This is your work, valued

Poland

Krzysztof Kolasinski

Elite
@kmkolasinski

A bio

AwesomeBump. AwesomeBump is a free program written using Qt library designed to generate normal, height, specular or ambient occlusion textures from a single image. Since the image processing is done in 99% on GPU the program runs very fast and all the parameters can be changed in real time. AB was made to be a new alternative to known gimp plugin called Insane Bump.

1.8k

deep-learning-notes. Experiments with Deep Learning

1.3k

nano-umap. Simplified implementation of UMAP like dimensionality reduction algorithm

55

clever-camera. A simple monitoring camera web server which uses deep learning model (classification) to filter or group camera events.

21

receptivefield. Gradient based receptive field estimation for Convolutional Neural Networks

14

tfrecords-reader. TFRecords reader with Random Access

6

superpoint-video-stitcher. Jupyter Notebook

6

fast-bfmatcher. Faster implementation of the OpenCV BFMatcher(cv2.NORM_L2) matcher for image keypoint matching

6

fast-ivf. Jupyter Notebook

5

tensorflow-nanoGPT. Example how to train GPT-2 (XLA + AMP), export to SavedModel and serve with Tensorflow Serving

5

keras_detection. Jupyter Notebook

2

models. Models and examples built with TensorFlow

2

ipython-llm-magics. Simple wrapper for OpenAI chat to nicely parse and display the results returned from the API client

2

repltilian. Interactive Swift REPL wrapper for Python

2

keras-llm-light. Jupyter Notebook

2

usearch. Fast Open-Source Search & Clustering engine × for Vectors & 🔜 Strings × in C++, C, Python, JavaScript, Rust, Java, Objective-C, Swift, C#, GoLang, and Wolfram 🔍

1

bubel-viewer. A simple 3D visualization tool for tight binding models

1

transformer-pytorch. Implementation of "Attention Is All You Need" using pytorch

1

triton-saved-model. Jupyter Notebook

1

caffenet-benchmark. Evaluation of the CNN design choices performance on ImageNet-2012.

1

QtnProperty. Extended properties for Qt5

1

checkered-cnn. A new technique to increase receptive field in convolutional neural networks without losing so much information.

1

approxkernel. approxkernel

1

Multicore-TSNE. Parallel t-SNE implementation with Python and Torch wrappers.

1

luminoth. Deep Learning toolkit for Computer Vision

1

DeepLearningImplementations. Implementation of recent Deep Learning papers

1

awesome-deep-learning-papers. The most cited deep learning papers

1

Practical_RL. A course in reinforcement learning in the wild

1

jplag. JPlag - Detecting Software Plagiarism

1