Ziyang

According to Swoole's guide book: When your server is listening to both TCP and UDP protocols, onPacket only receives UDP packets and onReceive ony receives TCP data. But if onPacket...

Is `$hander` actually the typo of `$handler`?

any chance we'd have this feature soon?

I just pushed a new commit to avoid calling S3 api when accessing thumbs. `Attributes()` method for the filesystem: [filesystem.go#LL128-L162](https://github.com/metricsubs/pocketbase/blob/7edad2828b56bf221bcd7d21fd8c6b6564c33c79/tools/filesystem/filesystem.go#LL128-L162). Basically, this is like what you suggested, which is sending...

Or a optional URL rewrite rule config, e.g.: `https://mycdn.com/{BUCKET}/{KEY}`

According to my napkin math, even the smallest model with 7B parameters will probably take close to 30GB of space. 8GB is unlikely to suffice. But I have no access...

> Could be possible with https://github.com/FMInference/FlexGen This project looks amazing 🤩. However, in its example, it seems like a 6.7B OPT model would still need at least 15GB of GPU...

To add to this, on macOS when double clicking the `data-tauri-drag-region` area to resize the window, the Webview will also not change its size during transition.

> I think this issue should be closed: https://twitter.com/ylecun/status/1629214720405143552 My guess is that the model architecture code is open sourced under GPL3, but the pretrained weights ain't. Technically Yann isn't...

> From another thread by @dizys I hereby officially transfer the copyright ownership of this logo artwork (originally [here](https://github.com/valkey-io/valkey-website/issues/7#issuecomment-2033284794)) to the Valkey project.