Elinx
Elinx
These two methods don't use `genericScheduler`'s member variable, how about removing the receiver like this: ```diff diff --git a/pkg/scheduler/core/generic_scheduler.go b/pkg/scheduler/core/generic_scheduler.go index 453df094..ec424489 100644 --- a/pkg/scheduler/core/generic_scheduler.go +++ b/pkg/scheduler/core/generic_scheduler.go @@ -145,7 +145,7...
> > These two methods don't use genericScheduler's member variable, how about removing the receiver like this: > > I think `selectCluster` is part of `genericScheduler`. Please confirm if [#2247...
> I get it. You just need `selectClusters`, the single method, not the `genericScheduler`. right? yes
Meet the same problem here `Unable to connect to the server: net/http: request canceled (Client.Timeout exceeded while awaiting headers)`, it happens because I use go1.19.2 at the first run and...
There are several ways you could try to make it work depends on you situation: 1. set the lp job to be preemptable by adding annotation: ` volcano.sh/preemptable: "true"` 2....
#2225 Did this solve your problem?
> We have some thoughts on network topology aware scheduling as well, `BUT` limited in the same IDC. > > ## User Story > Saying we have a 10,000 GPU...
> The tokenizers in transoformers do not support this function automatically when calling decode function > > The [standard ](https://github.com/huggingface/transformers/blob/d628664688b05cabdd69f4e7e295bc4aee0a8d31/src/transformers/generation/streamers.py#L111) way of going about this is holding tokens in cache...
> this picture is for training. in inference, MTP draft worker takes same input and previous hidden state, then outputs draft token Thanks for the reply, that makes sense. So...
> > > this picture is for training. in inference, MTP draft worker takes same input and previous hidden state, then outputs draft token > > > > > >...