Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on May 4, 2021

• Fix model names in README

  nikitakit

  committed
  24435a1

Commits on Feb 14, 2021

• Release v0.2.0

  Show description for 357b2bf
  nikitakit

  committed
  357b2bf

Commits on Feb 5, 2021

• Support spaCy v3.0 (in addition to v2)

  nikitakit

  committed
  4041dde

• New documentation and trained models for v0.2.0a0

  nikitakit

  committed
  23015e2

• Add --no-predict-tags flag for evaluation

  Show description for 192b306
  nikitakit

  committed
  192b306

• Delete code for exporting to tensorflow

  nikitakit

  committed
  01e5993

• New code to export models for release

  nikitakit

  committed
  4cb51d6

• Support Hebrew data in Unicode (no translit)

  nikitakit

  committed
  5eea505

• Auto-detect whether to always label the root node

  nikitakit

  committed
  547e548

• Better error messages when sentences are too long

  nikitakit

  committed
  28a62e6

• Major rewrite of benepar package for v0.2.0a0

  Show description for f376b21
  nikitakit

  committed
  f376b21

• Add decoding option to not label the root node

  nikitakit

  committed
  8198cf4

• Update EXPERIMENTS.md for new version of the code

  nikitakit

  committed
  d6407a6

• Don't require separate training files for raw text

  Show description for b751e0f
  nikitakit

  committed
  b751e0f

• Add CTB 5.1 data processing pipeline

  nikitakit

  committed
  9d4011f

• Add SPMRL data processing pipeline

  nikitakit

  committed
  f6ca5aa

• Check for sentence length mismatch

  nikitakit

  committed
  998fbdd

• Add --output-path flag for writing predicted trees

  nikitakit

  committed
  a95b8e3

• Fix output when provided POS tags contain '+' char

  nikitakit

  committed
  189dc1e

• Change hyperparameter defaults

  nikitakit

  committed
  06063fc

• Allow different names for the root symbol

  nikitakit

  committed
  ae6866a

• Support multi-token fields in CONLL-U input files

  nikitakit

  committed
  9c91c6b

• Support GPT-2 pretrained model

  nikitakit

  committed
  588d813

• Option to split pre-trained model across GPUs

  Show description for a42733e
  nikitakit

  committed
  a42733e

• Support more pre-trained model types

  Show description for fa02356
  nikitakit

  committed
  fa02356

• Major re-write and cleanup of the pytorch code

  Show description for 9b58e09
  nikitakit

  committed
  9b58e09

• Add Penn Treebank data processing pipeline

  Show description for c8db611
  nikitakit

  committed
  c8db611

• Delete unused viz folder and subcommand

  nikitakit

  committed
  3cee9f1

• Release v0.1.3

  nikitakit

  committed
  148b6fc

Commits on Jul 1, 2019

• Fix import to support pytorch v1

  nikitakit

  committed
  1ee43a8

Commits on Jun 28, 2019

• Better sub-batching when using BERT

  nikitakit

  committed
  9ec778f

• Add tree_from_str for loading a single tree

  nikitakit

  committed
  9864130

• Additional un-escaping during training

  nikitakit

  committed
  589825e

• Better generalization to text with en/em dashes

  nikitakit

  committed
  6f9094a

Commits on Feb 10, 2019

• Link to Hungarian models for spaCy in README (nikitakit#17)

  nikitakit

  committed
  d58e8c4

Pagination

PreviousNext