Issues
- 1
Performance issue in the difinition of zero_state, layers/eight_mile/tf/layers.py
#856 opened by DLPerf - 2
- 0
Solve TF eager mode serving
#530 opened by dpressel - 1
Add support to export to pytorch/serve
#569 opened by dpressel - 0
Update TF serving docs
#573 opened by dpressel - 0
- 0
Use stdlib csv tools for quotes
#734 opened by dpressel - 0
Better SoC for LRS + make easier to use
#653 opened by dpressel - 0
More logging points
#655 opened by dpressel - 0
make pair classification use `make_input`
#733 opened by dpressel - 2
- 1
Make TF default to eager mode
#619 opened by dpressel - 0
- 1
Data cache order problem
#697 opened by dpressel - 0
Pruning ONNX inputs
#717 opened by blester125 - 0
Fix inconsistent naming
#708 opened by blester125 - 1
Update TensorBoard reporting hooks
#654 opened by dpressel - 1
- 1
Fix TF LRS quirks
#639 opened by dpressel - 3
Test/update API examples
#585 opened by dpressel - 0
- 0
seq2seq inference example was broken in TF2
#622 opened by dpressel - 0
seq2seq TF1 broken on change to embeddings for mixin
#623 opened by dpressel - 0
seq2seq TF1 transformer API breakage
#625 opened by dpressel - 1
Layer weight conversion compat matrix for between TensorFlow and PyTorch layers
#533 opened by dpressel - 1
Add support back for LMs which generate a different feature than the input ones
#602 opened by dpressel - 7
Test/update mead configs
#581 opened by dpressel - 0
- 1
Update hub docs
#526 opened by dpressel - 1
Update mead docs for 2.0
#527 opened by dpressel - 1
change BERT tagger examples to use core library
#589 opened by dpressel - 1
Update vecs.json to container BERT wordpiece models
#583 opened by dpressel - 0
updates for tf distribute mode
#578 opened by dpressel - 5
How to reload models that use hub addons.
#496 opened by blester125 - 1
ONNX compat matrix of all PyTorch core models
#532 opened by dpressel - 0
- 0
- 2
Loading BERT checkpoints into TLM from TensorFlow
#528 opened by dpressel - 1
Run black on code
#561 opened by dpressel - 1
Simplify Embeddings
#546 opened by dpressel - 3
Try and handle token type embeddings more elegantly
#536 opened by dpressel - 1
- 0
Remove estimator fit functions
#555 opened by dpressel - 1
- 0
Update LM API to be consistent with the others
#553 opened by dpressel - 0
Move all 3rd party lib dependent addons to hub
#531 opened by dpressel - 1
- 0
Simplify baseline models by not using 8mi versions
#541 opened by dpressel - 0
- 0