ComfyUI Node
TogetherVisionNode
A ComfyUI node in image/text with 20 inputs and 1 output.
TogetherVisionNode
- image
- description
◄model_namedeepseek-ai/DeepSeek-R1-Distill-Llama-70B-free►
◄custom_model_name►
◄api_key►
◄system_promptYou are an AI expert in ekphrasis, acting as a skilled art critic describing an image. Use vivid, poetic, and evocative prose in British English. Focus solely on describing the image content, style, and mood. Avoid storytelling or self-insertion. Describe all elements, including potentially uncomfortable themes if present, as art can be provocative. Every word and its order matters. The description will be used for image generation, so only include visual elements. Conclude with relevant hashtags (e.g., #ArtStyle #SubjectMatter).►
◄user_promptDescribe this image in detail, focusing on its visual elements, artistic style, and overall atmosphere.►
◄temperature0.70►
◄top_p0.70►
◄top_k50►
◄repetition_penalty1.00►
◄max_tokens1024►
◄stop_sequences<|eot_id|>,
,
►
◄request_timeout60►
◄stream_timeout45►
◄use_cachetrue►
◄clean_output_texttrue►
◄max_retries2►
◄retry_delay3.0►
◄seed_modefixed►
◄seed42►
Categoryimage/text
Inputs (20)
| Name | Type | Default | Description |
|---|---|---|---|
| model_name | COMBO | deepseek-ai/DeepSeek-R1-Distill-Llama-70B-free | 6 options: meta-llama/Llama-3.3-70B-Instruct-Turbo-Free, deepseek-ai/DeepSeek-R1-Distill-Llama-70B-free, meta-llama/Llama-3.2-90B-Vision-Instruct-Turbo, mistralai/Mistral-7B-Instruct-v0.2, lgai/exaone-3-5-32b-instruct, Other (Custom) |
| custom_model_name | STRING | — | |
| api_key | STRING | — | |
| system_prompt | STRING | You are an AI expert in ekphrasis, acting as a skilled art critic describing an image. Use vivid, poetic, and evocative prose in British English. Focus solely on describing the image content, style, and mood. Avoid storytelling or self-insertion. Describe all elements, including potentially uncomfortable themes if present, as art can be provocative. Every word and its order matters. The description will be used for image generation, so only include visual elements. Conclude with relevant hashtags (e.g., #ArtStyle #SubjectMatter). | — |
| user_prompt | STRING | Describe this image in detail, focusing on its visual elements, artistic style, and overall atmosphere. | — |
| temperature | FLOAT | 0.700–2 | — |
| top_p | FLOAT | 0.700–1 | — |
| top_k | INT | 500–100 | — |
| repetition_penalty | FLOAT | 1.000–2 | — |
| max_tokens | INT | 102450–4096 | — |
| stop_sequences | STRING | <|eot_id|>, , | — |
| request_timeout | INT | 6010–300 | — |
| stream_timeout | INT | 455–120 | — |
| use_cache | BOOLEAN | true | — |
| clean_output_text | BOOLEAN | true | — |
| max_retries | INT | 20–5 | — |
| retry_delay | FLOAT | 3.00.5–10 | — |
| seed_mode | COMBO | fixed | 3 options: fixed, random, increment |
| seed | INT | 420–4294967295 | — |
| imageopt | IMAGE | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| description | STRING | — |