rockerBOO/ltx-2.5-nvfp4-convrot

LTX-2.5 — NVFP4 + ConvRot INT8 quant for ComfyUI

Quantized version of Lightricks' LTX-2.5 (22B audio-video diffusion transformer) for ComfyUI, using NVFP4 for most weights and INT8+ConvRot for attention — smaller and faster than the bf16 original with minimal quality loss. Unlike Lightricks' own official quants, blocks 0, 1, 46 and 47 are kept full bf16 precision here.

dev variant is planned but not started yet.

Files

File	Size	Status
ltx-2.5-22b-distilled-transformer_nvfp4_convrot_int8.safetensors	22 GB	Uploaded, tested in ComfyUI
ltx-2.5-22b-dev-transformer_nvfp4_convrot_int8.safetensors	~22 GB (est.)	Planned

Only the transformer is requantized here. Grab the VAEs and text encoders straight from Lightricks/LTX-2.5 and Comfy-Org/gemma-4 — those are used as-published, not touched by us.

Requirements

Blackwell GPU (SM >= 10.0/12.0) required for inference — RTX 50-series, B100/B200. Same as our other NVFP4 releases.

License

LTX-2 Community License Agreement (included in this repo), inherited from Lightricks/LTX-2.5.

---

Quantization method

No quantization was run ourselves for this file — it's spliced together from tensors in Lightricks' own two official pre-quantized releases, since both already quantize the same tensor set we wanted (verified against the real safetensors headers, not assumed):

• Base: ltx-2.5-22b-distilled-transformer-comfy-int8-convrot.safetensors — keeps its attention tensors (INT8+ConvRot) and everything Lightricks itself never quantizes (norms, gate-logits, adaln, patchify/proj_out, scale-shift tables).
• Feed-forward layers (ff.net.0.proj/ff.net.2, audio_ff.net.0.proj/audio_ff.net.2, incl. embeddings_connector) swapped in from ltx-2.5-22b-distilled-transformer-nvfp4.safetensors.
• Blocks 0, 1, 46, 47 overridden back to full bf16 (attention, feed-forward, everything) from the bf16 source, superseding both official files for just those 4 blocks — Lightricks quantizes all 48 blocks uniformly; we chose not to for the first/last two.

Two things surfaced during verification that are worth recording:

1. Lightricks' own NVFP4 file isn't uniform either. 56 of its 224 feed-forward tensors are plain bf16, not NVFP4: blocks 42-47 (both video and audio FF) and all of embeddings_connector (video+audio, 8 sub-blocks each). This wasn't documented anywhere we could find — only visible by reading the file's own tensor headers. Our output inherits this (blocks 42-45 end up FF-bf16/attention-INT8; blocks 46-47 are already fully bf16 from our own override anyway).
2. A missing .comfy_quant marker silently breaks NVFP4 loading in a mixed-format file. Lightricks' standalone NVFP4 file has no per-tensor .comfy_quant blob at all — ComfyUI evidently auto-detects a uniformly-NVFP4 checkpoint some other way. But comfy/ops.py's generic per-layer loader (if layer_conf is None: load as plain weight, no error) is what actually runs once a checkpoint contains any .comfy_quant-bearing tensors (i.e. as soon as INT8 layers are mixed in) — so NVFP4 tensors without their own marker silently loaded as raw packed weights, producing a mat1 and mat2 shapes cannot be multiplied error deep in the first sampler node, not a load-time failure. Fixed by adding a synthetic {"format": "nvfp4"} .comfy_quant tensor to all 160 NVFP4-tagged feed-forward tensors. Confirmed working via a full ComfyUI API workflow run (image-to-video + audio, video_ltx2_5_i2v_api.json) after the fix — first attempt failed with exactly this error, second attempt (post-fix) completed and produced valid output video/audio files.

Final composition (verified against the assembled file's own tensor headers):

Blocks	Attention	Feed-forward
0, 1, 46, 47	BF16 (our override)	BF16 (our override)
2-41	INT8 + ConvRot	NVFP4
42-45 + embeddings_connector	INT8 + ConvRot	BF16 (inherited from Lightricks' own NVFP4 file)

Quality (beyond functional load/generate correctness) has not been separately assessed.