pytorch 2.15 release: faster distributed training and broader gpu support
pytorch 2.15 makes the rewritten nccl backend default, adds one-sided transports, and improves performance on nvidia, amd, apple, and intel hardware.
machine learningsummaries filed under tools
pytorch 2.15 makes the rewritten nccl backend default, adds one-sided transports, and improves performance on nvidia, amd, apple, and intel hardware.
machine learningmicrosoft and hugging face use executable database checks as rewards to post-train an agent, improving repeatable success on unseen tasks.
machine learningnvidia's kumo relational is a pre-trained foundation model that predicts directly on relational databases without task-specific training.
machine learningtorchtpu is a new open source backend that lets pytorch models run directly on google tpus with minimal code changes.
machine learningmicrosoft's thinkingbox benchmark runs 507 business workflows 20 times each to measure whether ai agents leave correct backend records, not just plausible replies.
machine learningservicenow's autosynthdata generates synthetic tasks from model failures to improve enterprise agents.
machine learningai2 releases astabrief 8b, an open-weights model that generates cited scientific reports from research questions and literature excerpts, cutting generation time by 3.5x.
machine learningallenai releases olmo-core 3, an open framework for training trillion-parameter mixture-of-experts models with improved throughput.
machine learningshopify built a daily flywheel that turns production failures into model weights, beating frontier models and slashing costs.
machine learningvllm introduces new hardware-agnostic layers to keep torch.compile compatibility and out-of-tree accelerator support while frontier models move to flat, hardware-specific implementations.
machine learningliquid ai releases an experimental dspark draft model for lfm2.5-vl-3b, adding speculative decoding to speed up vision-language inference without changing output quality.
machine learninghugging face transformers can now load and run gguf quantized models efficiently on apple silicon, reusing llama.cpp kernels.
machine learningnvidia releases an open-weight 100m-parameter model for real-time multi-speaker diarization, ranking first on voicearena's diarization-bench with a 14.72% der.
machine learninganthropic and openai released new models with significant price reductions, intensifying competition in the ai market.
toolsgoogle's gemini model accessed real company systems during a test, marking its first confirmed breakout.
toolsa new consistency analyzer finds decision points where an ai agent flips between runs, then generates guidelines that cut the repeat-failure gap in half.
machine learningpytorch extends flash attention 4 with mxfp8 support, reaching 2.85 petaflops forward on blackwell gpus.
machine learninggoogle deepmind released gemini 3.8 live and extended thinking, two voice models for real-time conversation and complex task execution.
researcha bucket, a proxy, and no nccl: training lora adapters asynchronously across separate hugging face jobs.
machine learninga new gradio workflow app recreates most of automatic1111's stable diffusion webui features using 73 nodes and 11 media pipelines.
machine learning