yorko commited on
Commit
504460c
1 Parent(s): 8209bca

initial commit

Browse files
README.md CHANGED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # SciBERT Longformer finetuned to SDG classification
2
+
3
+ This is a Lonformer version of the [SciBERT uncased](https://huggingface.co/allenai/scibert_scivocab_uncased) model by Allen AI, finetuned to Sustainable Development Goals classification. The model is slower than SciBERT (~2.5x in my benchmarks) but can allow for 8x wider `max_seq_length` (4096 vs. 512) which is handy in case of working with long texts, e.g. scientific full texts.
4
+
5
+ The conversion to Longformer was performed with a [tutorial](https://github.com/allenai/longformer/blob/master/scripts/convert_model_to_long.ipynb) by Allen AI: see a [Google Colab Notebook](https://colab.research.google.com/drive/1NPTnMkeAYOF2MWH3_uJYesuxxdOzxrFn?usp=sharing) by [Yury](https://yorko.github.io/) which closely follows the tutorial.
6
+
7
+ Note:
8
+
9
+ - no additional MLM pretraining of the Longformer was performed, the [collab notebook](https://colab.research.google.com/drive/1NPTnMkeAYOF2MWH3_uJYesuxxdOzxrFn?usp=sharing) stops at step 3, and step 4 is not done. The model can be improved with this additional MLM pretraining, better to do so with scientific texts, e.g. [S@ORC](https://github.com/allenai/s2orc), again by Allen AI.
10
+ - no extensive benchmarks SciBERT Longformer vs. SciBERT were performed in terms of downstream task performance
11
+
12
+ Links:
13
+ - the original [SciBERT repo](https://github.com/allenai/scibert)
14
+ - the original [Longformer repo](https://github.com/allenai/longformer)
15
+
16
+
17
+ If using these models, please consider citing the following papers:
18
+ ```
19
+ @inproceedings{beltagy-etal-2019-scibert,
20
+ title = "SciBERT: A Pretrained Language Model for Scientific Text",
21
+ author = "Beltagy, Iz and Lo, Kyle and Cohan, Arman",
22
+ booktitle = "EMNLP",
23
+ year = "2019",
24
+ publisher = "Association for Computational Linguistics",
25
+ url = "https://www.aclweb.org/anthology/D19-1371"
26
+ }
27
+
28
+ @article{Beltagy2020Longformer,
29
+ title={Longformer: The Long-Document Transformer},
30
+ author={Iz Beltagy and Matthew E. Peters and Arman Cohan},
31
+ journal={arXiv:2004.05150},
32
+ year={2020},
33
+ }
34
+ ```
config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "BertForMaskedLM"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "attention_window": [
7
+ 512,
8
+ 512,
9
+ 512,
10
+ 512,
11
+ 512,
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512
19
+ ],
20
+ "gradient_checkpointing": false,
21
+ "hidden_act": "gelu",
22
+ "hidden_dropout_prob": 0.1,
23
+ "hidden_size": 768,
24
+ "initializer_range": 0.02,
25
+ "intermediate_size": 3072,
26
+ "layer_norm_eps": 1e-12,
27
+ "max_position_embeddings": 4096,
28
+ "model_type": "bert",
29
+ "num_attention_heads": 12,
30
+ "num_hidden_layers": 12,
31
+ "pad_token_id": 0,
32
+ "type_vocab_size": 2,
33
+ "vocab_size": 31090
34
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48db774448d6c458effaee86de8b6d656b6571f4de8df2f148542a6c7db8b7c7
3
+ size 450822016
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"model_max_length": 4096, "special_tokens_map_file": null, "full_tokenizer_file": null}
vocab.txt ADDED
The diff for this file is too large to render. See raw diff