ComfyUI Node
🖼️ V2 Local Vision Analysis
A ComfyUI node in 🤖 GGUF-VLM/v2/Local with 10 inputs and 1 output.
🖼️ V2 Local Vision Analysis
- model
- image
- video
- context
◄promptDescribe this image in detail.►
◄max_tokens512►
◄temperature0.7►
◄top_p0.90►
◄top_k40►
◄seed0►
◄system_promptYou are a helpful assistant that describes images and videos accurately and in detail.►
Category🤖 GGUF-VLM/v2/Local
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| model | VISION_MODEL | 视觉语言模型配置 | |
| prompt | STRING | Describe this image in detail. | 用户提示词 |
| max_tokens | INT | 5121–4096 | 最大生成 token 数 |
| temperature | FLOAT | 0.70–2 | 温度参数 |
| top_p | FLOAT | 0.900–1 | Top-p 采样 |
| top_k | INT | 400–100 | Top-k 采样 |
| seed | INT | 00–18446744073709550000 | 随机种子 |
| imageopt | IMAGE | 输入图像（与视频二选一） | |
| videoopt | IMAGE | 输入视频帧序列（与图像二选一） | |
| system_promptopt | STRING | You are a helpful assistant that describes images and videos accurately and in detail. | 系统提示词（可自定义模型行为） |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| context | STRING | — |