AI Infrastructure

Auto Added by WPeMatico

AMD Releases Instella-MoE-16B-A3B: A Fully Open Mixture-of-Experts LLM With 2.8B Active Parameters Trained On Instinct GPUs

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Applications, Artificial Intelligence, Editors Pick, Language Model, Large Language Model, Machine Learning, New Releases, Open Source, Python, Tech News, Technology

AMD released Instella-MoE-16B-A3B, a fully open Mixture-of-Experts language model trained from scratch on Instinct MI300X and MI325X GPUs. The model holds 16B total parameters but activates only 2.8B per token. AMD is publishing weights from every training stage, along with data mixtures, training configs, and inference code. Two systems-level choices carry the release: Gated Multi-head […]

AMD Releases Instella-MoE-16B-A3B: A Fully Open Mixture-of-Experts LLM With 2.8B Active Parameters Trained On Instinct GPUs Read More »

Accelerating Transformer Training with NVIDIA Transformer Engine, Fused Kernels, BF16, FP8, and GPU Benchmarking

ai, AI (Artificial Intelligence), AI Infrastructure, Artificial Intelligence, Editors Pick, Machine Learning, Staff, Technology, Tutorials

In this tutorial, we explore how NVIDIA Transformer Engine accelerates transformer workloads by combining fused GPU kernels, BF16 computation, and hardware-aware FP8 execution. We begin by installing Transformer Engine and detecting the active GPU architecture so that we can determine whether the runtime supports TE kernels, FP8 tensor cores, or only the pure-PyTorch fallback path.

Accelerating Transformer Training with NVIDIA Transformer Engine, Fused Kernels, BF16, FP8, and GPU Benchmarking Read More »

DeepSeek Upgrades DeepSeek-V4-Flash-0731 with Major Agentic and Coding Gains

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Paper Summary, AI Shorts, Applications, Artificial Intelligence, Editors Pick, Language Model, Large Language Model, Machine Learning, New Releases, Staff, Tech News, Technology

DeepSeek published DeepSeek-V4-Flash-0731 on Hugging Face and moved the official V4-Flash API into public beta on July 31, 2026. The model card is explicit that this is the official release superseding the preview, and that the architecture and size are unchanged. The gains come from re-post-training, not a new design. The checkpoint ships with the

DeepSeek Upgrades DeepSeek-V4-Flash-0731 with Major Agentic and Coding Gains Read More »

JetBrains Open-Sources KotlinLLM: Smart Macros That Generate Kotlin Source Code at Runtime and Hot-Reload It Through JDI

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Applications, Artificial Intelligence, Editors Pick, New Releases, Open Source, Software engineering, Staff, Tech News, Technology

JetBrains Research Open-Sources KotlinLLM. KotlinLLM is an IntelliJ IDEA plugin for Kotlin/JVM projects that adds a language feature called Smart macros. A Smart macro is a regular Kotlin function call whose body is generated Kotlin code. The public API is deliberately small. asLlm<F, T>(from, hint) converts an input of type F into a typed value

JetBrains Open-Sources KotlinLLM: Smart Macros That Generate Kotlin Source Code at Runtime and Hot-Reload It Through JDI Read More »

Google DeepMind Ships Three Physical AI Models For Whole Body Control, Dexterity And Multi Robot Collaboration

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, Applications, Artificial Intelligence, Editors Pick, Language Model, Large Language Model, New Releases, physical ai, Robotics, Staff, Tech News, Technology, Vision Language Model

Google DeepMind has released Gemini Robotics 2, the intelligence layer for its next generation of robots. The release moves the stack past table-top manipulation into whole body control, five finger dexterity and multi robot teamwork. It ships as three separate models with three different access tiers. Most robots today are pre-programmed or tele-operated for narrow,

Google DeepMind Ships Three Physical AI Models For Whole Body Control, Dexterity And Multi Robot Collaboration Read More »

Tencent Open-Sources AngelSpec: A Unified Training Framework for MTP and Block-Parallel Speculative Decoding on Hy3 Models

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Paper Summary, AI Shorts, Artificial Intelligence, Editors Pick, Language Model, New Releases, Open Source, Software engineering, Staff, Tech News, Technology, Uncategorized

Tencent has released AngelSpec, an open-source, torch-native training framework for speculative-decoding draft models. The release covers both autoregressive multi-token prediction (MTP) and the block-parallel DFlash family. Most speculative-decoding work searches for one drafter that scores well on an averaged benchmark mixture. Real serving traffic does not look like that mixture. AngelSpec treats workload heterogeneity as

Tencent Open-Sources AngelSpec: A Unified Training Framework for MTP and Block-Parallel Speculative Decoding on Hy3 Models Read More »

Moonshot AI Open-Sources MoonEP: A Perfectly Balanced Expert Parallelism Library for MoE Training

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Applications, Artificial Intelligence, Editors Pick, Open Source, Software engineering, Staff, Tech News, Technology

Moonshot AI has open-sourced MoonEP, an Expert Parallelism (EP) communication library for distributed Mixture-of-Experts (MoE) workloads. The team announced the release as a library built to make expert-parallel communication more efficient at scale. It ships under an MIT license. MoonEP arrived as part of Kimi K3 Open Day. Alongside the K3 model weights and technical

Moonshot AI Open-Sources MoonEP: A Perfectly Balanced Expert Parallelism Library for MoE Training Read More »

Liquid AI Releases LFM2.5-Encoder-230M and LFM2.5-Encoder-350M: Bidirectional Encoders That Stay Fast at 8K Context on CPU

Liquid AI has released two open-weight bidirectional encoders, LFM2.5-Encoder-230M and LFM2.5-Encoder-350M. Both are masked language models built on the LFM2 hybrid backbone. Both carry an 8,192-token context. Encoders sit underneath classifiers, intent routers, safety filters, and PII detectors. Those jobs run continuously, usually without a GPU, and increasingly on longer inputs. BERT established the class.

Liquid AI Releases LFM2.5-Encoder-230M and LFM2.5-Encoder-350M: Bidirectional Encoders That Stay Fast at 8K Context on CPU Read More »

Fireworks AI Releases Fireworks Nexus: A Drop-In Routing and Cost-Control Layer That Moves Routine Coding Work to Open-Weight Models

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, AI Shorts, Applications, Artificial Intelligence, Editors Pick, Language Model, New Releases, Software engineering, Tech News, Technology

Fireworks AI has released Fireworks Nexus, an AI management and routing platform aimed at engineering organizations. It connects the coding tools developers already use to a managed layer of open-weight models. The problem it targets is well documented. Forbes reported that Uber exhausted its entire 2026 AI budget in four months. Claude Code had reached

Fireworks AI Releases Fireworks Nexus: A Drop-In Routing and Cost-Control Layer That Moves Routine Coding Work to Open-Weight Models Read More »

Deploying a 1-Bit Bonsai-27B Model with PrismML llama.cpp and OpenAI-Compatible Local Inference Workflows

agentic ai, ai, AI (Artificial Intelligence), AI Infrastructure, Applications, Artificial Intelligence, Editors Pick, Technology, Tutorials

In this tutorial, we deploy the 1-bit Bonsai-27B language model using the PrismML fork of llama.cpp, which provides the specialized CUDA kernels required to decode the model’s Q1_0_g128 GGUF quantization format. We begin by validating the GPU runtime, installing the required Python dependencies, compiling the CUDA-enabled inference binaries, and downloading the compressed model weights from

Deploying a 1-Bit Bonsai-27B Model with PrismML llama.cpp and OpenAI-Compatible Local Inference Workflows Read More »