Every dequant shape and block geometry comes from a real transformer config,
one source of truth per model.
- flux.1: attention linears 3072x3072, feed-forward 12288x3072, block at
3072 wide, 24 heads, 12288 ff, 4608 joint tokens
- krea 2: standalone wq 6144x6144, swiglu gate 16384x6144, block at 6144
wide, 48 heads, 4608 joint tokens
- the block section measures every geometry in block_geometries; buyback
costs are judged at the block matching the reference shape, with the
geometry named in the reason
- te shape lookup keys on its label instead of a list index
Krea 2 runs joint attention over one text plus image stream with a segment
mask: text is padded to a fixed 512 tokens and the padded tail is masked for
queries and keys both, so fully masked query rows yield nan under sdpa. The
preset carries the transformer's exact (B, 1, L, L) mask, nan-guards the
error path the way the model does, and uses the 48 kernel-level heads left
after gqa expansion at 4608 joint tokens.
- fix the drift fields crashing when save_report rebuilt the run info
Per-preset include lists trimmed video and masked shapes for runtime, which
left the composition check without its smooth_hadamard row there. Only hard
technical exclusions remain, as an exclusion map: sd15 skips hadamard configs
because compiling hadamard with a non pow2 head dim hangs torch inductor.
- wan22, wan22-cfg, ltx2 and masked gain smooth_hadamard, fp16qk, fp8qk and
fp8full rows; availability gates and the per-config timeout still apply
Repeat-pair runs measured 2-4% between-run drift on one machine, enough to
flip threshold verdicts near the margin on every rerun.
- bench keeps its iteration samples: rows carry a median sigma, and per-shape
sentinel re-measurements sample run-level clock drift
- on/off verdicts are three-zone at the run's own noise level; too close to
the margin keeps the current setting and says so
- pv candidates (now including fp16) are tested independently against the
margin at a sidak-adjusted z instead of min-then-threshold
- unmeasured toggle stacks are estimated additively in the composition check
- per-shape qk verdicts print alongside the reference-shape verdict
- cross-gpu error sanity bands flag corrupted measurements
- on/off verdicts share one speed margin (recommend_speed_margin, 10%) across
attention, dequant, compile, te and conv rows
- re-check the recommended toggle stack against unquantized; individually passing
buybacks can eat a marginal qk gain
- judge smooth k and hadamard cost at block scope when measured: kernel rows hand
the prep contiguous q/k/v, real models hand it strided views from the fused qkv
projection; reasons cite both scopes
- give bare float8 qk its own shot at the margin before disabling, the verdict
must not hinge on int8 alone
- compare triton flash against the recommended config, not always int8
- note self-attention shapes that disagree with the reference verdict
- fold the enable rows into the type rows: disabled, enabled or an explicit dtype
- pv keep-unquantized value is disabled; enabled means int8 pv
- add a text encoder row judged at te geometry
- update locale hints, drop the orphaned checkbox hint
The native loader entry log repeated the name and full file path already
printed one line earlier by network_load. Remove it and fold cache-hit
status into the network_load announce line, so a native load emits one
starting line plus the result line instead of three with a duplicated
path.
The Krea 2 transformer keeps checkpoint-style module names while the
official krea/Krea-2-LoRA releases are saved with upstream-diffusers
names, so all 264 modules failed to bind and the LoRAs silently did
nothing. Krea 2 is the only native-LoRA arch with an sdnext-owned
transformer, so its module names diverge from the diffusers ecosystem.
- native_adapter.resolve_group_targets consults the arch resolver first
for passthrough prefixes, falling back to verbatim binding; a no-op
for arches that load the diffusers class
- krea2_lora maps diffusers attn/ff/text_fusion/embedder names onto the
checkpoint module tree; checkpoint-named LoRAs still bind verbatim
- add test/test-krea2-native-adapters.py
Qwen3.5 runs most of its layers as gated delta linear attention. Without
flash-linear-attention, transformers falls back to a per-token torch loop that
runs sequentially over prefill and decode, so a caption takes minutes with no
indication of why.
ToriiGate 0.5 is a Qwen3.5 vision fine-tune trained on a single system prompt
and a single query structure, and it degrades on anything else. The shared qwen
handler strips angle brackets and underscores from the question, which mangles
the model's format templates, and the generic caption instructions are not what
it was trained on.
- hold the model's caption formats, system prompt and query builder in modules/caption/toriigate.py
- build the system prompt and user query from those templates in the qwen handler
- offer the native formats in the task dropdown and in the caption api prompt groups
- drop Normal Caption for this model, which has no format between short and long
"sage" is a different kernel per gpu: sageattention dispatches the pv dtype and
accumulator by arch and cuda version, and modules/attention.py forces the fp16-pv
cuda kernel on sm86, so one label across shared reports compares unlike kernels.
- resolve the row label from the dispatch: sm89 with cuda 12.8+ reads
"sage int8 qk + fp8 pv, fp32+fp16 accum", sm86 reads "fp16 pv, fp32 accum"
- name the sm86-only fp16-accum baseline for what it is
Notes ran to paragraph length and wrapped in the terminal, burying the numbers.
- lead with the topic, keep the numbers and the fix, drop the restatement
- one line per note at 140 columns, except the sd15 compile warning
sdnq no longer skips compile for fp8 storage, it upcasts e4m3 to the scale dtype
before the compiled dequant, so the tool's mirror of the deleted skip_fp8_compile
gate went dead and fed the raw e4m3 weight to its compiled dequant: on pre-ada gpus
that reports a compile failure for rows the webui runs fine.
- mirror the upcast in dequant_args, keep the hardware probe on the raw weight
- drop dequant_gated_to_eager and the gated-to-eager row, fp8 rows are measured now
- say upcast, not eager fallback, in the environment panel and dequant notes
Hiding the sdnext startup log by pointing fds 1 and 2 at a file kills the process
when stdout is a real console: sys.stdout writes through the win32 console api on
the handle behind fd 1, so the first bootstrap log line raises OSError 'the handle
is invalid', and with stderr broken the same way the interpreter aborts with no
message, no traceback and no results.
- swap sys.stdout/sys.stderr to a buffer along with the fd redirect
- capture the startup log instead of discarding it to devnull, replay it on failure
- catch SystemExit: loader.py and installer.py exit on fatal startup errors
A full-weight extraction on Z-Image bound 308 modules and applied 172 of
them, silently dropping the rest, and left 71 more unmapped.
assign_network_names_to_compvis_modules puts every transformer module in
network_layer_mapping but skips stamping network_layer_name on norms,
which is the attribute the apply pass keys off. try_load_full bound those
modules through the mapping and they then never applied; stamp them
loader-locally, as try_load_norm already does.
Z-Image also names three module groups differently from the diffusers
tree: the qk-norms (q_norm/k_norm vs norm_q/norm_k), and the patch
embedder and final layer, which live in ModuleDicts keyed by
"{patch_size}-{f_patch_size}" and so carry a key the checkpoint has no
notion of. Read that key from the live model rather than hardcoding it.
The counts close exactly: 68 qk-norms plus 3 non-block targets are the 71
that went unmapped.
The linear layers bind their scaled-mm function at import, so
SDNQ_USE_TRITON_MM freezes the backend per process and comparing triton
against the torch fallback meant two runs, carrying clock drift into the
delta. --mm-backends rebinds the function on the consuming modules
between benches, so both rows measure the same quantized layer under the
same clock state.
- swap targets are the four linear modules that from-import a scaled-mm
function; every swap resets dynamo, since the layer forwards are
compiled and the traced graph pins the previous function
- the torch row is captured from kernel_wrappers rather than
reimplemented, so it is unavailable where triton is the device default
(rerun with SDNQ_USE_TRITON_MM=0) and says so
- rounds alternate order and keep the fastest per row, so drift cancels
instead of favouring whichever backend runs second
- flag output error when backends disagree past 1e-4: they are meant to
be numerically equivalent
probe_fp8 and the broken-compile bench path forced eager input prep by
toggling torch._dynamo.config.disable. torch 2.13+ raises 'found no
compiled frames' when a fullgraph-compiled function is called inside a
disable window, so the probe failed before reaching the attention kernel
and reported an environment error instead of the kernel's real verdict,
on fp8-capable gpus too. Swap the module-global prep function for its
eager inner instead, matching the correctness section.
Vladimir Mandic
CalamitousFelicitousness
Disty0
Oleksandr Liutyi