Sevag H

@diamantisk I found the following in the documentation that might be relevant: >Note that if you do not care about learning about updated devices, you might call this function only...

It took me a total of 2 or 3 days to integrate Stripe subscriptions into my website for delivering premium content, in a flow that "just works" using simple components...

I think I'm running into a similar issue on Thanos 0.14.0 The promql is `sum(irate(my_app_counter[5m]))`, where `my_app_counter` is a counter. Thanos with deduplication unchecked:  Thanos with deduplication checked shows...

Thanks - looks like I need to upgrade to 0.15.0+ to be able to use this new flag? What's a good value to pick with a scrape interval of 60s?...

Looks like upgrading to 0.16.0-rc0 and modifying `--query.lookback-delta` isn't fixing the above situation.

Actually, this is an error on my part. It's related to a new librosa version. I'm patching it as we speak (commit to follow).

Forgot to follow up - I will try to replicate later today.

The regression test uses the 7s MUSDB18 excerpt (which may not even be in the wav format). I wonder if that's related.

Can the pretrained model download be a single zip containing the 4 training logs, separator.json, and the 4 pth files? The working UMX setup. Or is it split up in...

I just saw the other suggestion to do inference in 30s-sized chunks so I'll do it that way.