Hi,
1. can you please explain how the evaluation data is given?
is it comming in chunks of 360 samples or like the csv given for training, in which i need to crop segments of 360 samples
2. in metadata file (sample submission) there is python model.py input output
the output shall be file? number? what is the expected ouput.
Thanks in advance
Posted by: avig @ April 18, 2022, 6:15 p.m.Hi
1. The test data format consists of chunks with 360 samples each, the predict function receives one chunk at a time and return one prediction value for each chunk.
2. The code execution is creating the output file automatically, you don’t need to change the metadata or create the output file.
Thank you and good luck,
MAFAT challenge team.
can we assume the chunks will come in the right order (time wise), or they can be shuffled?
Posted by: lionb @ May 2, 2022, 7:45 a.m.Hi,
Within a single chunk the samples are ordered according to time (“the right time”), but the order of the chunks is shuffled.
MAFAT Challenge Team
Hi,
Related questions in regards to the test format:
In the participate section it mentions that `Every window contains RSSI values from one device only (one device-router channel), and the participants are required to perform a prediction with it`
Do we get multiple devices to aggregate the data from each time window or do we need to estimate from a single device reading how many people are in the room?
Thanks, Yair
Posted by: yairb @ May 15, 2022, 9:20 a.m.Hi,
Each window's prediction should be based on a single device reading.
MAFAT Challenge Team