Here we share a pretrained BERT model that is aware of math tokens. The math tokens are treated specially and tokenized using
pya0
, which adds very limited new tokens for latex markup (total vocabulary is just 31,061).
This model is trained on 4 x 2 Tesla V100 with a total batch size of 64, using Math StackExchange data with 2.7 million sentence pairs trained for 7 epochs.
Modify the test examples in
test.txt
to play with it.
The test file is tab-separated, the first column is additional positions you want to mask for the right-side sentence (useful for masking tokens in math markups). A zero means no additional mask positions.
azbert-base huggingface.co is an AI model on huggingface.co that provides azbert-base's model effect (), which can be used instantly with this castorini azbert-base model. huggingface.co supports a free trial of the azbert-base model, and also provides paid use of the azbert-base. Support call azbert-base model through api, including Node.js, Python, http.
azbert-base huggingface.co is an online trial and call api platform, which integrates azbert-base's modeling effects, including api services, and provides a free online trial of azbert-base, you can try azbert-base online for free by clicking the link below.
castorini azbert-base online free url in huggingface.co:
azbert-base is an open source model from GitHub that offers a free installation service, and any user can find azbert-base on GitHub to install. At the same time, huggingface.co provides the effect of azbert-base install, users can directly use azbert-base installed effect in huggingface.co for debugging and trial. It also supports api for free installation.