<h3 id="comfyui-workflow">ComfyUI Workflow</h3><ul><li>required Custom Nodes:<ul><li><a target="_blank" rel="ugc" href="https://github.com/r-vage/ComfyUI_Eclipse">ComfyUI_Eclipse: Version 2.4.23, </a><a target="_blank" rel="ugc" href="https://registry.comfy.org/publishers/rvage/nodes/ComfyUI_Eclipse">ComfyReg (Nodes 2.0 compat, V3 API)</a></li><li><a target="_blank" rel="ugc" href="https://github.com/r-vage/ComfyUI_SmartLML">ComfyUI_SmartLML: Version 2.5.0</a>, <a target="_blank" rel="ugc" href="https://registry.comfy.org/publishers/rvage/nodes/ComfyUI_SmartLML">ComfyReg (Nodes 2.0 compat, V3 API)</a></li></ul></li><li>if you need help check the readme files in the repo <a target="_blank" rel="ugc" href="https://github.com/r-vage/ComfyUI_SmartLML#">readme</a><a target="_blank" rel="ugc" href="https://github.com/r-vage/ComfyUI_RvTools/tree/main/Readme"> </a>folder at github.</li><li>llama-cpp-python is needed for the loading_method: GGUF (llama-cpp-python) in the Smart LML v2 Node<ul><li><a target="_blank" rel="ugc" href="https://github.com/r-vage/ComfyUI_Eclipse/releases/download/wheels/llama_cpp_python-0.3.16-cp312-cp312-win_amd64.whl">https://github.com/r-vage/ComfyUI_Eclipse/releases/download/wheels/llama_cpp_python-0.3.16-cp312-cp312-win_amd64.whl</a></li></ul></li><li>switch the template in the Smart Language Model Loader v2 to download the model. the node detects if the model already exists, it is also not downloading florence2 anymore when models\florence2 already exists, except for models that are not found in this folder.</li></ul><ul><li>load images from folder</li><li>Smart LML v2 loads the selected model for the image description</li><li>ReplaceString v3 removes unwanted parts</li><li>save prompt (.txt, .csv, .json)</li><li>if you create one file e.g. random_prompts you can use this file in the wildcard processor. for that:<ul><li>copy the file to the wildcards folder e.g. comfyui\models\wildcards. reload the page or restart comfyui if that didnt work and load the file from the list.</li></ul></li></ul><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/9f00e16d-f987-4f16-a49f-e2f418a88bd6/original=true/9f00e16d-f987-4f16-a49f-e2f418a88bd6.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/e215d37d-7e26-4dbc-8563-b18cbac1922d/original=true/e215d37d-7e26-4dbc-8563-b18cbac1922d.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1f310384-fa3e-45f5-b79a-59504b4ed7a3/original=true/1f310384-fa3e-45f5-b79a-59504b4ed7a3.jpeg" />

Batch Captioning Workflow

<ul><li>Needs <a target="_blank" rel="ugc" href="https://github.com/r-vage/ComfyUI_Eclipse">Eclipse v1.4.1</a></li><li>the reupload contains the read prompt files node that allows to load all prompts from a file and use them either with a model to refine the prompts / remove unwanted parts (kind of works with text only models, can also be used with task chaining) or use replace string v3 which works much better now. connect the output to text input of Smart LM v2 or Replace String</li></ul>

Files with SHA256: 5fab550a0899d5933c13f6e8f9a716a13b09e17efba322512f2d5767a85b92d5