Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[docs] Fix the prompt link to the training script #3060

Merged
merged 1 commit into from
Nov 15, 2024
Merged

[docs] Fix the prompt link to the training script #3060

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion examples/training/prompts/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -96,7 +96,7 @@ Additionally, some research papers (`INSTRUCTOR <https://arxiv.org/abs/2212.0974
```eval_rst
See the following script as an example of how to train with prompts in practice:

* `training_nq_prompts.py <training_nq_prompts.py>`_: This script finetunes `mpnet-base <https://huggingface.co/microsoft/mpnet-base>`_ on 100k query-answer pairs from the `natural-questions <https://huggingface.co/datasets/sentence-transformers/natural-questions>`_ dataset using the :class:`~sentence_transformers.losses.CachedMultipleNegativesRankingLoss` loss. The model is evaluated during training using the :class:`~sentence_transformers.evaluation.NanoBEIREvaluator`.
* `training_nq_prompts.py <https://github.com/UKPLab/sentence-transformers/blob/master/examples/training/prompts/training_nq_prompts.py>`_: This script finetunes `mpnet-base <https://huggingface.co/microsoft/mpnet-base>`_ on 100k query-answer pairs from the `natural-questions <https://huggingface.co/datasets/sentence-transformers/natural-questions>`_ dataset using the :class:`~sentence_transformers.losses.CachedMultipleNegativesRankingLoss` loss. The model is evaluated during training using the :class:`~sentence_transformers.evaluation.NanoBEIREvaluator`.

This script has two variables that affect 1) whether prompts are used and 2) whether prompts are included in the pooling. I have finetuned both ``mpnet-base`` and ``bert-base-uncased`` under the various different settings, resulting in a 0.66% and 0.90% relative improvements on ``NDCG@10`` at no extra cost.

Expand Down