Hi all, version 0.0.75 is live! This release includes:

Web & Desktop:

• Fix a bug where a local character creation timestamp would get overwritten when you edited the character

Desktop:

• Fix a bug where the model wouldn't be downloaded correctly
• Add a button which will let you download a model immediately upon downloading the app
• Update how we package the llama.cpp binaries - if you see something break because of this, please let me know!
• Fix a bug in our token counting logic - this may help avoid model incoherence during longer conversations
• Improve error handling during model loading - we now will display better error messages and avoid infinite loading when things go wrong (in some cases)

Thank you!