Tom
Tom
I'm comparing performance of the built-in Dense layer with an equivalent Linear layer and a Dist layer. There are a couple of problems. Full code, output, Manifest, Project: https://gist.github.com/12b7cfc21482419aa39e7e9dfbb7cc32 The...
This PR adds a ShardExpander filter, a filter that will take shard specs of the form "prefix-{000..999}.tar" and expand them into the 1000 corresponding file names, like the shell brace...
This PR adds a filter that allows keys to be renamed in training samples represented as dictionaries. This is particularly useful for webdataset-style data sets, but can also be used...
This PR adds an ExtractKeys filter that turns samples represented as dictionaries into tuples. Tuples are constructed by selecting values from the dictionaries by matching the key against a given...
When trying to insert a Unicode character in an interactive Julia REPL, it inserts it instead into a source code buffer. Reproduce: - open a Julia source file - start...
PyTorch is currently rearchitecting their I/O pipelines because the indexed datasets don't scale well to large learning problems. Many pipelines will likely be based on IterableDataset. The changes to PyTorch...
The OS X input method doesn't function as a normal keyboard. I think it should only switch to LaTeX mode when "\" is used, but it requires every word to...
I would find it useful if typedlua allowed defaults for arguments; this often is useful for being able to specify a more restrictive type, e.g. "string" instead of "string?" ```...
I'd like to use typedlua to compile code for use in regular, dynamically typed Lua applications. That is, most of the code that will be using my code as a...
When I have jupyter-autopep8 installed and enabled (via jupyter_notebook_extensions), when I try to open any other kernel, I get a popup that the autopep8 Jupyter notebook extension does not work...