Sergei Miliaev
Sergei Miliaev
@ptrblck , I observed similar issues for training embeddings on classification task with a large number of classes. I tried both optimizations O1 and O2 and was able to solve...
@zlenyk , as I see in the current [documentation](https://nvidia.github.io/apex/amp.html#module-apex.amp) and [code](https://github.com/NVIDIA/apex/blob/master/apex/amp/frontend.py#L195), **max_loss_scale** should be a valid option.
@brimoor, are there any updates on an image fetching API implementation? It would greatly improve the usability of FiftyOne for various dataset formats.