how ever it takes some while to judge 2000 midi files
Nilabha Bhattacharya
@nilabha666_twitter
does anyone have the code to split mid files ? I have generated files which are slightly greater than the 3610 seconds and would want to trim the last few sections.
Nilabha Bhattacharya
@nilabha666_twitter
I have managed to manually trim the last portions of the track
Nilabha Bhattacharya
@nilabha666_twitter
but that is more hackish since I manually decide how many items from the last to trim
Nilabha Bhattacharya
@nilabha666_twitter
is there anyway to trim based on times?
Aozhi
@Aozhi
Hello
Does anybody know why the challenge is extended?
What is the exact date of the end?
@spMohanty
SP Mohanty
@spMohanty
@Aozhi: The challenge has been extended by 7 days. Exact date and time should be reflected on the challenge page. (PS: Sorry for the brevity. Travelling with sporadic internet connection)
Aozhi
@Aozhi
@spMohanty Thank you so much Monhanty!
Abdul Khader Jilani
@y2ee201
hi guys, anyone using techniques other than RNN based sequence prediction?
Aozhi
@Aozhi
@spMohanty Hello Monhanty, may I know that why the music can still be evaluated when the challenge is completed? Is there any record of the winner at the moment of Jan 7th UTC14:00?
SP Mohanty
@spMohanty
@Aozhi: Thanks for pointing out. We will be closing the evaluation interface soon; but only the evaluations submitted until Jan 7th, 14:00 will be used for chosing the winner. We will be doing a manual verification of all the submitted evaluations, to ensure that there was no participants used any unfair approaches, etc.
Aozhi
@Aozhi
@spMohanty Thank you so much Mohanty. It is a well organized challenge!
Qin Yongliang
@ctmakro
@nilabha666_twitter if you somehow generated sequences longer than 3600 seconds, that might be because some part in your generator includes a rounding operation.
since in Midi, the intervals are integer multiple of some small value.
if the rounding is round-to-zero, then accumulated error will result in generated sequence longer than expected.
or shorter.
my solution is to use dither, that converts for example 3.1 into 3 for 90% of the time and 4 for 10% of the time.
this way the errors never build up to more than 1 second/3600 seconds.
Qin Yongliang
@ctmakro
i don't know but somehow i became like the top3
Aozhi
@Aozhi
@spMohanty Hello Monhanty, it seems that the competition has restarted. Is there any updated prize? The prize is still an invitation to meeting at 2018.
SP Mohanty
@spMohanty
@Aozhi : we will be posting details soon. But the prize will most likely be an invitation to Applied ML days, EPFL
_
Aakash Bansal
@aakashbansal
Hi! I wish to know if this competition is still open in 2018. The crowdAI is saying 64 days left but not not accepting any submissions. Any ideas ?
Aakash Bansal
@aakashbansal
Oh Sorry! My bad. Just realised the submission is made through the python script in repo itself. :p
SP Mohanty
@spMohanty
@aakashbansal : there hasn't been much activity on this challenge
So we will have to consider to close this challenge up this year:
We will post updates about it soon.
Aakash Bansal
@aakashbansal
Oh. Too bad! I have been planning to participate in this challenge after some time. Anyways thanks for the update @spMohanty . :)
Gasinse
@Gasinse
I'm delighted I was able to discover a great piano without too many issues and problems, because you can search here and also look at this website to get good pianos for a reasonable price https://steinway.co.uk/pianos/special-collection/ . I've already purchased one and am really pleased with the build quality and sound. This is quite unusual, and I'm delighted I came across it.