4GitHub Trending (AI/LLM filtered)·
NVIDIA Model-Optimizer: unified library for quantization, pruning, distillation, and inference optimization
NVIDIA's Model-Optimizer is an open-source Python library consolidating state-of-the-art model compression techniques including quantization, distillation, pruning, neural architecture search, and speculative decoding. It targets downstream deployment on TensorRT-LLM, TensorRT, and vLLM to improve inference speed. The repository has accumulated 3,286 stars with modest recent activity (+8 today).
Related guides (3)
Related events (8)
Optimum-NVIDIA: One-Line LLM Inference Acceleration via TensorRT-LLM
Hugging Face's Optimum-NVIDIA integration wraps NVIDIA's TensorRT-LLM backend to enable high-performance LLM inference with minimal code changes. The library targets developers who want near-peak GPU throughput without manually configuring TensorRT-LLM pipelines. It positions as a bridge between the Hugging Face ecosystem and NVIDIA's optimized inference stack.
Introducing Optimum: The Optimization Toolkit for Transformers at Scale
Hugging Face announced Optimum, an optimization toolkit designed to accelerate Transformers models on various hardware backends. The toolkit aims to bridge the gap between Transformers model development and hardware-specific optimizations from partners. It provides a unified interface for quantization, pruning, and hardware-accelerated inference across different accelerators.
Optimize and Deploy with Optimum-Intel and OpenVINO GenAI
Hugging Face's Optimum-Intel library integrates with Intel's OpenVINO runtime to enable optimized inference of generative AI models on Intel hardware. The post covers quantization, model export, and deployment workflows using OpenVINO GenAI APIs. This targets edge and CPU-based inference scenarios where reducing model size and latency is critical.
Nvidia Nemotron 3 Ultra: hybrid Mamba-transformer open-weights model targeting agentic workloads
Nvidia released Nemotron 3 Ultra, a 550B parameter (55B active) hybrid Mamba-transformer mixture-of-experts model with a 1M token context window, publishing weights, training data, and RL environments under an open license. The model ranks as the highest-scoring U.S. open-weights model on the Artificial Analysis Intelligence Index (47.7-48.2) and is approximately three times faster than comparable open-weights rivals, though it trails leading Chinese models like Kimi K2.6 and DeepSeek V4 Pro on intelligence benchmarks. Nvidia used a novel Multi-Teacher On-Policy Distillation approach with 10+ specialized teacher models and trained using NVFP4 quantization. The release is strategically motivated by Nvidia's interest in a healthy open-weights ecosystem that drives AI semiconductor adoption.
Optimization story: Bloom inference
This Hugging Face blog post documents practical inference optimization techniques applied to the BLOOM large language model. It covers strategies for reducing latency and memory footprint during deployment, likely including quantization, tensor parallelism, and batching approaches. The post serves as a technical case study for serving very large open-weights models efficiently.
Unsloth: Web UI and Library for Efficient Fine-tuning of Open Models
Unsloth is an open-source Python library and web UI (Unsloth Studio) for efficient fine-tuning and local inference of open-weight models including Gemma 4, Qwen3, DeepSeek, and GPT-OSS variants. The project has accumulated over 64,000 GitHub stars with continued daily growth (+139 today), indicating strong community adoption. It targets practitioners who want to train and run large models locally with reduced memory and compute requirements.
Accelerate your models with Optimum Intel and OpenVINO
Hugging Face's Optimum Intel library integrates with Intel's OpenVINO toolkit to accelerate inference of transformer models on Intel hardware. The post covers how to export models to OpenVINO IR format and run optimized inference pipelines. This targets deployment efficiency for NLP and vision models on CPU and other Intel accelerators.
Nvidia releases Nemotron 3 Super 120B-A12B open-weights model with hybrid Mamba-2/MoE architecture
Nvidia released Nemotron 3 Super 120B-A12B, an open-weights LLM with a hybrid Mamba-2/transformer/MoE architecture that activates only 12B parameters per token and supports up to 1 million token context. The model claims the fastest inference speed in its size class at 442 tokens/second and leads open-weights models on PinchBench agentic task evaluation, outperforming larger models including Kimi K2.5 (1T parameters). Nvidia is releasing weights, training data, and recipes under a permissive commercial license, and plans a $26B five-year investment in open-weights models — framed partly as a strategic response to Chinese labs building capable open-weights models on non-Nvidia hardware.