ComfyUI Node
CLIP Text Encode (with Offload)
CLIPTextEncode with optional VRAM and RAM offload. Releases the CLIP model from VRAM and/or RAM after encoding to save memory. ComfyUI's output cache ensures the node does not re-execute on repeated generations with the same prompt, keeping performance impact minimal.
CLIP Text Encode (with Offload)
- clip
- CONDITIONING
◄text—►
◄offload_from_vramfalse►
◄offload_from_ramfalse►
Categoryconditioning
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| text | STRING | The text prompt to encode. | |
| clip | CLIP | The CLIP model used for encoding the text. | |
| offload_from_vram | BOOLEAN | false | When True, unloads the CLIP model from VRAM to CPU after encoding. Repeated generations with the same prompt use the cache, so no re-encoding occurs. |
| offload_from_ram | BOOLEAN | false | When True, also releases pinned (page-locked) memory after encoding. offload_from_vram is implied. Reduces RAM usage further but slightly slows the next load. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| CONDITIONING | CONDITIONING | A conditioning containing the embedded text used to guide the diffusion model. |