Tristan Rice

Seems like a reasonable feature to me. Is there also interest in global queue/multiregion support? That would be interesting to have as well though likely much more work to implement...

@starpit can share a bit more about why you want to override the `app_id`? We haven't supported setting `app_id` on the scheduler via the config options and always take it...

We don't document this but the output from `torchx run` will return only the job ID via stdout. All non-structured data is outputted to stderr. We use it for some...

@KexinFeng we've completely revamped the install process so it should be a lot simpler and not require building torch from source. Does the new install process work better for you?

closing since not up to date any more

I believe this is no longer relevant

outdated, already added

FYI for anyone following this: I'm currently implementing this in pytorch and following this RFC and the comments as much as possible. Once I get a good working POC I'll...

@Huxwell the issue is you're loading a quantized model via `quantize=True` and then mixing quantized and non-quantized layers. You need to load the unquantized model, modify/train/finetune and then apply the...

You can work around this by avoiding adding any functional methods to the class ie. avoid `self.foo = F.gelu`