World Needs AI
English
English日本語Españolעברית

10Eros-Max-h3-turbo-hybrid-beta3-w4a8

Source
Hugging Face
Type
Text to video
Asset type
Quantized weights
LoRA category
General
Released
Base model
TenStrip/10Eros-Max
Creator
berryber09
Downloads
342
Likes
3
minimax-h3comfyuiw4a8quantizedtext-to-videoturbo

Capabilities

text-to-video · W4A8 quantized weights · Roughly 68.8% smaller than bf16 source (per card)

Compatibility

ComfyUI v0.31.0 or later (per card, or provided patch) · comfy-kitchen AsymW4A8Int8Layout · CUDA SM 8.0 or higher · UNETLoader (convrot) loading

Models

  1. 10Eros-Max-h3-int8-convrot

    Hugging FaceQuantized weightsText to videoGeneralBase model: TenStrip/10Eros-MaxCreator: vdaular
    comfyuiint8convrotdiffusion-model
    int8 tensorwise quantizationConvRot groupsize 256skip-edges variant keeps blocks 0, 1, 48, 49 in BF16
    0Downloads

  2. 10Eros-Max-h3-turbo-hybrid-beta4-w4a8

    Hugging FaceQuantized weightsText to videoGeneralBase model: TenStrip/10Eros-MaxCreator: berryber09
    minimax-h3comfyuiw4a8quantizedtext-to-videoturbo
    text-to-video generationquantized inferenceturbo mode
    2,531Downloads

  3. 10Eros_Max_h3_TURBO-hybrid_beta3-GGUF

    Hugging FaceQuantized weightsText to videoGeneralCreator: electroglyph
    gguf
    text-to-video generationgguf format inference
    888Downloads

  4. 10Eros-Max-h3-int8-convrot

    Hugging FaceQuantized weightsText to videoGeneralBase model: TenStrip/10Eros-MaxCreator: chfm
    comfyuiint8convrotdiffusion-model
    image-to-video generation
    0Downloads

  5. 10eros_Max_Ref2va

    Hugging FaceCheckpoint / mergeText to videoGeneralBase model: MiniMaxAI/MiniMax-H3Creator: Stuubs
    text-to-video synthesisreference-to-video-audio conditioningadult content generation
    0Downloads

← All models