tensorflow / text

9 Pull requests merged by 2 people

• Support Android build for tf.text

  #384 merged Sep 25, 2020

• Change variable names for token offsets: "limit" -> "end".

  #406 merged Sep 24, 2020

• Extend documentation for class HubModuleSplitter.

  #405 merged Sep 23, 2020

• Update benchmark utils and add internal benchmarks.

  #404 merged Sep 22, 2020

• Add class HubModuleSplitter, a Hub module-based Splitter.

  #403 merged Sep 22, 2020

• End offset is preserved when an unknonwn token of long size is found.

  #380 merged Sep 22, 2020

• Update setup.py. TensorFlow has switched to the default package being GPU, and having users explicitly call out when wanting just CPU.

  #400 merged Sep 22, 2020

• Add ops benchmarks to the continuous monitoring

  #398 merged Sep 21, 2020

• Add Spliter / SplitterWithOffsets abstract base classes.

  #399 merged Sep 21, 2020

1 Pull request opened by 1 person

• Allow for WhitespaceTokenizer.tokenize to be convertible to TF.Lite.

  #401 opened Sep 22, 2020

2 Issues closed by 2 people

• How to generate test_oss_model.model file?

  #394 closed Sep 25, 2020

• import tensorflow_text gives errors in 2.3.0 on mac

  #402 closed Sep 22, 2020

3 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• tensorflow.python.framework.errors_impl.NotFoundError: /usr/local/lib/python3.6/dist-packages/tensorflow_text/python/metrics/_text_similarity_metric_ops.so: undefined symbol: _ZN10tensorflow8OpKernel11TraceStringEPNS_15OpKernelContextEb

  #385 commented on Sep 22, 2020 • 2 new comments

• End offset in `wordpiece_tokenizer` inacurrate when `max_bytes_per_token` is set along with `use_unknown_token`

  #344 commented on Sep 22, 2020 • 1 new comment

• Problem with wheels in older OSX versions

  #375 commented on Sep 26, 2020 • 1 new comment