calpt commited on
Commit
3b1a93d
1 Parent(s): 2dbd4fa

Initial version.

Browse files

README.md ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - bert
4
+ - adapterhub:qa/squad2
5
+ - adapter-transformers
6
+ datasets:
7
+ - squad_v2
8
+ language:
9
+ - en
10
+ ---
11
+
12
+ # Adapter `AdapterHub/bert-base-uncased-pf-squad_v2` for bert-base-uncased
13
+
14
+ An [adapter](https://adapterhub.ml) for the `bert-base-uncased` model that was trained on the [qa/squad2](https://adapterhub.ml/explore/qa/squad2/) dataset and includes a prediction head for question answering.
15
+
16
+ This adapter was created for usage with the **[adapter-transformers](https://github.com/Adapter-Hub/adapter-transformers)** library.
17
+
18
+ ## Usage
19
+
20
+ First, install `adapter-transformers`:
21
+
22
+ ```
23
+ pip install -U adapter-transformers
24
+ ```
25
+ _Note: adapter-transformers is a fork of transformers that acts as a drop-in replacement with adapter support. [More](https://docs.adapterhub.ml/installation.html)_
26
+
27
+ Now, the adapter can be loaded and activated like this:
28
+
29
+ ```python
30
+ from transformers import AutoModelWithHeads
31
+
32
+ model = AutoModelWithHeads.from_pretrained("bert-base-uncased")
33
+ adapter_name = model.load_adapter("AdapterHub/bert-base-uncased-pf-squad_v2", source="hf")
34
+ model.active_adapters = adapter_name
35
+ ```
36
+
37
+ ## Architecture & Training
38
+
39
+ The training code for this adapter is available at https://github.com/adapter-hub/efficient-task-transfer.
40
+ In particular, training configurations for all tasks can be found [here](https://github.com/adapter-hub/efficient-task-transfer/tree/master/run_configs).
41
+
42
+
43
+ ## Evaluation results
44
+
45
+ Refer to [the paper](https://arxiv.org/pdf/2104.08247) for more information on results.
46
+
47
+ ## Citation
48
+
49
+ If you use this adapter, please cite our paper ["What to Pre-Train on? Efficient Intermediate Task Selection"](https://arxiv.org/pdf/2104.08247):
50
+
51
+ ```bibtex
52
+ @inproceedings{poth-etal-2021-what-to-pre-train-on,
53
+ title={What to Pre-Train on? Efficient Intermediate Task Selection},
54
+ author={Clifton Poth and Jonas Pfeiffer and Andreas Rücklé and Iryna Gurevych},
55
+ booktitle = "Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
56
+ month = nov,
57
+ year = "2021",
58
+ address = "Online",
59
+ publisher = "Association for Computational Linguistics",
60
+ url = "https://arxiv.org/abs/2104.08247",
61
+ pages = "to appear",
62
+ }
63
+ ```
adapter_config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "adapter_residual_before_ln": false,
4
+ "cross_adapter": false,
5
+ "inv_adapter": null,
6
+ "inv_adapter_reduction_factor": null,
7
+ "leave_out": [],
8
+ "ln_after": false,
9
+ "ln_before": false,
10
+ "mh_adapter": false,
11
+ "non_linearity": "relu",
12
+ "original_ln_after": true,
13
+ "original_ln_before": true,
14
+ "output_adapter": true,
15
+ "reduction_factor": 16,
16
+ "residual_before_ln": true
17
+ },
18
+ "hidden_size": 768,
19
+ "model_class": "BertModelWithHeads",
20
+ "model_name": "bert-base-uncased",
21
+ "model_type": "bert",
22
+ "name": "squad_v2"
23
+ }
head_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "activation_function": "tanh",
4
+ "head_type": "question_answering",
5
+ "label2id": {
6
+ "LABEL_0": 0,
7
+ "LABEL_1": 1
8
+ },
9
+ "layers": 1,
10
+ "num_labels": 2
11
+ },
12
+ "hidden_size": 768,
13
+ "model_class": "BertModelWithHeads",
14
+ "model_name": "bert-base-uncased",
15
+ "model_type": "bert",
16
+ "name": "squad_v2"
17
+ }
pytorch_adapter.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5fc4eeca5f27553522a63f6fa93a564416d40a3bf4b36eac29200e12fff238c5
3
+ size 3594735
pytorch_model_head.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89c2c09ac50cd01489b967bdeaf029e2635afee9ea8fcebdd8c8b5e0ffb2ab67
3
+ size 7223