pytorch

PyTorch toolkit for streaming speech recognition, speech translation and simultaneous translation based on fairseq.

Training-free KV cache compression via E8 lattice VQ. 2-bit KV that preserves retrieval (30/30 NIAH vs TurboQuant 0/30). Calibration-free, 9 architectures validated.

Unified KV-cache compression for LLM inference: 12 Python-native methods, guarded add-on composition and routing, analytical capacity simulation, Godzilla KVarN/TriAttention with SM86/SM89 qualification, CUDA weight sharing, and multi-GPU planning.

AnimeGAN2 trained on Arcane

Detecting Fake News using AI

Chat interface for OpenBB based on a instruct finetuning of a LLM, similar to ChatGPT.

Empowering Data Driven insights through hands-on projects, SQL challenges and practical tools.

🚁 Can Vision-Language Models Think from the Sky? UAVReason for Aerial Reasoning and Generation

[ICRA 2023] Official implementation of the paper "ViPFormer: Efficient Vision-and-Pointcloud Transformer for Unsupervised Pointcloud Understanding"

Consumer-aware compression for embedding indexes and LLM KV caches — compress by the metric the downstream consumer actually uses. PCA-Matryoshka + TurboQuant (27x @ 99.8% recall@10), asymmetric K/V, CUDA/Triton kernels, vLLM plugin, replayable CI-gated claims. MIT.

Implementation for the different ML tasks on Kaggle platform with GPUs.

OCR system for handwritten medical prescriptions using Donut transformer and zero-shot classification

Lightweight representation engineering dataflow operations for agent developers.

MultiModal Sentiment Analysis (Text and Audio) (Pytorch)

This repository contains an attempt to incorporate Rasa Chatbot with state-of-the-art ASR (Automatic Speech Recognition) and TTS (Text-to-Speech) models directly without the need of running additional servers or socket connections.

Simple Python library, distributed via binary wheels with few direct dependencies, for easily using wav2vec 2.0 models for speech recognition

🥉 Codalab-Microsoft-COCO-Image-Captioning-Challenge 3rd place solution(06.30.21)

Local CLI for Google's TranslateGemma translation models with multi-platform support (MLX for Apple Silicon, PyTorch for CUDA/CPU).

Text simplification for a better world: Deep-Martin Transformer 🤗

Exploring NLP weak supervision approaches to train text classification models. The project is also a prototype for a semi-automated text data labelling platform. Approaches: Snorkel and Zero-Shot Learning.