Add model
Browse files- README.md +29 -0
- checkpoint/info.json +1 -0
- checkpoint/model.pth +3 -0
- config.json +1 -0
- definition.json +1 -0
README.md
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: xpmir
|
| 3 |
+
---
|
| 4 |
+
# monoBERT trained on MS-Marco
|
| 5 |
+
|
| 6 |
+
Passage Re-ranking with BERT (Rodrigo Nogueira, Kyunghyun Cho). 2019.
|
| 7 |
+
https://arxiv.org/abs/1901.04085
|
| 8 |
+
|
| 9 |
+
## Using the model)
|
| 10 |
+
The model can be loaded with [experimaestro IR](https://experimaestro-ir.readthedocs.io/en/latest/)
|
| 11 |
+
|
| 12 |
+
```py
|
| 13 |
+
from xpmir.models import AutoModel
|
| 14 |
+
|
| 15 |
+
# Model that can be re-used in experiments
|
| 16 |
+
model = AutoModel.load_from_hf_hub("xpmir/monobert")
|
| 17 |
+
|
| 18 |
+
# Use this if you want to actually use the model
|
| 19 |
+
model = AutoModel.load_from_hf_hub("xpmir/monobert", as_instance=True)
|
| 20 |
+
model.initialize(None)
|
| 21 |
+
model.rsv("walgreens store sales average", "The average Walgreens salary ranges from approximately $15,000 per year for Customer Service Associate / Cashier to $179,900 per year for District Manager...")
|
| 22 |
+
```
|
| 23 |
+
|
| 24 |
+
## Results
|
| 25 |
+
| Dataset | AP | P@20 | RR | RR@10 | nDCG | nDCG@10 | nDCG@20 |
|
| 26 |
+
|----| ---|------|------|------|------|------|------|
|
| 27 |
+
| msmarco_dev | 0.3574 | 0.0371 | 0.3624 | 0.3529 | 0.4640 | 0.4147 | 0.4370 |
|
| 28 |
+
| trec2019 | 0.4908 | 0.7233 | 0.9368 | 0.9368 | 0.6871 | 0.7046 | 0.6813 |
|
| 29 |
+
| trec2020 | 0.4803 | 0.6120 | 0.9380 | 0.9367 | 0.6865 | 0.6963 | 0.6626 |
|
checkpoint/info.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"epoch": 1504, "steps": 48128}
|
checkpoint/model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8803c9e357bceca55b6ff67cc08c2b0cf6081c1065c2a4804a1f1fd2af792d3
|
| 3 |
+
size 438008465
|
config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{}
|
definition.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
[{"id": 140279831147280, "module": "xpmir.text.huggingface", "type": "DualTransformerEncoder", "identifier": "945cb146c597c06064946f11253b127a1152c137ca7155d220107043d7fb55ae", "typename": "xpmir.text.huggingface.DualTransformerEncoder", "ignored": [], "fields": {"model_id": "bert-base-uncased", "trainable": true, "layer": 0, "dropout": 0.1, "maxlen": 512, "version": 2}}, {"id": 140279831215504, "module": "xpmir.neural.cross", "type": "CrossScorer", "identifier": "09134f4b01bc9472d5573db7abaa489b0f6d71dd3dec2826771f88c300870bc7", "fields": {"checkpoint": {"type": "path.serialized", "value": "checkpoint", "is_folder": true}, "encoder": {"type": "python", "value": 140279831147280}}}]
|