ComfyUI Node
BSAI Jev Structured Decision
A ComfyUI node in BSAI/Qwen Image 2.1 with 17 inputs and 3 outputs.
BSAI Jev Structured Decision
- jev_schema
- result_json
- confidence_json
- text
◄context►
◄batch_modefalse►
◄llm_model_name<No GGUF model found>►
◄mmproj_name►
◄chat_handlerauto (auto-detect Qwen3.8/Gemma4/GLM4.6 and other supported models from the model name)►
◄n_ctx8192►
◄n_gpu_layers-1►
◄load_mtpfalse►
◄n_cmoe0►
◄kv_cache_quantauto (f16, no quant)►
◄use_mmaptrue►
◄use_mlockfalse►
◄flash_attnauto►
◄threads0►
◄system_prompt►
◄force_offloadfalse►
CategoryBSAI/Qwen Image 2.1
Inputs (17)
| Name | Type | Default | Description |
|---|---|---|---|
| jev_schema | JEVSCHEMA | — | |
| context | STRING | — | |
| batch_mode | BOOLEAN | false | True: split context by line into multiple records; one forward pass decides all fields per record (records processed in order; fields are decided in parallel within a single forward pass) |
| llm_model_name | COMBO | <No GGUF model found> | 1 options: <No GGUF model found> |
| mmproj_name | COMBO | 1 options: | |
| chat_handler | COMBO | auto (auto-detect Qwen3.8/Gemma4/GLM4.6 and other supported models from the model name) | 15 options: auto (auto-detect Qwen3.8/Gemma4/GLM4.6 and other supported models from the model name), MTMD generic multimodal (recommended: Qwen3.5+/Gemma4/GLM with mtmd-mmproj), Qwen 3.8 / 3.6 / 3.5 vision (requires updated llama-cpp-python), Qwen 3 / 3-VL vision (requires updated llama-cpp-python), Qwen 2.5-VL / 2-VL vision, Gemma 4 vision, +9 |
| n_ctx | INT | 8192512–65536 | — |
| n_gpu_layers | INT | -1-1–200 | — |
| load_mtp | BOOLEAN | false | — |
| n_cmoe | INT | 00–256 | Split MoE experts across CPU/RAM layers (-ncmoe). Only effective for MoE models, not dense.\nQwen-35B-A3B is most stable around 24; 0 = off. Current 0.3.36 does not support it yet; it applies automatically after upgrading. |
| kv_cache_quant | COMBO | auto (f16, no quant) | KV cache quantization (-ctk/-ctv). Jev reads logits and is quality-sensitive, so it defaults to no quantization; pick q8_0/q4_0 only if you need the speedup. |
| use_mmap | BOOLEAN | true | use_mmap; can disable when RAM is tight / swapping (--no-mmap). |
| use_mlock | BOOLEAN | false | use_mlock (--mlock) locks the model in RAM to avoid page-out. |
| flash_attn | COMBO | auto | -fa. auto = enable when the model supports it. |
| threads | INT | 00–256 | -t thread count (video example: 14); 0 = auto. |
| system_prompt | STRING | — | |
| force_offload | BOOLEAN | false | Unload all cached local LLMs after inference. |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| result_json | STRING | — |
| confidence_json | STRING | — |
| text | STRING | — |