google-research / google-research/language

Inquiry about checkpoint of ORQA

Open
#173 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
1.8k
Forks
362
PR merge metrics
No merged PRs in 30d

Description

I'm reaching out to inquire about the pre-trained ICT models, specifically the weights for both the question encoder BERT_Q(q) and the block encoder BERT_B(b).

I came across the repository "https://github.com/google-research/language/tree/master/language/orqa" on GitHub, which suggests the site "gs://orqa-data/ict" as a potential source for these weights.
Upon checking, I found that "gs://orqa-data/ict" provides the pre-trained weight for only the question encoder BERT_Q(q) and a dense vector index of size (13353718, 128).
To be precise, the weight for BERT_Q(q) is located in "gs://orqa-data/ict/variables", and the dense vector index is in "gs://orqa-data/ict/encoded". However, I couldn't locate the pre-trained weight for the block encoder BERT_B(b).

Would it be possible for you to share the pre-trained weight of the block encoder BERT_B(b), if available?

I appreciate your assistance.

Thank you.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.