scherrmann commited on
Commit
23affb0
·
1 Parent(s): 11b0183

Rename README(1).md to README.md

Browse files
Files changed (1) hide show
  1. README(1).md → README.md +8 -8
README(1).md → README.md RENAMED
@@ -3,19 +3,19 @@ license: apache-2.0
3
  language:
4
  - de
5
  ---
6
- # German FinBERT For QuAD (Further Pre-trained Version, Fine-Tuned for Financial Question Answering)
7
 
8
  German FinBERT is a BERT language model focusing on the financial domain within the German language. In my [paper](https://arxiv.org/pdf/2311.08793.pdf), I describe in more detail the steps taken to train the model and show that it outperforms its generic benchmarks for finance specific downstream tasks.
9
 
10
- This model is the [further-pretrained version of German FinBERT](https://huggingface.co/scherrmann/GermanFinBert_FP), after fine-tuning on the [German Ad-Hoc QuAD dataset](https://huggingface.co/datasets/scherrmann/adhoc_quad).
11
 
12
  ## Overview
13
  **Author** Moritz Scherrmann
14
  **Paper:** [here](https://arxiv.org/pdf/2311.08793.pdf)
15
  **Architecture:** BERT base
16
  **Language:** German
17
- **Specialization:** Financial question answering
18
- **Base model:** [German_FinBert_FP](https://huggingface.co/scherrmann/GermanFinBert_FP)
19
 
20
 
21
  ### Fine-tuning
@@ -27,9 +27,9 @@ After finding the best model w.r.t the evaluation set, I report the mean result
27
 
28
  ### Results
29
 
30
- Ad-Hoc QuAD (Question Answering):
31
- - Exact Match (EM): 52.50%
32
- - F1 Score: 74.61%
33
 
34
 
35
  ## Authors
@@ -41,4 +41,4 @@ For additional details regarding the performance on fine-tune datasets and bench
41
  See also:
42
  - scherrmann/GermanFinBERT_SC
43
  - scherrmann/GermanFinBERT_FP
44
- - scherrmann/GermanFinBERT_SC_Sentiment
 
3
  language:
4
  - de
5
  ---
6
+ # German FinBERT For QuAD (Pre-trained From Scratch Version, Fine-Tuned for Financial Question Answering)
7
 
8
  German FinBERT is a BERT language model focusing on the financial domain within the German language. In my [paper](https://arxiv.org/pdf/2311.08793.pdf), I describe in more detail the steps taken to train the model and show that it outperforms its generic benchmarks for finance specific downstream tasks.
9
 
10
+ This model is the [pre-trained from scratch version of German FinBERT](https://huggingface.co/scherrmann/GermanFinBert_SC), after fine-tuning on a translated version of the [financial news phrase bank](https://arxiv.org/abs/1307.5336) of Malo et al. (2013).
11
 
12
  ## Overview
13
  **Author** Moritz Scherrmann
14
  **Paper:** [here](https://arxiv.org/pdf/2311.08793.pdf)
15
  **Architecture:** BERT base
16
  **Language:** German
17
+ **Specialization:** Financial sentiment
18
+ **Base model:** [German_FinBert_SC](https://huggingface.co/scherrmann/GermanFinBert_SC)
19
 
20
 
21
  ### Fine-tuning
 
27
 
28
  ### Results
29
 
30
+ Translated [Financial news phrase bank](https://arxiv.org/abs/1307.5336) (Malo et al. (2013)):
31
+ - Accuracy: 95.95%
32
+ - Macro F1: 92.70%
33
 
34
 
35
  ## Authors
 
41
  See also:
42
  - scherrmann/GermanFinBERT_SC
43
  - scherrmann/GermanFinBERT_FP
44
+ - scherrmann/GermanFinBERT_FP_QuAD