Topic
A parameter-efficient fine-tuning method that trains small low-rank adapter weights layered atop a frozen base model instead of updating all its parameters.
Black Forest Labs' open FLUX.3 Image Klein weights now load in Hugging Face Diffusers for self-hosted image generation and editing. Hugging Face times the distilled checkpoint at about 2 seconds per 1024 x 1024 image on an H100, and The Decoder reports an INT8 build for 8 GB cards.
Reality
vLLM 0.30.0 ignores the per-module rank and alpha patterns in PEFT LoRA adapters, and one test put the error against PEFT at 70 times the unpatterned baseline. The adapter loads without complaint, so only a config check or a side-by-side against PEFT will catch it.
Reality
Jared Palmer's Kev now ships as 0.8B, 4B and 9B variants that score typed answer options over frozen Qwen3.5 weights, and on his own development comparison hosted Jev still scores higher than the largest of them.
Reality
kev packs a document and every typed question into one sequence and reads them in a single forward pass on a Mac. Its author scored four checkpoints against the real Jev on frozen items and published the reads that missed a pre-declared gate.
Reality
Atlassian's own validator did the grading, on 25 app briefs, on the same workstation that trained the adapter. Anyone reusing the pass rate needs briefs that look like those 25.
Reality
Feyn Labs' MultiMatte updates 2.27 percent of SAM 3's weights with a rank-16 LoRA, and every pixel in the output now carries a continuous opacity value. The DIS-VD figures supporting it are Feyn Labs' own.
Reality
A reader's rule said to suspect the test set before the methods, and checking it showed the fine-tune was the arm the easy data flattered most, losing 33 points on the rebuilt set against prompting's 28.
Reality
A custom Qwen 3.5 2B adapter returned a shell command when OpenCode's prompt said September 1st, 2026. The harness was running with auto-approval, and the numbers are one researcher's.
Reality