Shi Yu

Error loading model in `query_gen_and_train.py`

1

Hello, I'm working on the GenQ setting and find that the model loading part seems incorrect. In `query_gen_and_train.py`, the model loading part is like: ```python #### Provide any sentence-transformers model...

Why to scale the loss when DDP trainning

5

Dear Luyu, Thank you for the great package simplifying the cumbersome DR training & evaluation process! I try using it in DDP mode and I read the code. I have...

[Feature request] Indexing datasets by a customly-defined id field to enable random access dataset items via the id

7

### Feature request Some datasets may contain an id-like field, for example the `id` field in [wikimedia/wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia) and the `_id` field in [BeIR/dbpedia-entity](https://huggingface.co/datasets/BeIR/dbpedia-entity). HF datasets support efficient random access via...

It seems that the 6th query "Which is the better type?" from topic 81, CAsT-20 is mistakenly put in `automatic_evaluation_topics_annotated_v1.1.json` since the query doesn't exist in either `2020_automatic_evaluation_topics_v1.0.json` or `2020_manual_evaluation_topics_v1.0.json`.

ANCE uses a layer norm over the embeddings: https://github.com/microsoft/ANCE/blob/master/model/models.py#L146

`strip` should be `rstrip` to avoid removing the leading `/` in the local path