Viewing profile — scv119

The tweet is referring to a paper that fine tunes Chinese dataset on english base model. I'm not surprised with LoRA's poor result in this setup.

I believe you can slightly change the flash attention kernel to implement the same kernel of this page attention, since both of them work on the key/value cache at block level.

Pretty cool stuff and the results are amazing. Hoping we will see virtual memory get standardized in pytorch or cuda.

We wanted to ensure everyone on the team has explored the full design space, and everyone agreed that the chosen solution is the best we can think of, given the current information…

Co-creator of Delos here, Let me give a concrete example for [18]. We have a "StorageEngine" interface that wraps around RocksDB, i.e. "RocksDbStorageEngine" to provide key-value i…

I guess that a hard question for user facing API: you need to support both the old and the new features during the migration. In our case, this design principle mostly applied to o…

The first engineer on the Delos team here. This design decision gave us an edge: we delivered the first production cluster in 8 months by layering over existing systems ; and we sm…

Agreed, technology-wise FDB and TIDB look almost the same.

Sorry, there is no DECENT internet connection in China