texar-pytorch
texar-pytorch copied to clipboard
asyml
Integrating the Best of TF into PyTorch, for Machine Learning, Natural Language Processing, and Text Generation. This is part of the CASL project: http://casl-project.ai/
Hi, while trying to use the following snippet: ```python import texar.torch as tx from texar.torch.run import * # (1) Modeling class BERTGPT2Model(nn.Module): """An encoder-decoder model with GPT-2 as the decoder."""...
- [ ] `TransformerDecoder.forward`: where does `self.training` come from? https://github.com/asyml/texar-pytorch/blob/d17d502b50da1d95cb70435ed21c6603370ce76d/texar/torch/modules/decoders/transformer_decoders.py#L448-L449 - [ ] All arguments should say their types **explicitly** in the docstring. E.g., what is the type of `infer_mode`?...
Provide a general-purpose example that can test the correctness of all pre-trained modules and their corresponding tokenizers.
Passage Re-ranking with BERT https://github.com/nyu-dl/dl4marco-bert
Hi, Thanks for adding vae_text example. I have a question related to MLPTransformConnector. [MLPTransformConnector ]( https://github.com/asyml/texar-pytorch/blob/43967ee238a5da2e996f4f71644a940a86cad009/examples/vae_text/vae_train.py#L115)expects an input of size `self.encoder.cell.hidden_size*2`. I don't understand why it should be multiplied by...
Does Texar support multi-gpus?
In a few places Texar-PyTorch would print out information (e.g., "download complete", "pre-trained weights loaded") using the Python builtin `print` function. The behavior might not be desired and is not...
There are several problems with the current data modules: 1. Filtering has to happen at data source level. This means one can't discard examples in `process`, but instead must supply...
The section is titled "Data Loaders" https://texar-pytorch.readthedocs.io/en/latest/code/data.html#data-loaders Would "Datasets" be better? Or does "Data Loaders" fit the Pytorch convention better? @huzecong @AvinashBukkittu