Card: shm_broadcast hang triage — pidfd_getfd/seccomp/Yama cause from the offload PR thread; Marlin note scoped to this image

Card: defaults failures reproduced on our box; meta-device overlay revision shipped, sysctl now only for the BF16 disk path

Overlay: build the n-gram table parameter on the meta device when VLLM_PLE_QUANT_DIR is set, and swap the stub Parameter instead of set_data — removes the 102 GB virtual reservation that the kernel's overcommit heuristic refuses on hosts with less RAM+swap than the table (field report, 64 GB host); validated under an emulated 67/99 GiB commit limit, sanity PASS, tool-calling 77.0 (n=3), 81 tok/s c1

Overlay: build the n-gram table parameter on the meta device when VLLM_PLE_QUANT_DIR is set, and swap the stub Parameter instead of set_data — removes the 102 GB virtual reservation that the kernel's overcommit heuristic refuses on hosts with less RAM+swap than the table (field report, 64 GB host); validated under an emulated 67/99 GiB commit limit, sanity PASS, tool-calling 77.0 (n=3), 81 tok/s c1

Card: field report (64 GB host) — pin --max-model-len/--max-num-seqs in the serve blocks, vm.overcommit_memory note, MTP+xgrammar caveat

Card: week-one audit (2026-09-03) — upstream status, dated field table, PR-branch wording for the image race fix

Upload README.md with huggingface_hub

ship the async-MRV2 connector fix as an overlay (upstream 4e8b849)

Add Primitive brand banner, logo and card thumbnail

Upload folder using huggingface_hub

Upload README.md with huggingface_hub

Upload worker_image_quant.py with huggingface_hub

Upload folder using huggingface_hub

Upload folder using huggingface_hub

Upload README.md with huggingface_hub

Upload ple_layer_quant.py with huggingface_hub

Upload worker_image_quant.py with huggingface_hub

initial commit