Reference only

mat1 and mat2 shapes cannot be multiplied: inspect the model component chain

Matrix dimensions help locate a failure during sampling, encoding or adapter use, but one numeric pair does not identify a unique missing model.

Symptoms and scope

Matrix dimensions help locate a failure during sampling, encoding or adapter use, but one numeric pair does not identify a unique missing model.

Error fragments for search; IDs, values and filenames may vary:

mat1 and mat2 shapes cannot be multiplied
size mismatch

Source-supported context

Official troubleshooting connects tensor-shape failures with incompatible model combinations, and issue #12076 supplies a concrete matrix-error report. Neither justifies reshaping tensors arbitrarily to hide an architecture mismatch. Source 1 Source 2

Cases to distinguish

1. The base model, encoder, ControlNet, IPAdapter or LoRA comes from an incompatible family or variant.

2. A node update changed its input structure while the workflow retained an older combination.

3. An extension constructs an unexpected tensor; a matched native baseline is needed to separate this from resource mismatch.

Diagnostic sequence

The sequence below is an editorial procedure based on the cited context, not a diagnosis already confirmed for your environment.

Step 1. Preserve complete dimensions, node, traceback and recent changes instead of reporting only size mismatch.

Step 2. List diffusion, text encoder, VAE and every adapter, recording the author's supported family for each.

Step 3. Remove optional adapters, test the author's original component combination, then restore one extension at a time.

Step 4. If the original combination fails too, compare core/node versions and template revision and submit a minimal reproduction. Do not blindly reshape, truncate embeddings or replace layers.

Completion check

The components have a documented compatibility basis and the minimal case passes. A repeatable failure after reintroducing one component narrows the problematic boundary.

Limitations and cautions

Shape edits can make execution proceed while changing model meaning. Values such as 768 or 2048 are not a timeless lookup table for all errors.

Original sources

Localization prepared: 2026-09-22. The source-check dates above were inherited from the existing article; this translation does not claim they were all rechecked today. No GPU run, minimum-memory measurement or runtime guarantee is supplied.

Didn't solve it?

Check the next possible cause

The same symptom can come from a different cause. Work through these related entries in order.

  1. LoRA has no effect or reports size mismatch: check the base model and both output brancheslora key not loadedA selectable LoRA can still target an incompatible model, or the workflow can bypass its modified outputs. Little visible change is an observation, not proof of a damaged file.
  2. ControlNet fails or has no visible effect: check the base, control image and apply pathControlNetDiagnose ControlNet compatibility, control-image preparation and conditioning separately; node-link type errors need their own check.
  3. CLIP input is invalid: None — trace where the text encoder should come fromCLIP input is invalid: NoneA wire reaching CLIPTextEncode does not prove that a valid encoder was loaded. Check upstream loading and whether the model uses a complete checkpoint or separate components.
Paste your full log into the error finder

Was this page helpful?

Anonymous. Only a yes/no counter is stored; no account, IP address or device details.

Sources & references

Localized from the existing Chinese research draft. Original source-check dates are retained; this translation is not new technical verification or a runtime test.

01ComfyUI model troubleshootingSource checked: 2026-09-2102Matrix-shape error reported in ComfyUI DesktopSource checked: 2026-09-21

Report an issue · bbdc75ee-d018-51e3-a404-443282e6b9de