Finding info about the loaded model at runtime #1672
-
I'm looking to get sd-webui-negpip back up and running, as it's an extension which I love dearly. I need to replace the original tokenize_line calls, which I'm thinking I can do with a ClassicTextProcessingEngine, but I need to get the tokenizer and text encoder values for the loaded model. Presently the old backend/model handling logic for doing this is commented out, and the new backend seems to be trying to import a module which isn't actually packaged with Forge, as far as VSCode can tell (huggingface_guess). So I'm at a loss of how to actually achieve this. I see |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
I found my way around with |
Beta Was this translation helpful? Give feedback.
I found my way around with
pprint(dir(<object>))
, in lieu of any actual documentation.