Follow-up analysis on composition in RNNG and transformers
-
Updated
Jul 4, 2021 - Python
Follow-up analysis on composition in RNNG and transformers
Simple Transformers Fork that supports T5TokenizerFast and umT5
A whatsapp assistant/bot that helps you chat with it and solve query along with answering questions.
Mô hình với nhiệm vụ dịch máy đơn giản English to Vietnamese
PyTorch re-implementation of "Generating Sentences from a Continuous Space" by Bowman et al. (2015), using BART encoder and GRU decoder with beam search.
Exploration on the effect of pooled layer on LLM. This project is conducted in sentiment dataset with three classes.
Finetune LLM using Pytorch Lightning + LoRA via peft / FSDP with Pytorch XLA on TPU and others
A lite zero-shot labeler
Certified Robustness & Uncertainty Quantification for AI using Interval Arithmetic. Supports PyTorch & NumPy.
Semantic website content search app with React frontend and FastAPI backend. Enter a URL and query to find the most relevant content chunks using transformer embeddings, FAISS, and Milvus vector database. Modern UI, easy setup, and fast semantic search.
Tools for merging pretrained speech models (TTS & ASR).
production-ready RAG (Retrieval-Augmented Generation) backend built with FastAPI, SentenceTransformers, and FLAN-T5. It demonstrates a modern approach to question-answering using both semantic search and language generation.
Quantize TinyLlama-1.1B-Chat from PyTorch to CoreML (float16, int8, int4) for efficient on-device inference on iOS 18+.
A Fine-Tuned BERT-Based Language Model for Hate Speech Detection in Telugu & Tenglish
Fork of Hugging Face TRL (Transformer Reinforcement Learning) with custom online DPO experiments, GRPO training scripts, evaluation tooling, and mobile-oriented deployment workflows.
A simple Python FastAPI-based backend for our AI Photoshop app
Système automatisé d'analyse et de visualisation des questions parlementaires françaises avec classification ML et dashboard BI interactif.
A FastAPI-based REST API service that provides OCR (Optical Character Recognition) for images, RAG (Retrieval-Augmented Generation) for PDF documents, and Speech-to-Text conversion for audio files. Built with PaddleOCR, LangChain, OpenAI Whisper, Qdrant vector search, and asynchronous job processing via Dramatiq and Redis.
Implementation of the Transformer model (as in the "Attention is All You Need" paper) in PyTorch
Add a description, image, and links to the transformers topic page so that developers can more easily learn about it.
To associate your repository with the transformers topic, visit your repo's landing page and select "manage topics."