Skip to content
Images

chroma-H3 LoRAs в формате ComfyUI

готовые файлы в bf16

1 min read

Это ComfyUI-конвертации LoRA из репозитория lodestones/chroma-h3. Автор crombobularBoop переложил ключи и структуру файлов, а обучение выполнил lodestone. Конвертации доступны в bf16 для text-to-image на базе MiniMax-H3.

Ключевые особенности

  • Два файла: основной LoRA и вариант vspace DMD8.
  • chroma-h3_lora_comfy.safetensors — применять со strength 1.
  • chroma-h3_lora_vspace_dmd8_comfy.safetensors — для stock MiniMax-H3, Euler, CFG 1, 4–8 шагов.
  • Формат ComfyUI, точность bf16.
  • В комплекте есть скрипт конвертации convert_chroma_h3_lora.py.

Пример

В ComfyUI загрузите chroma-h3_lora_vspace_dmd8_comfy.safetensors и примените к stock MiniMax-H3: sampler Euler, CFG 1, 4–8 шагов. Для обычного LoRA используйте strength 1.

ПодробнееДоступны два файла: основной LoRA и vspace DMD8 в формате ComfyUI.
Для vspace DMD8 указаны stock MiniMax-H3, Euler, CFG 1 и 4–8 шагов.
Все файлы в bf16; конвертация выполнена скриптом convert_chroma_h3_lora.py.
Это конвертация ключей и структуры, а не новое обучение LoRA.
Discuss Telegram 0 270

Will it run?

Working out the memory needs

This is a LoRA: it only runs on top of the base model lodestones/chroma-h3. The base model's size is still being collected; the estimate will appear within a day.

Massed ComputeOpen Massed Compute

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # crombobularBoop/chroma-H3-comfyui
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download lodestones/chroma-h3 --local-dir ~/models/chroma-h3
    hf download crombobularBoop/chroma-H3-comfyui --local-dir ~/models/chroma-H3-comfyui
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to chroma-h3
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • SageAttention 3v2.2.0from October 28, 2025

    even faster on RTX 50 / Blackwell

    FP4 attention for Blackwell. Faster than SageAttention 2, but the authors suggest v2 when accuracy matters.

    How to enable: Build from the sageattention3_blackwell folder of the repo; in ComfyUI via KJNodes.

    Earlier generations (2)
    • SageAttention 2++ — Quantised (INT8/FP8) attention with no visible quality loss. The main speed-up for image and video generation.
    • SageAttention 1 — First version: INT8 for Q·K. Still works, but 2++ is faster and more accurate.
  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (3)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
    • xFormers — The older way to speed up attention, works almost everywhere. Sage or Flash are faster where available.
  • cache-dit (DBCache)v1.5.6from November 20, 2026new release

    1.5–3× at nearly the same quality

    Reuses block outputs between diffusion steps. Supports most DiT models.

    How to enable: pip install cache-dit; one line cache_dit.enable_cache(pipe) for diffusers.

    Earlier generations (2)
    • MagCache — Skips steps by magnitude of change. More accurate than TeaCache, no calibration.
    • TeaCache — The first popular diffusion step cache. Works, newer methods are more accurate.
  • Nunchaku (SVDQuant 4-bit)v1.2.1from January 25, 2026

    ≈3× faster and 3.5× less memory

    4-bit versions of popular image models with almost no quality loss (FP4 on RTX 50).

    How to enable: Ready weights for supported models; in ComfyUI use ComfyUI-nunchaku.

  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

MiniMax H3 prompts

All prompts

Ready prompts with the resulting video: copy and try.

Model family

Requires the base model

lodestones/chroma-h3

This model is a LoRA / adapter. Download the base model first, then load this one on top.

New posts about versions of chroma-h3 (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

crombobularBoop/chroma-H3-comfyui
Author
crombobularBoop
Category
Images
Released
Posted

On NeyroLearn

Hugging Face

2700

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments