Karen Hambardzumyan

We also noticed that. Sigmoid disappeared after we switched from using Dense layers (with sigmoid activation) to SharedWeight layers. Now I'm trying to repeat training process. We hope can get...

Pull request is merged. Not closing the issue until we compare new scores. @MartinXPN please rerun the instruction steps from README.md

Sorry for late response, What Keras version do you use?

I suppose having a separate portable client (running without root access) should be even easier to accomplish, and (arguably) more valuable.

@laiqinghan Please find the instructions at #28

@shailja11 Are you sure you're getting the same "EADDRINUSE :::3000" error? If yes, try "lsof -i :3000" again and make sure there's no process using port 3000 anymore.

`aimrocks` implements `BackupEngine`, however, the feature is not tested well. Consider you have `db` instance of `aimrocks.DB`. To backup: ``` backup = rocksdb.BackupEngine("test.db/backups") backup.create_backup(db, flush_before_backup=True) ``` To restore: ``` backup...

Previously the right way of closing RocksDB was just deleting the instance (in C++). If you `del rocks_db` it will close it. But we plan to fix this in the...

I tried to go deeper and recover the errors causing exit. I ended up here: https://github.com/pytorch/pytorch/blob/db8abde9b6c4735d18d4681a1f70a55ff0b09f5b/torch/multiprocessing/spawn.py#L72-L76 Got an error: ``` Traceback (most recent call last): File "/home/USER/miniconda3/envs/MY_METASEQ_ENV/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 69, in...

> The rank 1, 2, 3 was exit before the loop of train_step. I print the every detailed log and find that the iter() inside more_itertools.peekable() kill all the non-master...