Skip to content
WorkflowListedExternal
Open source

Gemma 4 ComfyUI Text Encoder Heretic

Repackaged models for usage as text encoders in ComfyUI

Sourcehuggingface.co/pekkAi/gemma-4-comfyui-text-encoder-heretic/tree/62cd1639c7a904ed6845a6d8a2140e228eb800c4
VersionRelease 62cd1639c7a904ed6845a6d8a2140e228eb800c4 — Revision 62cd163Estimated VRAMNot yet characterizedCgrade

Runtime profile

Access
External workflow
Status
Listed
Version
Revision 62cd163
Source
Hugging Face
Updated
Oct 18, 2026

Source description

Gemma 4 - E2B/E4B - heretic - text encoders

Repackaged models for usage as text encoders in ComfyUI

Original model repository:

Place the files in the following folders:

📂 ComfyUI/
├── 📂 models/
│   ├── 📂 text_encoders/
│   │   ├── gemma4_e2b_heretic_bf16.safetensors
│   │   ├── gemma4_e2b_heretic_fp8_scaled.safetensors
│   │   ├── gemma4_e4b_heretic_bf16.safetensors
│   │   └── gemma4_e4b_heretic_fp8_scaled.safetensors

What you need

Hardware and download-size estimates are not available yet.

Requirements

Verified requirements are unavailable for this workflow variant. Nothing is marked as found without file-level evidence.

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

Purpose

  • Gemma 4 ComfyUI Text Encoder Heretic is a ComfyUI workflow that uses a CLIP input and text prompt to generate text from an image, video-derived images, or audio, then sends the result to preview.

Flow

  • LoadImage and LoadAudio feed TextGenerate directly; LoadVideo passes through GetVideoComponents, which sends images to TextGenerate.

  • CLIPLoader supplies TextGenerate’s CLIP input, and PreviewAny receives its generated text.

Inputs and outputs

  • TextGenerate takes a CLIP encoder, an image, video-derived images, and audio; its output is generated text (STRING).

Setup

  • For the packaged text encoder files, place these in ComfyUI/models/text_encoders/: gemma4_e2b_heretic_bf16.safetensors, gemma4_e2b_heretic_fp8_scaled.safetensors, gemma4_e4b_heretic_bf16.safetensors, and gemma4_e4b_heretic_fp8_scaled.safetensors.

    Suggestion · not verified

Dependencies

  • The workflow’s model link names gemma4_e4b_it_fp8_scaled.safetensors in Comfy-Org’s text_encoders folder.

Configuration

  • The saved TextGenerate prompt is “Describe the image, and transcribe the audio”.

  • CLIPLoader’s saved options are stable_diffusion and default.

Cautions

  • Before loading a model, compare these E4B FP8 names: the packaged file is gemma4_e4b_heretic_fp8_scaled.safetensors, the saved CLIPLoader filename is gemma4_e4b_ultra-uncensored-heretic_fp8_scaled.safetensors, and the linked Comfy-Org file is gemma4_e4b_it_fp8_scaled.safetensors.

    Suggestion · not verified

  • If troubleshooting, update ComfyUI first; Desktop and Cloud updates follow stable releases, so models available only in nightly builds may not yet be available there.

    Suggestion · not verified

Does this need to be edited?

Sign in to send an edit request.

Sources

1 source
Hugging FacePrimary
huggingface.co/pekkAi/gemma-4-comfyui-text-encoder-heretic

Verified Oct 18

Model artifacts

4 artifacts

gemma4_e2b_heretic_bf16.safetensors

safetensors · 9.57 GB · SHA-256 d0f2d06764c0…3c87

gemma4_e2b_heretic_fp8_scaled.safetensors

safetensors · 5.47 GB · SHA-256 76f1192100cf…5fd8

gemma4_e4b_heretic_bf16.safetensors

safetensors · 14.9 GB · SHA-256 6ec11a9cbe02…e24c

gemma4_e4b_heretic_fp8_scaled.safetensors

safetensors · 8.44 GB · SHA-256 df93f93f8181…3490

Source excerpts

3 excerpts
ReadmeJul 14huggingface.co/pekkAi/gemma-4-comfyui-text-encoder-heretic

--- license: apache-2.0 tags: - comfyui - text-encoder - fp8 base_model: - llmfan46/gemma-4-E4B-it-ultra-uncensored-heretic - llmfan46/gemma-4-E2B-it-ultra-uncensored-heretic --- # Gemma 4 - E2B/E4B - heretic - text encoders Repackaged models for usage as text encoders in ComfyUI Original model repository: - [llmfan46/gemma-4-E4B-it-ultra-uncensored-heretic](https://huggingface.co/llmfan46/gemma-4-E4B-it-ultra-uncensored-heretic) - [llmfan46/gemma-4-E2B-it-ultra-uncensored-heretic](https://huggingface.co/llmfan46/gemma-4-E2B-it-ultra-uncensored-heretic) Place the files in the following folders: ``` 📂 ComfyUI/ ├── 📂 models/ │ ├── 📂 text_encoders/ │ │ ├── gemma4_e2b_heretic_bf16.safetensors │ │ ├── gemma4_e2b_heretic_fp8_scaled.safetensors │ │ ├── gemma4_e4b_heretic_bf16.safetensors │ │ └── gemma4_e4b_heretic_fp8_scaled.safetensors ```

NoteAug 30huggingface.co/pekkAi/gemma-4-comfyui-text-encoder-heretic

Source context: 0 downloads · 2 likes · Repo pekkAi/gemma-4-comfyui-text-encoder-heretic

NoteJul 14huggingface.co/pekkAi/gemma-4-comfyui-text-encoder-heretic

Source context: 0 downloads · 1 likes · Repo pekkAi/gemma-4-comfyui-text-encoder-heretic