How to use GemmaAPITextEncode node in LTX 2.3 Workflow
Hello RuneXX,
Thank you for sharing the detailed LTX 2.3 workflows - it is well structured and works reliably.
I have been using your LTX-2.3_-_I2V_T2V_Basic_Custom_Audio.json workflow, and it runs perfectly on my setup. However, I am exploring ways to reduce memory usage.
Specifically, I am interested in integrating the GemmaAPITextEncode node into your workflow. Could you please guide me on how to properly configure it.
Will check how the GemmaAPITextEncode is in Comfy, see if it can be easily added ;-)
Tested it, works fine. +
I'll add and upload to the workflow you wnated, but if you want to try yourself its basically just bypass all the clip and prompt nodes that are in the workflow (select nodes and ctrl+b to bypass)
Then get the API key you need https://console.ltx.video/
And instead connect like this_
Or even simpler, just prompt directly in the GemmaAPI node itself:
Hello RuneXX, thanks for the quick reply, I have configuered the node as you suggested, but the checkpoint is not being recognized in GemmaAPITextEncode node, I have put ltx-2.3_text_projection_bf16.safetensors in ComfyUI\models\text_encoders folder.
At the bottom of that node, i think it looks in the checkpoint folder if i remember right.
So copy 2.3_text_projection_bf16.safetensors over to the checkpoint folder also, so that this gemma node can see it
EDIT:
Checked, yes it looks for ltx-2.3_text_projection_bf16.safetensors in the checkpoint folder.
So add the file there as well, and enter your api key. And all should work ;-)
At the bottom of that node, i think it looks in the checkpoint folder if i remember right.
So copy 2.3_text_projection_bf16.safetensors over to the checkpoint folder also, so that this gemma node can see itEDIT:
Checked, yes it looks for ltx-2.3_text_projection_bf16.safetensors in the checkpoint folder.
So add the file there as well, and enter your api key. And all should work ;-)
Works perfectly, thanks brother