Fix/tensorboard incompatible with numpy 1.24

README.md

@@ -355,6 +355,9 @@ If you need more visualizations, use `add_scalar('tag', data)`, `add_image('tag'
 
 **Note**: You don't have to specify current steps, since `WriterTensorboard` class defined at `logger/visualization.py` will track current steps.
 
+**known issue**: when using numpy>=1.24 and pytorch <=1.13 may encontered Type error: No loop matching the specified signature and casting was found for ufunc greater which will fixed by following update of pytorch.
+
+one fix is to install the pkg from source, since it has been fixed in source just havent released yet. another if following this hot fix that modify the torch/utils/tensorboard/summary.py from  cum_counts = np.cumsum(np.greater(counts, 0, dtype=np.int32)) to cum_counts = np.cumsum(np.greater(counts, 0)) one can find the pkg source by clicking the import line in vscode
 ## Contribution
 Feel free to contribute any kind of function or enhancement, here the coding style follows PEP8
 

parse_config.py

@@ -111,6 +111,10 @@ def __getitem__(self, name):
         """Access items like ordinary dict."""
         return self.config[name]
 
+    def __contains__(self, name):
+        """Use in operation like ordinary dict."""
+        return name in self.config
+
     def get_logger(self, name, verbosity=2):
         msg_verbosity = 'verbosity option {} is invalid. Valid options are {}.'.format(verbosity, self.log_levels.keys())
         assert verbosity in self.log_levels, msg_verbosity