MiniMax H3 MXFP8
MXFP8 quantization of the MiniMax H3 fl2va (first-last-frame to video+audio) diffusion checkpoint.
better audio at 5 (talking heads, less motions) & 8-steps (shift 8)
MiniMax H3 FP16attn
Balanced quality and performance
fp16attn 8-steps | shift: 8.0 Your browser does not support the video tag.
MiniMax H3 FP8 (E4M3FN) Quantized
Based on testing (this model and pruned_fp8), the minimal steps to get results similar to Lightx2v using 4-step LoRAs is 5step & 5steps. The sweet spot for quality and speed is 8step and 8steps. This quantization requires a minimum of 8step.
- Sampler: dpmpp_2m
- Scheduler: sgm_uniform
| 5-Step pruned fp8 | 5-Step mxfp8 | 8-Step fp8 |
|---|---|---|
| 8-steps fp16attn |
|---|
FP8 quantization of the MiniMax H3 fl2va (first-last-frame to video+audio)
diffusion checkpoint.
Original model: MiniMaxAI/MiniMax-H3
File sizes
| Variant | Download |
|---|---|
| minimax_h3_fl2va_mxfp8.safetensors | Download 47.6GB |
| minimax_h3_fl2va_fp16attn_fp8.safetensors | Download 28.7GB |
| minimax_h3_fl2va_fp8_e4m3fn.safetensors | Download 47GB |
| minimax_h3_fl2va_pruned_fp8_scaled.safetensors | Download 21GB |
Simple Upscalers
Old but gold upscale models. Lower image resolution for faster inference and more steps.
| Variant | Download |
|---|---|
| 2xNomosUni_esrgan_multijpg.pth | Download |
| RealESRGAN_x2plus.pth | Download |
736x416 | ~56 sec | 1472x832 60fps Your browser does not support the video tag.
Left at native precision (quality-critical)
| Tensors | Dtype | Note |
|---|---|---|
adaln_proj.linear.* (100 tensors) | BF16 | Per-block modulation, largest single weight |
token_refiner.* (17 tensors) | BF16 | Text embedding refinement, 2 layers only |
condition_proj.* (2 tensors) | BF16 | Cross-modal conditioning bridge |
All norm weights (norm1, norm2, q_norm, k_norm, final_norm) | BF16 | 104 tensors, tiny and precision-sensitive |
| All biases | BF16 | 56 tensors |
time_embedder.* (4 tensors) | FP32 | Timestep signal, original precision preserved |
video_patch_proj.*, audio_patch_proj.* (4 tensors) | FP32 | Input/output projections |
final_layer.*_out.* (4 tensors) | FP32 | Decoder output heads |
rope.inv_freq (1 tensor) | FP32 | Position encoding |
ComfyUI-MiniMaxH3-Text-Enhancer node
What this is: the result of looking into why male anatomy comes out misshaped in text-to-video scenes
or when a nude male appears in the I2V scene.
Along the way we learned some things worth knowing about how this model really works, read README.md
and notes.md in case of you are interested.
Q: Is it a text-encoder problem? Would an abliterated Qwen3-VL solve it?
- A: No. The stock encoder doesn't refuse NSFW prompts, the tokenizer and embedding table handle them normally. Its only strong divergence from stock is on harmful-instruction content (~11% embedding shift); sexual content shifts only ~1–3%. Abliteration targets the wrong thing here.
Q: Is the base model censored? Would a domain-specific LoRA be ineffective?
- A: No, the base is uncensored. The misshape comes from weak text conditioning: only ~3–5% of the embedding actually carries the prompt, so the DiT's generic body prior wins over the anatomy request. Modifying internal values (condition_proj ×1.2 + 12,658 FFN columns ×1.2) improves the result slightly and a properly trained LoRA would refine anatomy further on top of it.
Experimental: 1.2 strength in both nodes improves NSFW rhythmic animations in some scenarios