Merge pull request #100 from daita-technologies/develop

✨Release 1.1 `ai-tools`

.dockerignore

@@ -1,6 +1,6 @@
 # Ignore everything
 **
-# Allow heathcheck module
+# Allow healthcheck module
 !healthcheck/**
 # Allow augmentation module
 !augmentation/**
@@ -10,3 +10,5 @@
 !requirements*
 # Allow all Dockerfile
 !Dockerfile*
+# Allow all python files
+!*.py

README.md

@@ -14,7 +14,7 @@ List of preprocessing methods:
 - `NormalizeHue`: Adjust hue of image to match with a reference image if the difference `> 75%`.
 - `NormalizeSaturation`: Adjust saturation of image to match with a reference image if the difference `> 75%`.
 - `NormalizeSharpness`: Adjust sharpness of image to match with a reference image if the difference `> 75%`.
-- `NormalizeContrast`: Adjust constrast of image to match with a reference image if the difference `> 75%`.
+- `NormalizeContrast`: Adjust contrast of image to match with a reference image if the difference `> 75%`.
 - `EqualizeHistogram`: Equalize the histogram of an image if the contrast is low.
 - `IncreaseResolution`: Increase resolution of a tensor image given a reference image.
 
@@ -36,7 +36,7 @@ List of augmentation methods:
 - `random_brightness`: Changing brightness of image randomly. Brightness factor range: `0.75` to `1.5`.
 - `random_hue`: Change hue of image randomly. Hue factor range: `-0.5` to `0.5`.
 - `random_saturation`: Change saturation of image randomly. Saturation factor range: `0.5` to `1.5`.
-- `random_contrast`: Change constrast of image randomly. Contrast factor range: `0.5` to `1.5`.
+- `random_contrast`: Change contrast of image randomly. Contrast factor range: `0.5` to `1.5`.
 - `random_solarize`: Solarize an image randomly; `thresholds = 0.1`, `additions = 0.1`.
 - `random_posterize`: Posterize an image randomly; `bits = 3`.
 - `super_resolution`: Change resolution of an image; `scale = 2`.

SegFormer/Dockerfile.cpu

@@ -0,0 +1,33 @@
+FROM python:3.8.13-slim
+
+# Install necessary build tools
+RUN apt-get update -y \
+    && apt-get install -y --no-install-recommends git build-essential ffmpeg libsm6 libxext6 \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/*
+
+# Change default directory
+WORKDIR /app
+
+# Install dependencies
+COPY requirements-cpu.txt .
+RUN pip install --no-cache-dir \
+                -r requirements-cpu.txt \
+                --find-links=https://download.pytorch.org/whl/torch_stable.html \
+                --find-links=https://download.openmmlab.com/mmcv/dist/cpu/torch1.7.0/index.html
+
+# Set environment variable
+ENV PYTHONPATH="$PYTHONPATH:/app"
+
+# Copy model's checkpoint
+COPY checkpoints/segformer.b0.1024x1024.city.160k.pth checkpoints/segformer.b0.1024x1024.city.160k.pth
+
+# Install dependencies
+COPY mmseg mmseg
+COPY setup.py setup.py
+
+# Copy necessary files to docker image
+COPY local_configs local_configs
+COPY run.py run.py
+
+ENTRYPOINT ["python", "run.py"]

SegFormer/configs/_base_/datasets/ade20k.py

@@ -0,0 +1,60 @@
+# dataset settings
+dataset_type = "ADE20KDataset"
+data_root = "data/ade/ADEChallengeData2016"
+img_norm_cfg = dict(
+    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True
+)
+crop_size = (512, 512)
+train_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(type="LoadAnnotations", reduce_zero_label=True),
+    dict(type="Resize", img_scale=(2048, 512), ratio_range=(0.5, 2.0)),
+    dict(type="RandomCrop", crop_size=crop_size, cat_max_ratio=0.75),
+    dict(type="RandomFlip", prob=0.5),
+    dict(type="PhotoMetricDistortion"),
+    dict(type="Normalize", **img_norm_cfg),
+    dict(type="Pad", size=crop_size, pad_val=0, seg_pad_val=255),
+    dict(type="DefaultFormatBundle"),
+    dict(type="Collect", keys=["img", "gt_semantic_seg"]),
+]
+test_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(
+        type="MultiScaleFlipAug",
+        img_scale=(2048, 512),
+        # img_ratios=[0.5, 0.75, 1.0, 1.25, 1.5, 1.75],
+        flip=False,
+        transforms=[
+            dict(type="Resize", keep_ratio=True),
+            dict(type="RandomFlip"),
+            dict(type="Normalize", **img_norm_cfg),
+            dict(type="ImageToTensor", keys=["img"]),
+            dict(type="Collect", keys=["img"]),
+        ],
+    ),
+]
+data = dict(
+    samples_per_gpu=4,
+    workers_per_gpu=4,
+    train=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir="images/training",
+        ann_dir="annotations/training",
+        pipeline=train_pipeline,
+    ),
+    val=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir="images/validation",
+        ann_dir="annotations/validation",
+        pipeline=test_pipeline,
+    ),
+    test=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir="images/validation",
+        ann_dir="annotations/validation",
+        pipeline=test_pipeline,
+    ),
+)

SegFormer/configs/_base_/datasets/chase_db1.py

@@ -0,0 +1,66 @@
+# dataset settings
+dataset_type = "ChaseDB1Dataset"
+data_root = "data/CHASE_DB1"
+img_norm_cfg = dict(
+    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True
+)
+img_scale = (960, 999)
+crop_size = (128, 128)
+train_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(type="LoadAnnotations"),
+    dict(type="Resize", img_scale=img_scale, ratio_range=(0.5, 2.0)),
+    dict(type="RandomCrop", crop_size=crop_size, cat_max_ratio=0.75),
+    dict(type="RandomFlip", prob=0.5),
+    dict(type="PhotoMetricDistortion"),
+    dict(type="Normalize", **img_norm_cfg),
+    dict(type="Pad", size=crop_size, pad_val=0, seg_pad_val=255),
+    dict(type="DefaultFormatBundle"),
+    dict(type="Collect", keys=["img", "gt_semantic_seg"]),
+]
+test_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(
+        type="MultiScaleFlipAug",
+        img_scale=img_scale,
+        # img_ratios=[0.5, 0.75, 1.0, 1.25, 1.5, 1.75, 2.0],
+        flip=False,
+        transforms=[
+            dict(type="Resize", keep_ratio=True),
+            dict(type="RandomFlip"),
+            dict(type="Normalize", **img_norm_cfg),
+            dict(type="ImageToTensor", keys=["img"]),
+            dict(type="Collect", keys=["img"]),
+        ],
+    ),
+]
+
+data = dict(
+    samples_per_gpu=4,
+    workers_per_gpu=4,
+    train=dict(
+        type="RepeatDataset",
+        times=40000,
+        dataset=dict(
+            type=dataset_type,
+            data_root=data_root,
+            img_dir="images/training",
+            ann_dir="annotations/training",
+            pipeline=train_pipeline,
+        ),
+    ),
+    val=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir="images/validation",
+        ann_dir="annotations/validation",
+        pipeline=test_pipeline,
+    ),
+    test=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir="images/validation",
+        ann_dir="annotations/validation",
+        pipeline=test_pipeline,
+    ),
+)

SegFormer/configs/_base_/datasets/cityscapes.py

@@ -0,0 +1,60 @@
+# dataset settings
+dataset_type = "CityscapesDataset"
+data_root = "data/cityscapes/"
+img_norm_cfg = dict(
+    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True
+)
+crop_size = (512, 1024)
+train_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(type="LoadAnnotations"),
+    dict(type="Resize", img_scale=(2048, 1024), ratio_range=(0.5, 2.0)),
+    dict(type="RandomCrop", crop_size=crop_size, cat_max_ratio=0.75),
+    dict(type="RandomFlip", prob=0.5),
+    dict(type="PhotoMetricDistortion"),
+    dict(type="Normalize", **img_norm_cfg),
+    dict(type="Pad", size=crop_size, pad_val=0, seg_pad_val=255),
+    dict(type="DefaultFormatBundle"),
+    dict(type="Collect", keys=["img", "gt_semantic_seg"]),
+]
+test_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(
+        type="MultiScaleFlipAug",
+        img_scale=(2048, 1024),
+        # img_ratios=[0.5, 0.75, 1.0, 1.25, 1.5, 1.75],
+        flip=False,
+        transforms=[
+            dict(type="Resize", keep_ratio=True),
+            dict(type="RandomFlip"),
+            dict(type="Normalize", **img_norm_cfg),
+            dict(type="ImageToTensor", keys=["img"]),
+            dict(type="Collect", keys=["img"]),
+        ],
+    ),
+]
+data = dict(
+    samples_per_gpu=2,
+    workers_per_gpu=2,
+    train=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir="leftImg8bit/train",
+        ann_dir="gtFine/train",
+        pipeline=train_pipeline,
+    ),
+    val=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir="leftImg8bit/val",
+        ann_dir="gtFine/val",
+        pipeline=test_pipeline,
+    ),
+    test=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir="leftImg8bit/val",
+        ann_dir="gtFine/val",
+        pipeline=test_pipeline,
+    ),
+)

SegFormer/configs/_base_/datasets/cityscapes_768x768.py

@@ -0,0 +1,38 @@
+_base_ = "./cityscapes.py"
+img_norm_cfg = dict(
+    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True
+)
+crop_size = (768, 768)
+train_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(type="LoadAnnotations"),
+    dict(type="Resize", img_scale=(2049, 1025), ratio_range=(0.5, 2.0)),
+    dict(type="RandomCrop", crop_size=crop_size, cat_max_ratio=0.75),
+    dict(type="RandomFlip", prob=0.5),
+    dict(type="PhotoMetricDistortion"),
+    dict(type="Normalize", **img_norm_cfg),
+    dict(type="Pad", size=crop_size, pad_val=0, seg_pad_val=255),
+    dict(type="DefaultFormatBundle"),
+    dict(type="Collect", keys=["img", "gt_semantic_seg"]),
+]
+test_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(
+        type="MultiScaleFlipAug",
+        img_scale=(2049, 1025),
+        # img_ratios=[0.5, 0.75, 1.0, 1.25, 1.5, 1.75],
+        flip=False,
+        transforms=[
+            dict(type="Resize", keep_ratio=True),
+            dict(type="RandomFlip"),
+            dict(type="Normalize", **img_norm_cfg),
+            dict(type="ImageToTensor", keys=["img"]),
+            dict(type="Collect", keys=["img"]),
+        ],
+    ),
+]
+data = dict(
+    train=dict(pipeline=train_pipeline),
+    val=dict(pipeline=test_pipeline),
+    test=dict(pipeline=test_pipeline),
+)

SegFormer/configs/_base_/datasets/cityscapes_769x769.py

@@ -0,0 +1,38 @@
+_base_ = "./cityscapes.py"
+img_norm_cfg = dict(
+    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True
+)
+crop_size = (769, 769)
+train_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(type="LoadAnnotations"),
+    dict(type="Resize", img_scale=(2049, 1025), ratio_range=(0.5, 2.0)),
+    dict(type="RandomCrop", crop_size=crop_size, cat_max_ratio=0.75),
+    dict(type="RandomFlip", prob=0.5),
+    dict(type="PhotoMetricDistortion"),
+    dict(type="Normalize", **img_norm_cfg),
+    dict(type="Pad", size=crop_size, pad_val=0, seg_pad_val=255),
+    dict(type="DefaultFormatBundle"),
+    dict(type="Collect", keys=["img", "gt_semantic_seg"]),
+]
+test_pipeline = [
+    dict(type="LoadImageFromFile"),
+    dict(
+        type="MultiScaleFlipAug",
+        img_scale=(2049, 1025),
+        # img_ratios=[0.5, 0.75, 1.0, 1.25, 1.5, 1.75],
+        flip=False,
+        transforms=[
+            dict(type="Resize", keep_ratio=True),
+            dict(type="RandomFlip"),
+            dict(type="Normalize", **img_norm_cfg),
+            dict(type="ImageToTensor", keys=["img"]),
+            dict(type="Collect", keys=["img"]),
+        ],
+    ),
+]
+data = dict(
+    train=dict(pipeline=train_pipeline),
+    val=dict(pipeline=test_pipeline),
+    test=dict(pipeline=test_pipeline),
+)