Brian Lambert

Any plans to support multi-task tuning?

1

Some examples of using a single base model for multiple tasks: - https://github.com/NVIDIA/NeMo/blob/main/tutorials/nlp/Multitask_Prompt_and_PTuning.ipynb - https://openreview.net/forum?id=Nk2pDtuhTq

MIssing External Data File

1

Hey there! I just stumbled across your repo as I was debugging my own LoRA onnxruntime code - it looks like we independently took a really similar approach (identify reused...

Support vector search

2

**Describe the bug** Hey there! I noticed that you just bumped libcblite to 3.2.0 - do you have any plans to support the [vector search feature](https://docs.couchbase.com/couchbase-lite/current/cbl-whatsnew.html?_gl=1*j2zp7r*_gcl_aw*R0NMLjE3MjgzMTYzMDQuQ2owS0NRandqWTY0QmhDYUFSSXNBSWZjN1laU29nY0VqcFozNDRaOHQxeHJEbnZwVE52MW1ac1N6VW5pZEJ6T19qN28tYi1PWDZnaGJXOGFBcHFURUFMd193Y0I.*_gcl_au*MjA5NjExNjE1NS4xNzI4MDQ3NzMz) that was released in...

Request: Estimated RAM Requirement

1

It would be helpful if the `foundry model info ` command included a rough estimate of the RAM required to run each model so that we could gate some models...

Vector Search Library Not Loading Periodically

2

**Describe the bug** Sorry for another bug here.... Ocassionally on Windows 10, MacOS 15.2, and MacOS 15.3 I'm getting a crash when calling `createIndex` and it appears that it's failing...