Emil Lynegaard
Emil Lynegaard
For the model I am training, I am relying on a custom [Sampler](https://pytorch.org/docs/stable/data.html#torch.utils.data.Sampler), that returns variable batch sizes. My task at hand is translation, where I following [Attention is all...
I was unable to figure out exactly whether or not the Verify API was supported. It feels as though some of it lies in the `ExTwilio.ProgrammableChat.Service` module, but it is...
I apologize if this is somehow fixed in a newer version (I am using 4.5.4) but it looks like no changes have been made to any queuing logic recently: https://github.com/modernmt/modernmt/tree/master/src/decoder-neural/src/main/java/eu/modernmt/decoder/neural/queue...
If running MMT on a machine with multiple file systems, eg. could be a mounted volume for the runtime directory, using `os.rename` causes the `./mmt create` command to fail with...
Currently, the early stopping is implemented by looking at the TensorboardX loss at every N for `--save-interval-updates N`. This results in unexpected behavior if one tries to disable interval-based checkpoints,...
Right now, the Java code gets the available GPUs [using nvidia-smi](https://github.com/modernmt/modernmt/blob/fdea12c13699c9126a86cdf3233548536fbe61b0/src/commons/src/main/java/eu/modernmt/hw/Graphics.java#L29b), which does not respect the `CUDA_VISIBLE_DEVICES` envvar, that is typically used to hide/show GPUs. While I can specifiy `--gpus`...
As briefly discussed in #515 with @nicolabertoldi , `nbest` is not supported in ModernMT 4.x. While I can see the purposes of this not always being obvious, I will give...
I am slightly confused with regards to what the correct use of `"pyls.plugins.jedi.environment`" and `"pyls.plugins.jedi.extra_paths"`, and how I can use them to get consistent completion both when I have a...
We want to make sure the new configuraiton works on both dev/prod. Requires the `distillerty_configuraiton` branch to be merged, which it is ready for.