All our software is available on GitHub.
Genie was previously released under the Almond moniker. The last release of Almond includes the following components:
See the release notes and individual module changelogs for details. Additional pre-releases and unstable software are also available on GitHub.
Datasets associated with the following papers are also available:
Findings of EMNLP 2023
Please see the instructions at https://github.com/stanford-oval/WikiChat for code, models and datasets.
EMNLP 2023
Please see the instructions at https://github.com/stanford-oval/wikidata-emnlp23 for code, models and datasets.
Findings of ACL 2023
Please see the instructions at https://github.com/stanford-oval/dialogues to download the dataset, models, and software.
EACL 2023
Please see the instructions at https://github.com/stanford-oval/dialogues to download the dataset, models, and software.
EACL 2023
Please see the instructions at https://github.com/stanford-oval/dialogues to download the dataset, models, and software.
Findings of ACL 2022
Please see the instructions at https://github.com/stanford-oval/multiwoz-acl2022 to download the dataset, models, and software.
EMNLP 2020
Please see the instructions at https://github.com/stanford-oval/SPL to download the dataset, models, and software.
A new version of Schema2QA dataset has been released and AutoQA result has been updated as well, check here.
EMNLP 2020
train.tsv
files contain automatically synthesized training data. eval.tsv
files contain crowdsourced validation data.train.tsv
files contain automatically synthesized training data. eval.tsv
and test.tsv
files contain human-paraphrased validation and test data.A new version of Schema2QA dataset has been released with new model and result, check here.
CIKM 2020
Please download the dataset and models at https://oval.cs.stanford.edu/releases/cikm2020/
SHA256 sums of the dataset and models
55f9bbcbfe3ed94adc0f23f9c957a58ce64b87b507603fe62f71847ff8e0c856 train.zip
2e5278a7f467a23bddf2d3f82fcf11b95c0e14eb79968c8b90fc1f2f4081af47 paraphrase.zip
5356dcde2bca026ecfeb69dc5c0cc21b84e67553434129531b68f2f138b769cf eval.zip
9dd46cd51552f8a862f0cb463148ff79478ca89388465f1a405a812eaf1d171d models.zip
ACL 2020
Please use the instructions at https://github.com/stanford-oval/zero-shot-multiwoz-acl2020 to download the dataset and software necessary to reproduce this paper.
PLDI 2019
SHA256 sums of the artifacts:
c58209727093ed0b395c50fb7698edec398c1fa129b0bd79bd343108f512455a pldi19-artifact.tar.xz
68a0e7d24b7d9ebe6ccf58cc837c61fe8a0b5aa6649f769e62327b4a2d556186 pldi19-dataset.tar.xz
To download the continuously updated, live version of the main Genie natural language model, use the following links: