Four open supply tasks dominate LLM fine-tuning at present. Unsloth, Axolotl, TRL, and LLaMA-Factory all wrap the identical underlying PyTorch and Hugging Face stack. They diverge on the place they spend engineering effort. Unsloth rewrites kernels. Axolotl composes parallelism methods. TRL defines the coach APIs the others construct on. LLaMA-Factory optimizes for breadth of mannequin…
Poolside has launched Laguna S 2.1, a 118B-parameter open-weight mannequin constructed for agentic coding. It is a Mixture-of-Experts (MoE) mannequin with 8B activated parameters per token. It helps a context window of as much as 1M tokens in each pondering and no-thinking modes. The weights are on Hugging Face beneath an OpenMDW-1.1 license, and the…
In this tutorial, we discover NVIDIA’s srt-slurm framework and find out how we use srtctl to transform declarative YAML configurations into reproducible SLURM benchmark workflows for distributed LLM serving. We arrange the undertaking in Google Colab, examine its inside structure, outline a cluster configuration, dry-run built-in and customized recipes, and mannequin a disaggregated prefill-and-decode deployment…
A survey of over 300 CTOs, CISOs, and engineering leaders revealed a widening hole between AI adoption and id safety. 88% say AI has already outpaced their id infrastructure. 84% of probably the most assured organizations have already skilled a confirmed id incident. Shared SaaS id setups get hit twice as arduous. Our newest eBook…
NVIDIA has launched Cosmos 3 Edge, a 4-billion-parameter open world mannequin constructed to run on-device. It helps robots and imaginative and prescient AI brokers perceive environment, purpose in actual time, and generate robotic actions domestically. The Cosmos 3 family included Cosmos 3 Nano (16B) and Cosmos 3 Super (64B) shipped on May 31, 2026 at…
A single 24GB card is the sensible ground for critical native inference. It is sufficient for genuinely succesful fashions, and sufficiently small to sit down on one GPU. An RTX 3090 or RTX 4090 each land in this tier. The card you personal issues lower than the fashions you decide for it. The previous hobbyist…
In this tutorial, we construct an end-to-end NVIDIA NeMo AutoModel workflow in Google Colab and use a single GPU to discover the identical configuration-driven coaching structure that scales to distributed multi-GPU environments. We confirm the accessible CUDA {hardware} and precision assist, set up NeMo AutoModel immediately from its supply repository, load an official Qwen3-0.6B LoRA…
Backpropagation dominates deep studying, but it makes use of a mechanism the mind probably can’t. Specifically, the backward cross wants actual transposes of ahead weight matrices. This is the weight transport drawback. Sakana AI’s new paper, Diffusing Blame, confronts this constraint immediately. The analysis staff trains networks that obey Dale’s precept whereas avoiding weight transport…
Thinking Machines Lab simply launched Inkling, their first mannequin educated from scratch, weights are open, fine-tunable on Tinker. The lab pitches it as a base for personalisation. What is Inkling? Inkling is a Mixture-of-Experts transformer with 975B whole parameters and 41B energetic. It helps a context window of as much as 1M tokens. Pretraining coated…
A German analysis consortium has printed the pretraining report for Soofi S 30B-A3B. It is an open base mannequin for German and English. Training ran finish to finish on Deutsche Telekom’s Industrial AI Cloud in Munich. Preview weights are on Hugging Face. It is value noting that amongst among the totally open base fashions examined,…