Skip to content
Other

Universal Conditioner adapters

adapters for VRAM savings in ComfyUI

1 min read

This set of ComfyUI-UniversalConditioner adapters lets you use a single language model on a separate server as a text and visual encoder (data encoding module) for various image and video generation models. ComfyUI itself does not load the heavy text encoder into VRAM, which saves resources. The adapters convert the hidden states of the base language model to fit the target model space. The project is tailored for ComfyUI users looking to optimize their workflow with generative models.

Key features

  • Support for the Qwen-Image 2.1 target model
  • Operation via an external server running the Qwen3.8-27B language model
  • VRAM savings by avoiding encoder loading in ComfyUI
  • Experimental status and open adapter files

What you can do

  • Set up ComfyUI-UniversalConditioner with Qwen-Image 2.1
  • Connect an external server with the Qwen3.8-27B model for text and image processing
  • Reduce VRAM load during image and video generation

Example

Install the ComfyUI-UniversalConditioner node, place the adapter file in the models/universal_conditioner/ folder, and configure the address of the server handling hidden state requests.

Read moreDetails
Target model: Qwen-Image 2.1
Base model for the adapter: Qwen3.8-27B (neroued/Qwen3.8-27B-nvfp4-NInfer recommended)
License: Qwen Research License (non-commercial use only)
Status: experimental
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.3 GB · ready files0.8 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 GB
  • 4-bitweights 0.1 GB0.6 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

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.

    # igorls/universal-conditioner
    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 igorls/universal-conditioner --local-dir ~/models/universal-conditioner
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to universal-conditioner
  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.

  • 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 (2)
    • 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.
  • 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.

Model family

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

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

About the model

igorls/universal-conditioner
Author
igorls
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

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

Comments