Fast and accurate multilingual translations, all available offline for enhanced privacy and accessibility.

Advanced Speech Emotion Recognition, based on ExHuBERT: Enhancing HuBERT Through Block Extension and Fine-Tuning on 37 Emotion Datasets and 14 languages (Emotions: Disgust, Neutral, Kind, Anger, Su…

Turkish text to speech model that the part of Facebook's Massively Multilingual Speech

Speech synthesis (TTS) in low-resource languages by training from scratch with Fastpitch and fine-tuning with HifiGan

Image captioning with a benchmark of CNN-based encoder and GRU-based inject-type (init-inject, pre-inject, par-inject) and merge decoder architectures

NVIDIA GPU Cloud setup and building NVIDIA Containers for Jetson and JetPack

Post-training quantization on Nvidia Nemo ASR model

Verifying the identity of a person from characteristics of the voice independent from language via NVIDIA NeMo models (ECAPA-TDNN, SpeakerNet, TitaNet-L).

Extractive Question-Answering with BERT on SQuAD v2.0 (Stanford Question Answering Dataset) using NVIDIA PyTorch Lightning

Create conda environment and launch jupyter notebook in Anaconda docker container

Custom object detection on a video dataset using PyTorch Faster RCNN

YOLO Darknet: Traffic sign detection on image and video

Classifying custom image datasets by creating Convolutional Neural Networks and Residual Networks from scratch with PyTorch

Transfer learning using Inception V3 for custom image classification dataset with TensorFlow and Keras

Create light-weight conda environment for ARM64 devices

Download speech datasets (English and non-English) for Automatic Speech Recognition

Determining the probability of a sequence of words in Turkish using the KenLM language model with Python

Automatic image captioning on Android-based mobile application with CNN and multi-layer GRU encoder-decoder model

InceptionV3-Multi-layer GRU based automatic image captioning with Keras and TensorFlow frameworks

Package for cleaning and preprocessing text data is supported for all languages (some functions) and English (all fuctions).

Compute the MFCCs and measure (dis)similarity between two audio files using DTW