Jie Fu

support upload/download model to/from hubble

1

Since we move to `jcloud` deployment, it's necessary to support uploading/downloading PCA/PQ model to/from `Hubble`. Thus, we need to implement these APIs: self._projector_codec.upload(artifact='...') self._projector_codec.download(artifact='...') The `artifact` is determined by users...

acquire train data from lmdb

3

Sometimes we need to train the PCA model when we already created an indexer. (for example, there is a memory issue after we have indexed thousands or even millions of...

In this pr, we added a splitter and a merger based on `filesplit` package to support `backup_to_remote` and `restore_from_remote`.

demo code on huggingface is not runnable

1

I got error when running the sample code on hf: https://huggingface.co/docs/diffusers/api/pipelines/if#diffusers.IFImg2ImgSuperResolutionPipeline the error message: I already logging in to huggingface. It seems that there is a bug when mapping the...

Dimension mismatch when using Coca for VQA task

5

I use `generate` endpoint to do VQA task in Coca model, but got this error: It seems that this issue will not happen in `beam_search` mode but appear in `top_k`...

investigate how to pass `past_key_values` to generate()

1

We already implemented passing `past_key_values` to `step_generate()` to enable cache in LLM. `generate()` also support this but should have a different method to implement this since `generate()` will truncate the...