Michael Feil

infinity. Infinity is a high-throughput, low-latency serving engine for text-embeddings, reranking models, clip, clap and colpali

embed. A stable, fast and easy-to-use inference library with a focus on a sync-to-async API

hf-hub-ctranslate2. Connecting Transformers on HuggingFace Hub with CTranslate2

skyjo_rl. Multi-Agent Reinforcement Learning Environment for the card game SkyJo, compatible with PettingZoo and RLLIB

candle-flash-attn-v3. C++

commonroad_motionplaner_michaelf. Winning 2020 solution for the commonroad.io contest

radix-mlp. Resources for RadixMLP

iot_gateway_modbus. A MQTT Gateway connecting Modbus RTU and Google IoT Core

candle-index-select-cu. Rust

web-tfjs. An Tensorflow.js Web App running MobileNet on ImageNet in the Github Pages & my personal website.

LSTM-DLNN2020. Character-LSTM implementation in Numpy

lm-evaluation-harness. A framework for few-shot evaluation of language models.

sentence-transformers. Multilingual Sentence & Image Embeddings with BERT

llama-recipes. Scripts for fine-tuning Meta Llama3 with composable FSDP & PEFT methods to cover single/multi-node GPUs. Supports default & custom datasets for applications such as summarization and Q&A. Supporting a number of candid inference solutions such as HF TGI, VLLM for local or cloud deployment. Demo apps to showcase Meta Llama3 for WhatsApp & Messenger.

gpt-fast. Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

dynamo. A Datacenter Scale Distributed Inference Serving Framework

cognita. RAG (Retrieval Augmented Generation) Framework for building modular, open source applications for production by TrueFoundry

tokenizers. 💥 Fast State-of-the-Art Tokenizers optimized for Research and Production

bigcode-evaluation-harness. A framework for the evaluation of autoregressive code generation language models.

flash-deberta. Deberta, but Flash