NVIDIA releases the full Run:ai platform as open source
The release includes GPU workload scheduling, resource management and GPU-sharing components for Kubernetes.
Article subject labels can be suggestions and are not editorial confirmation.
The release includes GPU workload scheduling, resource management and GPU-sharing components for Kubernetes.
IBM said Mashreq is piloting its Bob platform for software integration and production troubleshooting.
The preview integration adds centralized visibility and gateway policy enforcement for MCP servers and tools.
Meta released weights for Muse Spark Open, a multimodal model it describes as having 816 billion total parameters and about 42 billion active per token.
In a development benchmark reported by NVIDIA, Siemens EDA cut agent turns from 155 to 53 while raising the composite verification score from about 55% to 95%.
Llama Windows is an open-source app for downloading open models from the Windows 11 tray and running them locally with llama.cpp.
NVIDIA reported that a Spark UDF rewrite cut operator time from 274.55 hours to 12.78 hours on one 600-billion-row workload at an unnamed consumer technology company.
NVIDIA said staged training reduced Nemotron ASR's word error rate from 77% to 22.46% on a two-hour ATCO2 test set.
Researchers said ReplaySSM, which buffers state-space-model inputs, made decoding kernels 1.46–1.80 times faster across four models at batch size 256.
AlphaEdge AI released Sparseup-fr, a French adaptation of Linkup's open-source method, together with a cleaned NanoBEIR-fr benchmark.
Global Emotions contains 525 hours of speech recorded by 10,420 speakers across 83 languages and 108 locales; the company made the dataset available for commercial and research use.
Ai2 said teams received 98% of the GPU hours they were owed during a 30-day evaluation of a new budgeting and fair-share system spanning thousands of GPUs.