End-to-end ASR with Wav2Vec2 + CTC — 4.15% WER on LibriSpeech test set, beam search decoding, warmup + cosine LR scheduling

Connectionist Temporal Classification (CTC) decoder with dictionary and language model for TensorFlow.

ASR-LE is an advanced ASR evaluation + observability toolkit that goes beyond WER: it shows where errors happen in time, estimates streaming p95 first-word latency, generates “worst moments” automatically, and produces reusable artifacts (report.json, timeline bins, moments, etc.)

TensorFlowASR: Almost State-of-the-art Automatic Speech Recognition in Tensorflow 2. Supported languages that can use characters or subwords

Convert handwriting images into text (OCR)

The goal of this project is to accurately transcribe handwritten English word images into text using a deep learning model. This is achieved through the use of: The IAM Handwriting Dataset is a labeled image corpus CNN + BLSTM + CTC architecture An end-to-end training and inference pipeline

Neural network trained to regonize short speech commands

a test application to evaluate coverage tools

Connectionist Temporal Classification (CTC) decoding algorithms: best path, prefix search, beam search and token passing. Implemented in Python and OpenCL.

CTF writeups and proof files from TCS HackQuest Season 10 cybersecurity challenges.

Sistema OCR con Deep Learning (CRNN + CTC) entrenado con 100k imágenes sintéticas, incluye app web con Gradio

Small CRNN that solves the Ikariam pirate fortress captcha at 97.3% accuracy. ONNX-ready.

Tensorflow-based CNN+LSTM trained with CTC-loss for OCR

Korean jamo-level STT for dysarthria assessment support with DeepSpeech2, Simple-Attention, and Transformer model comparison.

An Arduino Nano based fan spoofer

A small model for recognition of digits in audio clips

"LipNet: End-to-End Sentence-level Lipreading" in PyTorch

This python script will help you to get a sense of what you will get in hand salary for a given CTC

Implementation of Conformer CTC ASR model