Truncated response when generating

General
1

Hey there! I’ve noticed a few instances of AI Decompilation being truncated early. Is this known?

Example:

image
image1217×1868 249 KB

File hash: 4308ed406a8de64eb3ca4e6accc27d10794e960a4631590cab569cf734f3d3e9

2

We currently have a maximum input and output length in the early experimental models but we’ll be upgrading the context length during the next release (hopefully in December). This should also include much better performance and Windows support :muscle:

3

Glad to hear this! Thanks!