8 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
admin	e37c5aed40	probe(evidence): round two - W8A8 int8-dynamic passes the structure gate at BF16-parity PSNR; verdict flips to YES; upstream Marlin report draft staged	2026-10-08 09:06:49 -04:00
admin	704b8e29ae	probe(round two): W8A8 int8-dynamic scheme — kernel isolation, raw-op probe, artifact builder support	2026-10-08 08:58:45 -04:00
admin	bf65a41ceb	probe(evidence): int8-wo capacity matrix, structure gate, PSNR, Marlin 8-bit NaN isolation (issue 5)	2026-10-08 08:00:51 -04:00
admin	74b596cd05	probe: artifact round-trip verifier (unpack+dequant vs sealed BF16)	2026-10-08 07:45:20 -04:00
admin	67e177860a	probe: track issue-5 fixtures.json (ignore exception, matches issue-2 kit)	2026-10-08 07:39:55 -04:00
admin	321e8bb47f	fix(quant): load offline-quantized compressed-tensors checkpoints directly on CPU under offload ... The diffusion loader treated any quant config without the MXFP8/ is_checkpoint_quantized markers as online quantization: it loaded every component on the accelerator before offloading, so serving the int8 weight-only compressed-tensors artifact OOM'd at model construction (encoder 17 GiB + VAE + packed DiT skeleton > 24 GiB). compressed-tensors is always a serialized format with no on-load quantization pass — detect it (through component wrappers too) and load on the host, round-tripping each quantized module through the accelerator only for the post-load Marlin repack, which has no CPU path.	2026-10-08 07:36:01 -04:00
admin	e778d21f91	probe: int8 weight-only compressed-tensors DiT artifact kit (issue 5) ... quantize_dit_int8wo.py builds the offline W8A16 artifact (per-channel symmetric int8, pack-quantized via compressed-tensors' own packer, per-component config wrapper so encoder/VAE stay BF16, digest manifest); serve.sh/probe.py/gate.py/psnr.py adapt the issue-2 kit for the structure-gated capacity matrix (1-4 refs x seeds 2301-2304).	2026-10-08 07:01:55 -04:00
admin	2f8363c5fc	fix(quant): resolve hyphenated registry methods and disk-dict from_config in quant factory ... _normalize_method_name canonicalizes dashes to underscores for _OVERRIDES lookup, but vLLM's QUANTIZATION_METHODS keeps historical spellings ('compressed-tensors'), so both --quantization compressed-tensors and checkpoint auto-detection raised 'Unknown quantization method'. Resolve the normalized name back to the registry spelling, and fall back to from_config for registry configs whose __init__ does not take the serialized schema (compressed-tensors config_groups/format/ignore).	2026-10-08 07:00:05 -04:00