+1
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544779150
+1
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544754307
@mbarrett97 I see your point. If the problem is narrowed down to "skip some
tasks in a model when resuming the tuning that was accidently interrupted",
then your proposal is a lightweight working solution. Maybe we can file another
RFC focusing on a more general history reuse support.
Then talk
@comaniac I think I understand where our different approaches are coming from.
I was proposing that only the optimal configurations be permanently saved to
the config library (like with TopHub) and a temporary log file of tuning
configs would be maintained only during a tuning job. Storing all o
+1
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544614691
Thanks for the helpful discussion. Some of the common themes that I see
- Need for more meta data to inform the tuner, if possible
- The key question is whether meta data is mandatory or serve as an auxiliary
component.
- e.g. we may not want the general features to must depend on the
meta-d
Thanks for the reponses and I think they are valuable. I embedded my opinions
with yours and leave the dispatch context for @kevinthesun.
Also cc @tqchen and @icemelon9 for their inputs.
> > If we design this resume logic in a general way, we can also extend it to
> > tophub.
>
> Does it make
+1
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544583390
+1
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544539880
+1
> On Oct 21, 2019, at 12:07 AM, Tianqi Chen wrote:
>
> It has been great to see community growth over the past few years. One thing
> that was bought up by several community members recently was should we show
> organization logos that use and contribute to TVM.
>
> This is an RFC thread t
+1
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544497287
+1
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544489114
Thanks @kevinthesun and @comaniac for the responses!
> I'd prefer to keep the current AutoTVM dispatch context
I'm not intending to replace the existing dispatch context, only provide some
syntactic sugar. We could just override the `__enter__` method of ConfigLibrary
to do `apply_history_best`
+1
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544379584
+1
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/dmlc/tvm/issues/4162#issuecomment-544377190
15 matches
Mail list logo
