Hugging Face's logo

EMBO
/
bio-lm

Fill-Mask
Transformers
PyTorch
JAX
roberta
language model
Inference Endpoints
Model card Files Community
2
Thomas Lemberger commited on
Commit
5a9225d
·
1 Parent(s): d66184b

card

Browse files
Files changed (1) hide show
  1. README.md +9 -22
README.md CHANGED
@@ -1,18 +1,13 @@
1
  ---
2
  language:
3
- -
4
- -
5
  thumbnail:
6
  tags:
7
- -
8
- -
9
- -
10
  license:
11
  datasets:
12
- -
13
- -
14
  metrics:
15
- -
16
  -
17
  ---
18
 
@@ -28,7 +23,6 @@ This model is a [RoBERTa base model](https://huggingface.co/roberta-base) pre-tr
28
 
29
  The intended use of this model is to be fine-tuned for downstream tasks, token classification in particular.
30
 
31
-
32
  To have a quick check of the model as-is in a fill-mask task:
33
 
34
  ```python
@@ -74,22 +68,15 @@ Training code is available at https://github.com/source-data/soda-roberta
74
  - max_grad_norm: 1.0,
75
  - tensorboard run: lm-MLM-2021-01-27T15-17-43.113766
76
 
77
- End of training eval on validation set:
78
  ```
79
- {'loss': 0.8653350830078125, 'learning_rate': 6.708070119323685e-08, 'epoch': 2.995975157928406}
80
- {'eval_loss': 0.8192330598831177, 'eval_recall': 0.8154601116513597, 'epoch': 2.995975157928406}
81
  ```
82
 
83
-
84
  ## Eval results
85
 
86
  Eval on test set:
87
- {'test_loss': 0.8240728974342346, 'test_recall': 0.814471959728645}
88
-
89
- ### BibTeX entry and citation info
90
-
91
- ```bibtex
92
- @inproceedings{...,
93
- year={2020}
94
- }
95
- ```
 
1
  ---
2
  language:
3
+ - english
 
4
  thumbnail:
5
  tags:
6
+ - language model
 
 
7
  license:
8
  datasets:
9
+ - EMBO/biolang
 
10
  metrics:
 
11
  -
12
  ---
13
 
 
23
 
24
  The intended use of this model is to be fine-tuned for downstream tasks, token classification in particular.
25
 
 
26
  To have a quick check of the model as-is in a fill-mask task:
27
 
28
  ```python
 
68
  - max_grad_norm: 1.0,
69
  - tensorboard run: lm-MLM-2021-01-27T15-17-43.113766
70
 
71
+ End of training:
72
  ```
73
+ trainset: 'loss': 0.8653350830078125
74
+ validation set: 'eval_loss': 0.8192330598831177, 'eval_recall': 0.8154601116513597
75
  ```
76
 
 
77
  ## Eval results
78
 
79
  Eval on test set:
80
+ ```
81
+ recall: 0.814471959728645
82
+ ```