ComfyUI Node
Local AI Model (Advanced)
A ComfyUI node in WepeNerd/Local AI/Advanced with 20 inputs and 1 output.
Local AI Model (Advanced)
- config
◄model▾►
◄llama_serverauto►
◄context_size8192►
◄gpu_layers-1►
◄target_free_vram_mb24576►
◄aggressive_vram_handofffalse►
◄release_after_generatetrue►
◄mmproj▾►
◄startup_timeout_s300►
◄request_timeout_s600►
◄extra_server_args►
◄keep_alive_seconds0►
◄flash_attn▾►
◄cache_type_k▾►
◄cache_type_v▾►
◄image_min_tokens0►
◄image_max_tokens0►
◄cuda_visible_devices►
◄comfy_vram_handoff▾►
◄native_video_max_mb96►
CategoryWepeNerd/Local AI/Advanced
Inputs (20)
| Name | Type | Default | Description |
|---|---|---|---|
| model | COMBO | 1 options: <put .gguf models in ComfyUI/models/LLM> | |
| llama_server | STRING | auto | — |
| context_size | INT | 8192256–262144 | — |
| gpu_layers | INT | -1-1–999 | — |
| target_free_vram_mb | INT | 245760–262144 | — |
| aggressive_vram_handoff | BOOLEAN | false | — |
| release_after_generate | BOOLEAN | true | — |
| mmprojopt | COMBO | 1 options: (none) | |
| startup_timeout_sopt | FLOAT | 3005–1800 | — |
| request_timeout_sopt | FLOAT | 6005–7200 | — |
| extra_server_argsopt | STRING | — | |
| keep_alive_secondsopt | INT | 00–86400 | — |
| flash_attnopt | COMBO | 3 options: auto, on, off | |
| cache_type_kopt | COMBO | 2 options: f16, q8_0 | |
| cache_type_vopt | COMBO | 2 options: f16, q8_0 | |
| image_min_tokensopt | INT | 00–65536 | — |
| image_max_tokensopt | INT | 00–65536 | — |
| cuda_visible_devicesopt | STRING | — | |
| comfy_vram_handoffopt | COMBO | 3 options: auto, always, never | |
| native_video_max_mbopt | INT | 961–1024 | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| config | GGUF_LLM_CONFIG | — |