ondfa commited on
Commit
e1fa83b
1 Parent(s): 6714bd8

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +11 -3
README.md CHANGED
@@ -1,3 +1,11 @@
 
 
 
 
 
 
 
 
1
  # CZERT
2
  This repository keeps trained Czert-B model for the paper [Czert – Czech BERT-like Model for Language Representation
3
  ](https://arxiv.org/abs/2103.13031)
@@ -39,14 +47,14 @@ We evaluate our model on two sentence level tasks:
39
 
40
 
41
  <!-- tokenizer = BertTokenizerFast.from_pretrained(CZERT_MODEL_PATH, strip_accents=False)
42
- model = TFAlbertForSequenceClassification.from_pretrained(CZERT_MODEL_PATH, num_labels=1)
43
 
44
  or
45
 
46
  self.tokenizer = BertTokenizerFast.from_pretrained(CZERT_MODEL_PATH, strip_accents=False)
47
  self.model_encoder = AutoModelForSequenceClassification.from_pretrained(CZERT_MODEL_PATH, from_tf=True)
48
  -->
49
-
50
  ### Document Level Tasks
51
  We evaluate our model on one document level task
52
  * Multi-label Document Classification.
@@ -102,7 +110,7 @@ Comparison of F1 score achieved using pre-trained CZERT-A, CZERT-B, mBERT, Pavlo
102
 
103
  | | mBERT | Pavlov | Albert-random | Czert-A | Czert-B | dep-based | gold-dep |
104
  |:------:|:----------:|:----------:|:-------------:|:----------:|:----------:|:---------:|:--------:|
105
- | span | 78.547 ± 0.110 | 79.333 ± 0.080 | 51.365 ± 0.423 | 72.254 ± 0.172 | **81.861 ± 0.102** | \- | \- |
106
  | syntax | 90.226 ± 0.224 | 90.492 ± 0.040 | 80.747 ± 0.131 | 80.319 ± 0.054 | **91.462 ± 0.062** | 85.19 | 89.52 |
107
 
108
  SRL results – dep columns are evaluate with labelled F1 from CoNLL 2009 evaluation script, other columns are evaluated with span F1 score same as it was used for NER evaluation. For more information see [the paper](https://arxiv.org/abs/2103.13031).
 
1
+ ---
2
+ tags:
3
+ - cs
4
+ - bert
5
+ - Transformers
6
+ - Tensorflow
7
+ ---
8
+
9
  # CZERT
10
  This repository keeps trained Czert-B model for the paper [Czert – Czech BERT-like Model for Language Representation
11
  ](https://arxiv.org/abs/2103.13031)
 
47
 
48
 
49
  <!-- tokenizer = BertTokenizerFast.from_pretrained(CZERT_MODEL_PATH, strip_accents=False)
50
+ \tmodel = TFAlbertForSequenceClassification.from_pretrained(CZERT_MODEL_PATH, num_labels=1)
51
 
52
  or
53
 
54
  self.tokenizer = BertTokenizerFast.from_pretrained(CZERT_MODEL_PATH, strip_accents=False)
55
  self.model_encoder = AutoModelForSequenceClassification.from_pretrained(CZERT_MODEL_PATH, from_tf=True)
56
  -->
57
+ \t
58
  ### Document Level Tasks
59
  We evaluate our model on one document level task
60
  * Multi-label Document Classification.
 
110
 
111
  | | mBERT | Pavlov | Albert-random | Czert-A | Czert-B | dep-based | gold-dep |
112
  |:------:|:----------:|:----------:|:-------------:|:----------:|:----------:|:---------:|:--------:|
113
+ | span | 78.547 ± 0.110 | 79.333 ± 0.080 | 51.365 ± 0.423 | 72.254 ± 0.172 | **81.861 ± 0.102** | \\- | \\- |
114
  | syntax | 90.226 ± 0.224 | 90.492 ± 0.040 | 80.747 ± 0.131 | 80.319 ± 0.054 | **91.462 ± 0.062** | 85.19 | 89.52 |
115
 
116
  SRL results – dep columns are evaluate with labelled F1 from CoNLL 2009 evaluation script, other columns are evaluated with span F1 score same as it was used for NER evaluation. For more information see [the paper](https://arxiv.org/abs/2103.13031).