors,
-                                     overlay_img=overlay_img, prediction_width=prediction_width,
-                                     prediction_height=prediction_height)
-
     if out_fname is not None:
+        seg_img = visualize_segmentation(pr, inp, n_classes=n_classes, colors=colors,
+                                 overlay_img=overlay_img, prediction_width=prediction_width,
+                                 prediction_height=prediction_height)
         cv2.imwrite(out_fname, seg_img)
 
     return pr
@@ -540,7 +540,6 @@ def train_hyperparameters_tuning(model, train_images, train_annotations, batch_s
 
             start_time = time.time()
             hist=model.fit_generator(train_gen, steps_per_epoch, epochs=epochs, workers=1, use_multiprocessing=False)
-            #model.fit_generator(train_gen, steps_per_epoch, epochs=epochs, workers=1, use_multiprocessing=False)
             total_time += time.time()-start_time
             print("Fit number: ", ctr, " ==> Time Taken for Training in seconds --> ", time.time()-start_time)
             if best_config["accuracy"] < hist.history["accuracy"][0]:          

src/utils.py

@@ -42,8 +42,8 @@
                      "vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5"
 
 cryptogen = SystemRandom()
-class_colors = [(cryptogen.randint(0, 255), cryptogen.randint(
-    0, 255), cryptogen.randint(0, 255)) for _ in range(5000)]
+class_colors = [np.array((cryptogen.randint(0, 255), cryptogen.randint(
+    0, 255), cryptogen.randint(0, 255))) for _ in range(5000)]
 
 
 def get_colored_segmentation_image(seg_arr, n_classes, colors=None):
@@ -56,9 +56,8 @@ def get_colored_segmentation_image(seg_arr, n_classes, colors=None):
     seg_img = np.zeros((output_height, output_width, 3))
 
     for c in range(n_classes):
-        seg_img[:, :, 0] += ((seg_arr[:, :] == c) * (colors[c][0])).astype('uint8')
-        seg_img[:, :, 1] += ((seg_arr[:, :] == c) * (colors[c][1])).astype('uint8')
-        seg_img[:, :, 2] += ((seg_arr[:, :] == c) * (colors[c][2])).astype('uint8')
+        mask = (seg_arr == c)
+        seg_img[mask] += colors[c].astype('uint8') 
 
     return seg_img
 
@@ -140,7 +139,7 @@ def get_segmentation_array(image_input, nclasses, width, height, no_reshape=Fals
         raise Exception(f"get_segmentation_array: Can't process input type {str(type(image_input))}")
 
     img = cv2.resize(img, (width, height), interpolation=cv2.INTER_NEAREST)
-    img = img[:, :, 0]
+    img = img.mean(axis=-1)
 
     for c in range(nclasses):
         seg_labels[:, :, c] = (img == c).astype(int)
@@ -302,7 +301,6 @@ def evaluate(model=None, inp_images=None, annotations=None, inp_images_dir=None,
                 "output_width": output_width
             }, f)
 
-    if checkpoints_path is not None:
         latest_checkpoint = checkpoints_path  # find_latest_checkpoint(checkpoints_path)
         if latest_checkpoint is not None:
             print("Loading the weights from latest checkpoint ",
@@ -333,7 +331,9 @@ def evaluate(model=None, inp_images=None, annotations=None, inp_images_dir=None,
             fn[cl_i] += np.sum((pr != cl_i) * (gt == cl_i))
             n_pixels[cl_i] += np.sum(gt == cl_i)
 
-    cl_wise_score = tp / (tp + fp + fn + 0.000000000001)
+    cl_wise_score = tp / (tp + fp + fn)
+    cl_wise_score = np.nan_to_num(cl_wise_score, nan=0.0)
+
     n_pixels_norm = n_pixels / np.sum(n_pixels)
     frequency_weighted_iu = np.sum(cl_wise_score * n_pixels_norm)
     mean_iu = np.mean(cl_wise_score)
@@ -463,11 +463,10 @@ def predict(model=None, inp=None, out_fname=None, checkpoints_path=None, overlay
     pr = model.predict(np.array([x]))[0]
     pr = pr.reshape((output_height, output_width, n_classes)).argmax(axis=2)
 
-    seg_img = visualize_segmentation(pr, inp, n_classes=n_classes, colors=colors,
-                                     overlay_img=overlay_img, prediction_width=prediction_width,
-                                     prediction_height=prediction_height)
-
     if out_fname is not None:
+        seg_img = visualize_segmentation(pr, inp, n_classes=n_classes, colors=colors,
+                                overlay_img=overlay_img, prediction_width=prediction_width,
+                                prediction_height=prediction_height)
         cv2.imwrite(out_fname, seg_img)
 
     return pr
@@ -540,7 +539,6 @@ def train_hyperparameters_tuning(model, train_images, train_annotations, batch_s
 
             start_time = time.time()
             hist=model.fit_generator(train_gen, steps_per_epoch, epochs=epochs, workers=1, use_multiprocessing=False)
-            #model.fit_generator(train_gen, steps_per_epoch, epochs=epochs, workers=1, use_multiprocessing=False)
             total_time += time.time()-start_time
             print("Fit number: ", ctr, " ==> Time Taken for Training in seconds --> ", time.time()-start_time)
             if best_config["accuracy"] < hist.history["accuracy"][0]: