Evidence sheets for a MiniMax H3 packing question: when a keyframe guide and a Ref2VA reference are used on the same conditioning, where does the guide actually land? Two sheets, twenty cells at one seed: four ComfyUI builds by five reference configurations, sixteen of which render and four of which raise. Kept here rather than on the front page because it is a note for one upstream discussion, not part of the Motion Lab.
Guide rows are positioned from the text span, while reference blocks are packed afterwards and advance the cursor the target video sits on. So every reference slides the guide earlier relative to the clip it is meant to anchor. The size of the slide is the reference's own cursor advance:
| reference | guide lands at | visible? |
|---|---|---|
| none | +0.0 units | on target frame 0 |
| one image | −1.0 unit (−0.6 frames) | sub-frame, no |
| 107-frame video plate | −150 units (−90 frames) | outside a 73-frame clip |
Row 1 is the input for each column. Below it, generated frame 0
measured against the guide image (mean absolute error, 0–255). The image
reference columns hold the guide everywhere, because the slide is under one
frame. The video plate of the same scene is where it breaks: 33.91, and the
beaker in the guide is simply gone. Stock master has no row because it raises
RuntimeError: shape mismatch on all four reference configurations.
Frame 72 of 73. Swapping an image reference for a restyled version of the same subject moves the output by about 12/255; swapping the video plate moves it by about 51/255, and what transfers is framing, scale and staging rather than identity. A single seed and a single subject, so read these as a direction, not a law.
Mean absolute error on 0–255 RGB, full frame. About 1.7 is seed-level noise between two runs of the same configuration, about 5 is a cleanly anchored frame, and a guide the model ignores lands near 70. The fix that produces the clean rows is one small branch: derive the guide origin from the same reference advance the target uses. It renders pixel-identical whether applied to master or on top of the upstream PR.
back to the Motion Lab · repo · MIT