Open yuming-l2 opened 2 years ago
Is it intended to be 240 rather than 280? Seems 240 would make more sense for the dataset and in real life.
https://github.com/noise-lab/ml-systems/blob/7fb6a90cb6629b657b62fbfaf7c02df727183a07/docs/assignments/1-Video-Quality-Inference.ipynb?short_path=41ca219#L140-L148
Yes, definitely, we will fix this!
Is it intended to be 240 rather than 280? Seems 240 would make more sense for the dataset and in real life.
https://github.com/noise-lab/ml-systems/blob/7fb6a90cb6629b657b62fbfaf7c02df727183a07/docs/assignments/1-Video-Quality-Inference.ipynb?short_path=41ca219#L140-L148