Update README.md
Browse files
README.md
CHANGED
|
@@ -18,10 +18,77 @@ Use https://github.com/nlmatics/nlm-utils
|
|
| 18 |
|
| 19 |
Click on each model to see details:
|
| 20 |
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
Click on each model to see details:
|
| 20 |
|
| 21 |
+
### roberta.large.boolq
|
| 22 |
+
|
| 23 |
+
Location: [roberta.large.boolq](https://huggingface.co/ansukla/roberta/tree/main/roberta.large.boolq)
|
| 24 |
+
|
| 25 |
+
Trained with MNLI + Boolq
|
| 26 |
+
|
| 27 |
+
Trained by: Evan Li
|
| 28 |
+
|
| 29 |
+
Application: Given a passage and a question, answer the question with yes, no or unsure.
|
| 30 |
+
|
| 31 |
+
Training Process: https://blogs.nlmatics.com/2020/03/12/Boolean-Question-Answering-with-Neutral-Labels.html
|
| 32 |
+
|
| 33 |
+
### roberta.large.qa
|
| 34 |
+
See folder: [roberta.large.qa](https://huggingface.co/ansukla/roberta/tree/main/roberta.large.qa)
|
| 35 |
+
|
| 36 |
+
Trained with SQuAD 2.0 + Custom Dataset preferring shorter spans better suited for data extraction
|
| 37 |
+
|
| 38 |
+
Trained by: Ambika Sukla
|
| 39 |
+
|
| 40 |
+
Application: Given a passage and a question, pick the shortest span from the passage that answers the question
|
| 41 |
+
|
| 42 |
+
Training Process: start, end location head on the top of Roberta Base
|
| 43 |
+
|
| 44 |
+
### roberta.large.stsb
|
| 45 |
+
See folder: [roberta.large.stsb](https://huggingface.co/ansukla/roberta/tree/main/roberta.large.stsb)
|
| 46 |
+
|
| 47 |
+
Trained with STSB dataset
|
| 48 |
+
|
| 49 |
+
Trained by: Meta/Fairseq
|
| 50 |
+
|
| 51 |
+
Application: Given two passages, return a score beteen 0 and 1 to evaluate their similarity
|
| 52 |
+
|
| 53 |
+
Training Process: regression head on top of Roberta Base
|
| 54 |
+
|
| 55 |
+
### roberta.large.phraseqa
|
| 56 |
+
See folder: [roberta.large.phraseqa](https://huggingface.co/ansukla/roberta/tree/main/roberta.large.phraseqa)
|
| 57 |
+
|
| 58 |
+
Trained with Roberta 2.0 with the question words removed from the question
|
| 59 |
+
|
| 60 |
+
Trained By: Batya Stein
|
| 61 |
+
|
| 62 |
+
Application: Given a passage and phrase (key), extract a value from the passage
|
| 63 |
+
|
| 64 |
+
Training Process: https://blogs.nlmatics.com/2020/08/25/Optimizing-Transformer-Q&A-Models-for-Naturalistic-Search.html
|
| 65 |
+
|
| 66 |
+
### roberta.large.qasrl
|
| 67 |
+
|
| 68 |
+
See folder: [roberta.large.qasrl](https://huggingface.co/ansukla/roberta/tree/main/roberta.large.qasrl)
|
| 69 |
+
|
| 70 |
+
Trained with QASRL dataset
|
| 71 |
+
|
| 72 |
+
Application: Given a passage, get back values for who, what, when, where etc.
|
| 73 |
+
|
| 74 |
+
Trained By: Nima Sheikholeslami
|
| 75 |
+
|
| 76 |
+
### roberta.large.qatype.lower.RothWithQ
|
| 77 |
+
|
| 78 |
+
See folder: [roberta.large.qatype.lower.RothWithQ](https://huggingface.co/ansukla/roberta/tree/main/roberta.large.qatype.lower.RothWithQ)
|
| 79 |
+
|
| 80 |
+
Trained with the Roth Question Type dataset.
|
| 81 |
+
|
| 82 |
+
Application: Given a question, return one of the answer types e.g. number, location. See the Roth dataset for full list.
|
| 83 |
+
|
| 84 |
+
Trained By: Evan Li
|
| 85 |
+
|
| 86 |
+
### roberta.large.io_qa
|
| 87 |
+
|
| 88 |
+
See folder: [roberta.large.io_qa](https://huggingface.co/ansukla/roberta/tree/main/roberta.large.io_qa)
|
| 89 |
+
Trained with SQuAD 2.0 dataset
|
| 90 |
+
|
| 91 |
+
Trained By: Nima Sheikholeslami
|
| 92 |
+
|
| 93 |
+
Training Process: Use io head to support multiple spans.
|
| 94 |
+
|