Name: Dave Xanatos
Type: User
Company: XANATOS
Bio: Freelancer. Python, JavaScript, PHP, VBScript/ASP, HTML5, CSS, PBASIC, SPIN, Perl. Robotics, AI, AGI, Computer Vision, Deep Learning... fun.
Location: Western MA, USA
Blog: http://www.xanatos.com/
Dave Xanatos's Projects
modest natural-language processing in javascript
A TensorFlow Implementation of DC-TTS: yet another text-to-speech model
A TensorFlow implementation of Baidu's DeepSpeech architecture
Code for the paper "Language Models are Unsupervised Multitask Learners"
Dataset of GPT-2 outputs for research in detection, biases, and more
Disk Image for Raspberry Pi 4B 8 Gig with functional implementation of GPT-2 with Tensorflow 1.13.1
Repository for Autonomous Robotic Human-Mimetic Host Builds
A set of Python classes that transport OpenCV images from one computer to another using PyZMQ messaging.
Generates Lenticular Images
Models and examples built with TensorFlow
β¨State-of-the-art coreference resolution based on neural nets and spaCy
NGC-Learn: Predictive Coding and Neurobiologically-Motivated Learning in Python
A practical and feature-rich paraphrasing framework to augment human intents in text form to build robust NLU models for conversational engines.
Speech Recognition using PocketSphinx on Raspberry Pi3
An interpreter module for AIML (Artificial Intelligence Markup Language), implemented in pure Python
Python code and documents
Speech recognition module for Python, supporting several engines and APIs, online and offline.
Switchboard Dialog Act Corpus with Penn Treebank links
How to train a TensorFlow Object Detection Classifier for multiple object detection on Windows
This is the code for 'How to Make a Simple Tensorflow Speech Recognizer' by @Sirajology on Youtube
Total Text Dataset - ICDAR 2017. It consists of 1555 images with more than 3 different text orientations: Horizontal, Multi-Oriented, and Curved, one of a kind.
π€ Transformers: State-of-the-art Natural Language Processing for TensorFlow 2.0 and PyTorch.
Perspective recovery of text using transformed ellipses
UPDATE 20210705: Adding ZeroMQ functions in to allow full-duplex comms with deep parts of the hosts' cognitive processes, and expose speech recognition background processes, visual processes, etc. ORIGINAL DESCRIPTION >>>This is a working version of the iconic Behavior Pad "Rose Graph" Interface that the Westworld Techs use to set Host Behavioral Attributes (See SCREENSHOT.jpg). It is written in HTML5 using Canvas, CSS, JavaScript and PHP. It has only been tested in Chrome. I have a copy running on an Apache server "in the wild" as well as a copy running on Raspberry Pi SBCs that serve as "Host Brains" in my robots. Those Raspberries are running NGINX server and VSFTPC. This project allows you to set and save host parameters (reads and writes text files to the server if you have R/W perms). The parameters file (ATTRIBS.txt) allows you to fully customize the Attribute Names. The system auto-scales to fit whatever screen size you are using, I've tested on my Android phone (LG V20) up to my 32" monitor. Feel free to add, customize as you like, just please provide credit to me for the hard work I did to get it to this point, and will continue to do as I add more features. The eventual goal is reproducing working functionality of all the Behavior Pad Screens we've seen in the awesome HBO series Westworld.
Image File for Raspberry Pi 4 8Gig with functioning implementation of GPT-2 using Tensorflow 1.13.1
YOLOv2 in PyTorch
tensorflow implementation of 'YOLO : Real-Time Object Detection'