Uploaded to PyPi

build/lib/sightseer/__init__.py

@@ -0,0 +1,6 @@
+# __init__.py
+__version__ = "1.0.1"
+
+from sightseer.sightseer import Sightseer
+from sightseer.zoo import *
+from sightseer.proc import *

main.py

@@ -1,5 +1,5 @@
-from sight.sight import Sightseer
-from sight.zoo import YOLOv3Client
+from sightseer.sight import Sightseer
+from sightseer.zoo import YOLOv3Client
 
 # downloading and configuring weights and hyperparams
 yolo = YOLOv3Client()

setup.py

@@ -0,0 +1,18 @@
+import pathlib
+from setuptools import setup
+
+HERE = pathlib.Path(__file__).parent
+README = (HERE / "README.md").read_text()
+
+setup(
+  name="sightseer",
+  version="1.0.1",
+  description="State-of-the-art Computer Vision and Object Detection for TensorFlow.",
+  long_description=README,
+  long_description_content_type="text/markdown",
+  author="",
+  author_email="",
+  license="ASF",
+  packages=["sightseer"],
+  zip_safe=False
+)

sightseer.egg-info/PKG-INFO

@@ -0,0 +1,116 @@
+Metadata-Version: 2.1
+Name: sightseer
+Version: 1.0.1
+Summary: State-of-the-art Computer Vision and Object Detection for TensorFlow.
+Home-page: UNKNOWN
+Author: 
+Author-email: 
+License: ASF
+Description: <p align="center">
+            <br>
+        	<img src="./Assets/logo.png" width=200>
+            <br>
+        <p>
+
+        <p align="center">
+            <a href="https://github.com/rish-16/sight/blob/master/LICENSE">
+        		<img alt="AUR license" src="https://img.shields.io/badge/License-Apache%202.0-yellow.svg">
+            </a>
+        </p>
+
+        <h3 align="center">
+        <p>State-of-the-art Computer Vision and Object Detection for TensorFlow.</p>
+        </h3>
+
+        *Sight* provides state-of-the-art general-purpose architectures (YOLO9000, MaskRCNN, Fast/Faster RCNN, SSD...) for Computer Vision and Object Detection tasks with 30+ pretrained models written in TensorFlow 1.15.
+
+        ## Installation
+
+        `sight` is written in Python 3.5+ and TensorFlow 1.15. 
+
+        Ideally, `sight` should be installed in a [virtual environments](https://docs.python.org/3/library/venv.html). If you're unfamiliar with Python virtual environments, check out this [tutorial](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/) on getting started.
+
+        ### Via PyPi
+
+        To use `sight`, you must first have TensorFlow installed. To do so, follow the instructions on the [TensorFlow installation page](https://www.tensorflow.org/install/pip?lang=python3).
+
+        When your virtual environment is set up with TensorFlow, you can install `sight` using `pip`:
+
+        ```bash
+        pip install sight
+        ```
+
+        ### From Source
+
+        Again, to install from source, you need TensorFlow 1.15 and above running in a virtual environment. You can install the package by cloning the repo and installing the dependencies:
+
+        ```bash
+        git clone https://github.com/rish-16/sight
+        cd sight
+        pip install .
+        ```
+
+        ### Model Architectures
+
+        1. YOLOv3 (Darknet by Joseph Redmon)
+        2. Mask R-CNN (Facebook AI Research)
+
+        ## Usage
+
+        <strong>1a. Loading images</strong>
+
+        ```python
+        from sight import Sightseer
+
+        ss = Sightseer()
+        image = ss.load_image("path/to/image")
+        ```
+
+        <strong>1b. Loading videos</strong>
+
+        ```python
+        from sight import Sightseer
+
+        ss = Sightseer()
+        frames = ss.load_vidsource("path/to/video", return_data=True)
+        ```
+
+        <strong>1c. Loading webcam footage</strong>
+
+        ```python
+        from sight import Sightseer
+
+        ss = Sightseer()
+        image = ss.load_webcam()
+        ```
+
+        <strong>1d. Loading screen grab footage</strong>
+
+        ```python
+        from sight import Sightseer
+
+        ss = Sightseer()
+        image = ss.screen_grab()
+        ```
+
+        <strong>2. Using models from `sight.zoo`</strong>
+
+        Once installed, any model offered by `sight` can be accessed in less than 10 lines of code. For instance, the code to use the YOLOv3 (Darknet) model is as follows:
+
+        ```python
+        from sight import Sightseer
+        from sight.zoo import YOLOv3Client
+
+        yolo = YOLOv3Client()
+        yolo.load_model() # downloads weights
+
+        # loading images from local system
+        ss = Sightseer("path/to/img")
+        image = ss.load_image()
+
+        # returns array of labels, confidence, and bounding box info
+        preds, pred_img = yolo.predict(image, return_image=True)
+        ss.render_image(pred_img)
+        ```
+Platform: UNKNOWN
+Description-Content-Type: text/markdown

sightseer.egg-info/SOURCES.txt

@@ -0,0 +1,12 @@
+README.md
+setup.py
+sightseer/__init__.py
+sightseer/blocks.py
+sightseer/proc.py
+sightseer/sightseer.py
+sightseer/zoo.py
+sightseer.egg-info/PKG-INFO
+sightseer.egg-info/SOURCES.txt
+sightseer.egg-info/dependency_links.txt
+sightseer.egg-info/not-zip-safe
+sightseer.egg-info/top_level.txt

sightseer.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

sightseer.egg-info/not-zip-safe

@@ -0,0 +1 @@
+

sightseer.egg-info/top_level.txt

@@ -0,0 +1 @@
+sightseer

sightseer/__init__.py

@@ -0,0 +1,6 @@
+# __init__.py
+__version__ = "1.0.1"
+
+from sightseer.sightseer import Sightseer
+from sightseer.zoo import *
+from sightseer.proc import *

sightseer/blocks.py

@@ -0,0 +1,123 @@
+import struct
+import numpy as np
+import tensorflow as tf
+from tensorflow.keras.layers import Conv2D, ZeroPadding2D, BatchNormalization, LeakyReLU, add	
+
+class BoundingBox(object):
+    def __init__(self, xmin, ymin, xmax, ymax, objectness=None, classes=None):
+        self.xmin = xmin
+        self.ymin = ymin
+        self.xmax = xmax
+        self.ymax = ymax
+
+        self.objectness = objectness
+        self.classes = classes
+
+        self.label = -1
+        self.confidence = -1
+
+    def get_label(self):
+        if self.label == -1:
+            self.label = np.argmax(self.classes)
+        return self.label
+
+    def get_confidence(self):
+        if self.confidence == -1:
+            self.confidence = self.classes[self.get_label()]
+        return self.confidence		
+
+class SightLoader():
+    def __init__(self, weights_path):
+        """
+        Weights loading framework for all Sight models
+        """
+        with open(weights_path, 'rb') as wf:
+            major, = struct.unpack('i', wf.read(4))
+            minor, = struct.unpack('i', wf.read(4))
+            revision, = struct.unpack('i', wf.read(4))
+
+            if (major*10+ minor) >= 2 and major < 1000 and minor < 1000:
+                wf.read(8)
+            else:
+                wf.read(4)
+
+            transpose = (major > 1000) or (minor > 1000)
+
+            binary = wf.read()
+
+        self.offset = 0
+        self.all_weights = np.frombuffer(binary, dtype="float32")
+
+    def read_bytes(self, chunk_size):
+        self.offset = self.offset + chunk_size
+        return self.all_weights[self.offset - chunk_size:self.offset]
+
+    def load_weights(self, model, verbose=True):
+        for i in range(106): # standard darknet layer count
+            try:
+                conv_layer = model.get_layer("conv_" + str(i))
+
+                if verbose:
+                    print ("Loading Convolution #{}".format(i))
+
+                if i not in [81, 93, 105]:
+                    norm_layer = model.get_layer("bnorm_" + str(i))
+
+                    size = np.prod(norm_layer.get_weights()[0].shape)
+
+                    beta = self.read_bytes(size)
+                    gamma = self.read_bytes(size)
+                    mean = self.read_bytes(size)
+                    var = self.read_bytes(size)
+
+                    weights = norm_layer.set_weights([gamma, beta, mean, var])
+
+                if len(conv_layer.get_weights()) > 1:
+                    bias = self.read_bytes(np.prod(conv_layer.get_weights()[1].shape))
+                    kernel = self.read_bytes(np.prod(conv_layer.get_weights()[0].shape))
+
+                    kernel = kernel.reshape(list(reversed(conv_layer.get_weights()[0].shape)))
+                    kernel = kernel.transpose([2, 3, 1, 0])
+                    conv_layer.set_weights([kernel, bias])
+                else:
+                    kernel = self.read_bytes(np.prod(conv_layer.get_weights()[0].shape))
+                    kernel = kernel.reshape(list(reversed(conv_layer.get_weights()[0].shape)))
+                    kernel = kernel.transpose([2, 3, 1, 0])
+                    conv_layer.set_weights([kernel])
+
+            except ValueError:
+                if verbose:
+                    print ("No Convolution #{}".format(i))
+                else:
+                    pass
+
+        if verbose:
+            print ("Finished loading weights into model. Predicting on input data...")
+
+    def reset_offset(self):
+        self.offset = 0		
+
+class ConvBlock():
+	def get_conv_block(inp, convs, skip=True):
+		x = inp
+		count = 0
+
+		for conv in convs:
+			if count == (len(convs) - 2) and skip:
+				skip_conn = x
+			count += 1
+
+			if conv['stride'] > 1: x = ZeroPadding2D(((1,0),(1,0)))(x)
+
+			x = Conv2D(conv['filter'], 
+					conv['kernel'], 
+					strides=conv['stride'], 
+					padding="valid" if conv['stride']>1 else "same", 
+					name="conv_"+str(conv['layer_idx']), 
+					use_bias=False if conv['bnorm'] else True)(x)
+
+			if conv['bnorm']: x = BatchNormalization(epsilon=0.001, name="bnorm_"+str(conv['layer_idx']))(x)
+
+			if conv['leaky']: x = LeakyReLU(alpha=0.1, name="leaky_"+str(conv['layer_idx']))(x)
+
+		return add([skip_conn, x]) if skip else x