mfatih7
mfatih7
depth_clean = f['depth'].value gives errors and can be corrected like **depth_clean = f['depth'][:]** ______________________________________________________________ Then f2 = plt.imshow(data['depth'].astype(np.float)) can be corrected like **# f2 = plt.imshow(data['depth'])** ______________________________________________________________ changing plt.imshow(rgb.transpose((1, 2,...
Hello I use Spyder IDE 5.3.2. Sometimes I cannot move the edge bars of panes. Even if they move the movement is very restricted. I also observe that close all...
Hello I have models trained using multi-core TPU option. I have saved their checkpoints using ``` import torch_xla.core.xla_model as xm xm.save(checkpoint, path_checkpoints_file, master_only=True ) ``` Now, on new models to...
Hello For a model that can be sharded in model parallelization in TPUv4 (4x32) device, I am getting the error below at the beginning of the training on TPUv3 (8x16)...
Hello I am using CheckpointManager to save and load checkpoints for models with SPMD. In my training setup, in the first epoch, the model is trained with lossA only. After...
Hello While implementing checkpoint operations for SPMD by examining the [documentation](https://github.com/pytorch/xla/blob/master/docs/spmd.md) I get a couple of warnings. ``` /home/THE_USER/env3_8/lib/python3.8/site-packages/torch/distributed/checkpoint/state_dict_saver.py:29: UserWarning: 'save_state_dict' is deprecated and will be removed in future versions.Please...
Hello Without SPMD, we can train 8 duplicates of a model on 8 TPU cores on the same v3 device and optimize model weights concurrently. With SPMD, is it possible...
Hello I am getting errors since HardSwish has not been implemented