You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Outputting images is useless without being able to generate valid images. The standard way to Python to generate images appears to be the Pillow library. The Pillow library cannot easily be added to eval-py as it depends on native code. I experimented with trying to have Claude generate image generating code (without third-party libraries) and was largely unsuccessful, either the outputted image was not valid or it did not match the prompt.
This can be picked back up when the python-pdk supports adding native libraries and/or AI models are better at generating image files from scratch.
No description provided.
The text was updated successfully, but these errors were encountered: