Studio has known which words to weight since 2.0 —
_v2_concept_units_for_run splits every run into phrases, single words and
n-grams; _v2_update_phrase_memory trains each against the rating's
missing/wrong axes. The only consumer was _v2_build_attn2_patch, and H3
skips it: one packed self-attention stream, no cross-attention. So the
learning ran on every generation and reached nothing.
It now becomes an attention-logit bias on the prompt's token positions in
the packed stream, which is the same statement the K/V x1.25 emphasis was
making, at a place H3 has. Boost only by default.
Located with the tokenizer's offset mapping (verified against ComfyUI's
H3 vocabulary), not by re-encoding each phrase through Qwen3-VL-32B the
way _v2_find_phrase_token_ranges does.
The bias is a mask and SLA routes masked calls to dense, so a weighted
run is a dense run. Said once, at install.
Qwen never applies (word:1.2): ComfyUI's H3 tokenizer hardcodes every
weight to 1.0 and runs the raw tokenizer over the literal string, so the
syntax reaches the encoder as punctuation inside the sentence and
sd1_clip's weighting arm never runs.
Scaling embeddings would not have worked either — Qwen's rows are
contextual hidden states, not per-word vectors. But H3 is one packed
self-attention stream with the prompt's tokens in it, so log(w) added to
the logits at that phrase's key positions multiplies its share of the
attention mass by w, which is what a weight means.
Module and tests only; nothing is wired to a node yet.
The check was stricter than the merge it was protecting. comfy merges with
mm(mat1.flatten(1), mat2.flatten(1)).reshape(weight.shape), so the element
count is the constraint, not the two dimensions — and a locon `mid` or a
`reshape` entry changes what the delta has to fit at all. Those two now
abstain instead of dropping, and the count is what is compared.
The message carries both shapes. "Trained against a different variant" says
nothing you can act on; "LoRA 5376x5376 into weight 16128x5376" says whether
every layer is off by one factor (a different width, nothing to do) or only
some are (a fusion or naming difference, fixable).
Reverts 6bf533e, 78c9625, 4bacc9c, 993e5e4, ae337b5 and c4d8018 —
gguf_support.py, loaders.py and their tests are byte-identical to
6f04642. Nothing else is touched.
The quantized path loaded fast and gave the sampler a model it could not
run: shape [21504, 5376] against an input of 22848x5376, one 1344-token
unit apart. Expanding at load is slow and works.
ComfyUI-GGUF hands back either the state dict or a (state_dict, arch)
pair depending on version. Passed on whole, ComfyUI iterates the tuple,
gets the dict as its first element and raises "'dict' object has no
attribute 'startswith'" — an error naming nothing that led back here.
Unwrapped, and the shape check now runs on every load rather than only
when FunPack had overridden the architecture list, which is why a pack
that already knew the architecture went through unvalidated.
The expansion cache is gone. It wrote the model's full dequantized size
beside it, and rented disk is billed. A leftover from it is named with
its size on the next load; deleting it is left to the user.
Dequantizing produces float32 — four times the width the weight ends up
at. Every tensor was collected before any was cast, so the peak was the
whole checkpoint in float32 plus a copy of each: tens of GB of transient
host allocation on a video model, enough to take the machine down rather
than just the load. ThreadPoolExecutor.map made it worse by submitting
everything at once and buffering results in order.
Now cast-and-release per tensor, with a fixed in-flight window. Peak is
the finished model plus a handful of tensors. Pool capped at 4.
Also prints the path and tensor count before the wait, so a load killed
part way leaves evidence of which backend it chose.
No env var. Stepping around the pack's guard means doing its job here:
the returned state dict is validated, and anything unexpected raises
into the existing fallback rather than reaching ComfyUI.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A bare "'dict' object has no attribute 'startswith'" names neither the
step nor the file. Each stage now labels itself in the re-raise. The
ComfyUI-GGUF architecture override is opt-in, not default.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Add an unlisted architecture to ComfyUI-GGUF's image list so the
quantized path works, and cache the dequantized weights beside the
file when it does not, so the expansion is paid once.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Attention and SLA are applied to a clone of a one-entry weight cache
keyed on file identity + dtypes, so a sparsity change skips the disk.
GGUF dequantization now runs across threads instead of serially.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
pip install -r upgrades anything below its floor; moving numpy or
transformers under ComfyUI's compiled extensions segfaults hours later.
Absent packages are installed by name; outdated ones are reported with
the command and left alone.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Unreadable modality tags widened the erase to every position including
an image span; a partly-failed erase left some scenes changed; an
unreadable sigma disabled sharpening while still reporting it active.
A malformed reference slot claimed nothing was marked.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The panel replaced the whole log with the fetch error — destroying the
last lines the server sent at the exact moment they matter, and what
Copy would hand over. It now holds them and backfills across a restart
from ComfyUI's own log file.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Native crashes print a Python traceback instead of nothing, and
kill -USR1 dumps every thread's stack from a hung process without
attaching anything or restarting it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
MiniMax H3 encodes with Qwen3-VL; the stand-in path fetched a 12B Gemma
tokenizer regardless. Local HF cache is now tried before any download,
and a download announces itself — it blocks the run with no timeout.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Comfy materialises the full delta before discovering it will not
reshape: 51 blocks x 96768x2688 on a curve-form H3 checkpoint, ~25 GB
churned in bf16, for adapters that could never have applied.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
FUNPACK_PINNED_MEMORY=off|<GB>|<pct>%, applied at import — after the
budget is computed, before any model is staged. Only ever lowers it.
Unevictable pinned pages wedge a host rather than killing the process.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Freeze before and after the requirements install and print the diff.
A silent bump of numpy or transformers under ComfyUI's compiled
extensions segfaults hours later with nothing pointing at the update.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Page-locked pages cannot be swapped or reclaimed, so a budget near
total RAM is what wedges a Linux host rather than killing the process.
Warn when ComfyUI's budget exceeds 80% of RAM.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
RSS, host RAM, swap, VRAM and what ComfyUI is holding (with how much
sits on the CPU), before and after the decode — the call that evicts
model weights from VRAM into host RAM.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A decode is one uninterruptible call with no progress of its own, so a
slow decode looked identical to a hung one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A token pointing along the negative is left with rounding error after
projection; restoring its norm amplified that noise ~1000x into the
conditioning. Cap the gain and refuse a non-finite result.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
On H3 it is a Qwen3-VL-32B forward whose output CFG 1.0 never reads.
It still has to happen, but it must be visible when accounting for
what a run loaded and when.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Log latent shape, estimated CPU pixel size, free VRAM/RAM, and empty
torch's cache before decoding — a decode that kills the process leaves
no traceback. Sharpness proxy releases its retained prediction.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Render geometry, sampler + schedule, second pass when running, Studio
and Chain Sampler overrides. Only the selected algorithm's block is
printed, and pass 2 only when it is actually on.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Erase the negative's pooled direction from the positive conditioning
(projection by default, never the vision span). Plus an option to let a
wired positive CONDITIONING own the prompt while CLIP keeps its jobs.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Pass 2 can run a different algorithm. Unwired it reuses pass 1; in the
Editor it is behind Second pass > Own sampler, and Studio mirrors the
high pass into that output when off, so existing projects are unchanged.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Wire a socket to "Reference image 1" and re-order marks in the bin
without opening a node page. Per-kind numbering; an unmarked slot is
left unconnected silently, never auto-wired to a substitute.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A bypassed slot claims no input, is not a second source, and needs no
pass-through for a consumer that is fed elsewhere or does not require
one. Long errors wrap, scroll and copy instead of running off-canvas.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Loaders, LoRAs, typed-in custom-node values and host torch/CUDA/GPU.
Wired inputs are marked, not printed. Themed, watermarked, and carries
the report as JSON in a tEXt chunk.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Lift the unsharp out of the sampler loop through a denoiser proxy, the
same way ALG does. Window is a fraction of the schedule; audio excluded.
Off by default; an unwrappable sampler keeps sampling and says so.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Clone from a git URL, ff-only pull, delete. node_dir() is the whole safety
story for the delete: one path segment, resolved, must be a direct child of
custom_nodes, never FunPack itself.
Pulled code but never dependencies, so a release adding one broke the
import silently. pip runs against sys.executable, opt-in from the route,
off the event loop; failures name the manual command.
The panel is the Editor's Properties column: collapsed there it stayed
display:none, and its folds are hidden in Simple mode, so the button opened
nothing. ComfyUI-GGUF refusing an arch now falls back to dequantizing.
A .gguf renamed .safetensors reached the safetensors parser (UTF-8 decode
error); content now decides. ComfyUI-GGUF's loader uses relative imports, so
its directory is mounted as a package instead of loaded file-by-file.
Curve-form H3 checkpoints read adaLN from a compact basis, so a full-width
turbo LoRA matches all 51 adaLN keys and merges into none. Named explicitly
instead of 51 generic warnings under a clean-looking status line.
Core lists no .gguf, so the files were invisible. Backend is whichever is
installed: ComfyUI-GGUF keeps weights quantized, the gguf package
dequantizes at load. Family detection reads GGUF containers too.
Both are structurally inert on H3; context windowing also mis-reported the
cause as an old ComfyUI. alg_anchor deliberately kept: a continuation scene
carries real latent frames for it to act on.
H3 keys keyframe pins by frame index and takes only first/last, so memory
frame 0 evicted the anchor and the rest were refused. Forced off with a
reason; hidden in Engine Settings. Tooltips now state the LoRA requirement.
The prompt is anchor + transitions + scene texts, so clearing only the
bodies left it non-empty. Postfix and variables are separate settings and
are untouched.
The bar never recorded which template was applied, so there was nothing to
rename, delete or leave; deletePromptTemplate had no caller at all. Clearing
is now explicit because applyGlobalPromptQuiet refuses empty text.
Built-in loaders show an icon and a plain name instead of their class;
third-party nodes keep theirs. The add-node panel reuses the node page's
multi-destination editor, gate included.
