LongSpec: Long-Context Lossless Speculative Decoding with Efficient Drafting and Verification

[COLM 2025] Official PyTorch implementation of "Quantization Hurts Reasoning? An Empirical Study on Quantized Reasoning Models"

Adaptive Parallel PDF Parsing and Resource Scaling Engine

Meta-Repository for Bespoke Silicon Group's Manycore Architecture (A.K.A HammerBlade)

Experiments in Joint Embedding Predictive Architectures (JEPAs).

Python implementation of the TMscore program

PyTorchSim is a Comprehensive, Fast, and Accurate NPU Simulation Framework

Official PyTorch implementation of "Denoising MCMC for Accelerating Diffusion-Based Generative Models", ICML 2023 Oral Paper

A simple Python script for running LLMs on Intel's Neural Processing Units (NPUs)

A high-throughput and memory-efficient inference and serving engine for LLMs

Created and enhanced a local LLM training system on Apple Silicon with MLX and Metal API, overcoming the absence of CUDA support. Fine-tuned the Llama3 model on 16 GPUs for streamlined solution of …

KV cache compression via sparse coding

[ISCA 2025] Official Implementation of "MicroScopiQ: Accelerating Foundational Models through Outlier-Aware Microscaling Quantization"