Raffi Khatchadourian
Raffi Khatchadourian
I see the description of the `files` option as follows: Name | Type | Description -- | -- | -- files | String | Newline-delimited globs of paths to assets...
The warning for chktex is missing the warning number, e.g., "Warning 8," which is useful for suppressing warnings. I think that a fix may go in this code: https://github.com/dense-analysis/ale/blob/f0887d3e6178482255f11aa378124aef3699245f/ale_linters/tex/chktex.vim#L29-L47
All tensors initially have the dimensions of the MNIST dataset regardless if they actually have those dimension.
The initial tensor shapes should be unknown. Currently, they are set to the MNIST dataset shape. This makes the current tests pass.
Per #267, we track initial tensor shapes from literal expressions. However, we also need to track them from non-literal expressions, such as function arguments and return values. This is important...
Since `expectedNumberOfTensorParameters` must be equal to `expectedTensorParameterValueNumbers.length`, we can remove the `expectedNumberOfTensorParameters` parameter and just use `expectedTensorParameterValueNumbers`. https://github.com/wala/ML/blob/53d2f0a82446e4913682249af7de1b270addb32a/com.ibm.wala.cast.python.ml.test/source/com/ibm/wala/cast/python/ml/test/TestTensorflow2Model.java#L3496-L3500
Per #331, we want to refactor the print statements to instead use logging. But, some print statements are OK, like the ones used in the CLI. Once #331 is fixed,...
There are many places in the code that use `System.out.println()` or `System.err.println()` that should instead use the logging we've already setup. On the other hand, there is a CLI where...
Switch the ML code analysis to use JEP if if not done so already.
I wonder if there's a bug in the PYTHONPATH code that is necessitating this change ... _Originally posted by @khatchad in https://github.com/wala/ML/pull/300#discussion_r2320058929_