Initial version.

README.md

@@ -0,0 +1,63 @@
+---
+tags:
+- bert
+- adapterhub:nli/multinli
+- adapter-transformers
+datasets:
+- multi_nli
+language:
+- en
+---
+
+# Adapter `AdapterHub/bert-base-uncased-pf-mnli` for bert-base-uncased
+
+An [adapter](https://adapterhub.ml) for the `bert-base-uncased` model that was trained on the [nli/multinli](https://adapterhub.ml/explore/nli/multinli/) dataset and includes a prediction head for classification.
+
+This adapter was created for usage with the **[adapter-transformers](https://github.com/Adapter-Hub/adapter-transformers)** library.
+
+## Usage
+
+First, install `adapter-transformers`:
+
+```
+pip install -U adapter-transformers
+```
+_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)_
+
+Now, the adapter can be loaded and activated like this:
+
+```python
+from transformers import AutoModelWithHeads
+
+model = AutoModelWithHeads.from_pretrained("bert-base-uncased")
+adapter_name = model.load_adapter("AdapterHub/bert-base-uncased-pf-mnli", source="hf")
+model.active_adapters = adapter_name
+```
+
+## Architecture & Training
+
+The training code for this adapter is available at https://github.com/adapter-hub/efficient-task-transfer.
+In particular, training configurations for all tasks can be found [here](https://github.com/adapter-hub/efficient-task-transfer/tree/master/run_configs).
+
+
+## Evaluation results
+
+Refer to [the paper](https://arxiv.org/pdf/2104.08247) for more information on results.
+
+## Citation
+
+If you use this adapter, please cite our paper ["What to Pre-Train on? Efficient Intermediate Task Selection"](https://arxiv.org/pdf/2104.08247):
+
+```bibtex
+@inproceedings{poth-etal-2021-what-to-pre-train-on,
+    title={What to Pre-Train on? Efficient Intermediate Task Selection},
+    author={Clifton Poth and Jonas Pfeiffer and Andreas Rücklé and Iryna Gurevych},
+    booktitle = "Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
+    month = nov,
+    year = "2021",
+    address = "Online",
+    publisher = "Association for Computational Linguistics",
+    url = "https://arxiv.org/abs/2104.08247",
+    pages = "to appear",
+}
+```

adapter_config.json

@@ -0,0 +1,23 @@
+{
+  "config": {
+    "adapter_residual_before_ln": false,
+    "cross_adapter": false,
+    "inv_adapter": null,
+    "inv_adapter_reduction_factor": null,
+    "leave_out": [],
+    "ln_after": false,
+    "ln_before": false,
+    "mh_adapter": false,
+    "non_linearity": "relu",
+    "original_ln_after": true,
+    "original_ln_before": true,
+    "output_adapter": true,
+    "reduction_factor": 16,
+    "residual_before_ln": true
+  },
+  "hidden_size": 768,
+  "model_class": "BertModelWithHeads",
+  "model_name": "bert-base-uncased",
+  "model_type": "bert",
+  "name": "glue_mnli"
+}

head_config.json

@@ -0,0 +1,20 @@
+{
+  "config": {
+    "activation_function": "tanh",
+    "bias": true,
+    "head_type": "classification",
+    "label2id": {
+      "contradiction": 2,
+      "entailment": 0,
+      "neutral": 1
+    },
+    "layers": 2,
+    "num_labels": 3,
+    "use_pooler": false
+  },
+  "hidden_size": 768,
+  "model_class": "BertModelWithHeads",
+  "model_name": "bert-base-uncased",
+  "model_type": "bert",
+  "name": "glue_mnli"
+}

pytorch_adapter.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:784fae664e1c22e8ea26a897d1303a3cb9151127ec2cf87e55fa53de1b15fe60
+size 3594799

pytorch_model_head.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e85951319c78948dfae57e3e1eed6a190772125385b6661c933160563e276eac
+size 2373247