-
### To Reproduce
My hparams:
```python
{
'n': [8000],
'k': [30],
'batch_size': 512,
'data_dir': '/Users/kyoungrok/Resilio Sync/Dataset/2019 TREC/passage_ranking/dataset',
'max_nb_epoc…
-
In run_qe_classifier.py there is only accuracy when evaluating the model. Can you provide any code on how you do nDCG@20 evaluation in your paper?
-
Hi @emmileaf I'm getting these MS MARCO passage regression errors:
This is on `tuna`:
```
2019-08-11 03:55:02,107 - regression_test - ERROR - !!!!!{"actual": 0.1518, "collection": "msmarco-pass…
-
Hi,
Just discovered your code.
Seems super interesting, in particular the TK model.
I am wondering if there is a paper on TK model?
May be some scores and info on speed? (I have found nothing …
-
I might miss something but could you grant access to the above bucket? so that I can run experiments. Thank you.
-
-
I am new to text retrieval. The two tasks seems are similar.
Thank you very much.
-
Full ranking means scoring all the query-passage pairs.
Thank you!
@rodrigonogueira4
-
There's some ideas in #456 which are worth pursuing to improve fallback performance. Specifically, it added caching for results of various fontconfig operations like `font_sort`. Additionally, it migh…
jwilm updated
4 years ago
-
I reproduce VanillaBERT
but only get the result
****NDCG@20**:0.3889
**P@20** :0.3180
**optimizer** :AdamW
**batch size** :1
lr = 1e-5**
Train by HingeLoss and
by following code's pro…