WeVisOne-VL

WeVisOne-VL

Packs Tencent's WeVisOne-VL model files for ComfyUI, in 4B and 2B sizes with BF16 and int8 versions. For ComfyUI users who need the files in their text encoders folder.

Made by Comfy-Org on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: WeVisOne-VL here is a repackaged set of model files that Comfy-Org made for ComfyUI. The files come from Tencent's WeVisOne-VL, and the card points to the original model page for the model itself. It lists a 4B and a 2B size, each in BF16 and in an int8 build.

Who it suits: It suits ComfyUI users who need the WeVisOne-VL files in a form ComfyUI can read.

Industry | Who uses it | How they use it | Example
Image workflows | ComfyUI users | Put the files where ComfyUI looks for text encoders | A workflow that reads one of the four files
Research | Model testers | Compare the 2B and 4B files | A side-by-side of the BF16 and int8 builds
Tooling | Workflow builders | Prepare a ComfyUI install with the right files | A setup script that places each file

People who do not use ComfyUI can skip it, since these files are packaged for that tool.

Where to use it:

• ComfyUI, in its models folder for text encoders
• Windows, macOS or Linux, where you download the files with Python

Use case | Who | What they get back
Install the files | ComfyUI users | Four safetensors files placed in the right folder
Pick a size | Workflow builders | A choice between the 4B and 2B files
Pick a precision | Testers | A choice between the BF16 and int8 files

How to set it up: You need these first:

• Python and the huggingface_hub package
• ComfyUI installed on your computer

1. Install the Hugging Face hub package.
2. Download the model files to your computer.
3. Copy the files you want into the text encoders folder inside ComfyUI's models folder.

pip install -U huggingface_hub
hf download Comfy-Org/WeVisOne-VL

How to use it: After the files are in place, start ComfyUI and choose one of them where a workflow asks for a text encoder file. The card gives no workflow of its own, so read the original model page to see how WeVisOne-VL is meant to be used.

Example: You copy the 4B BF16 file into the text encoders folder and restart ComfyUI. ComfyUI can then find that file in the folder. What the model does in your workflow depends on the original WeVisOne-VL, which this card does not describe.

For newcomers and experts: Repackaged means the same model, saved in a layout that a tool can load directly. A first small thing to try is downloading one file and putting it in the folder, then restarting ComfyUI.

For experts, the file names show a BF16 build and an int8 build named convrot for each size. The card does not say how the int8 files were made. Hugging Face lists WeVisOne-VL-2B as the base, while the card links the 4B original, so check which one you need.

Download the repository files
Pick a size and a precision
Copy the file into ComfyUI's text encoders folder
Restart ComfyUI
Use the file in a workflow

Good to know:

• Files: four safetensors files, a BF16 and an int8 one for each of the 4B and 2B sizes
• Base: Hugging Face lists WeVisOne-VL-2B, and the card links the 4B original
• License: apache-2.0
• Cost to run: none named, no API key or account
• Data it reaches: it downloads model files from Hugging Face and the files sit in your ComfyUI folder
• Not for: use outside ComfyUI, since the card says these files are repackaged for it

On Hugging Face, 13 November 2026

• Downloads in the last 30 days: 0
• Likes: 21
• License: apache-2.0
• Based on: tencent/WeVisOne-VL-2B
• Tags: comfyui

Links

• Model: https://huggingface.co/Comfy-Org/WeVisOne-VL
• Files: https://huggingface.co/Comfy-Org/WeVisOne-VL/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download Comfy-Org/WeVisOne-VL

Then follow the usage in its model card. Nothing in this kit runs by itself.