feat(gallery): add minimax-h3-ref2va-q4 - the Ref2VA partition, verified coherent - #11439
Merged
Conversation
…ied coherent MiniMax-H3 ships two independently-served DiT partitions and each refuses the other's tasks. The gallery had only the FL2VA one (t2va + fl2va); this adds the Ref2VA checkpoint, which is what serves REFERENCE conditioning - a reference image, a reference clip, or reference audio prepended as their own blocks. It is a genuinely separate 19.9 GB download, not a mode flag, which is why it needs its own entry. Everything except the DiT is shared with the FL2VA entry. Q4_K_M specifically, and the description says why: the NVFP4 Ref2VA weights render a multicolour patch grid. Three upstream investigations attributed that to the ref2va reference-row path before the checkpoint was varied; on Q4_K_M the identical code renders coherently (period-16 seam 1.13 against 2.28 broken, VAE-input latent adjacent-cell cosine 0.8526 against 0.06 white). Pointing users at NVFP4 here would hand them that dead end. known_input_modalities carries image, video AND audio, unlike the FL2VA entry: the ABI's ref2va slots are ref_image (PPM), ref_video (a directory of frame_%06d.ppm) and ref_audio (16-bit PCM WAV), and the backend's checkPartitionConditioning already routes all three. Checksum verified two ways: sha256 computed over the downloaded 19,864,208,064 byte file, and HuggingFace's x-linked-etag for the same URI - both 17925612821ea3037ffaf5f7f9789f5460e87025385bd45e9ec6c7d536684d56. Not verified here: an end-to-end render through LocalAI itself. The render was proven with vllm.cpp's own driver on this exact checkpoint; the backend path is the same library seam but has not been exercised with these weights. Assisted-by: Claude Code:claude-opus-5 [ClaudeCode]
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
MiniMax-H3 ships two independently-served DiT partitions and each refuses the other's tasks. The gallery had only the FL2VA one (t2va + fl2va). This adds the Ref2VA checkpoint — the one that serves reference conditioning.
It's a genuinely separate 19.9 GB download rather than a mode flag, which is why it needs its own entry. Everything except the DiT is shared with the FL2VA entry.
Why Q4_K_M, explicitly
The NVFP4 Ref2VA weights render a multicolour patch grid. Three upstream investigations attributed that to the ref2va reference-row path before anyone varied the checkpoint; on Q4_K_M the identical code renders coherently:
The description warns users off NVFP4 rather than leaving that dead end for them to rediscover.
Modalities
known_input_modalitiescarries image, video and audio (the FL2VA entry has only text+image). The ABI's ref2va slots areref_image(PPM),ref_video(a directory offrame_%06d.ppm) andref_audio(16-bit PCM WAV), andcheckPartitionConditioningalready routes all three.Checksum verified two ways
x-linked-etagfor the same URIBoth
17925612821ea3037ffaf5f7f9789f5460e87025385bd45e9ec6c7d536684d56.Not verified
An end-to-end render through LocalAI itself. The render was proven with vllm.cpp's own driver on this exact checkpoint (512x512/124f, coherent office scene, no grid). The backend drives the same library seam, but has not been exercised with these weights.