aapot commited on
Commit
21208d6
1 Parent(s): 0f719af

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -2
README.md CHANGED
@@ -98,16 +98,17 @@ Training code was from the Google's Jax/Flax based [t5x framework](https://githu
98
 
99
  Evaluation was done by fine-tuning the model on a downstream text classification task with two different labeled Finnish datasets: [Yle News](https://github.com/spyysalo/yle-corpus) and [Eduskunta](https://github.com/aajanki/eduskunta-vkk). Classification fine-tuning was done with a sequence length of 128 tokens.
100
 
101
- When fine-tuned on those datasets, this model (the third row of the table) achieves the following accuracy results compared to our other T5 models and their parameter counts:
102
 
103
  | | Model parameters | Yle News accuracy | Eduskunta accuracy |
104
  |-------------------------------------------------------|------------------|---------------------|----------------------|
105
  |Finnish-NLP/t5-tiny-nl6-finnish | 31 million |92.80 |69.07 |
106
  |Finnish-NLP/t5-mini-nl8-finnish | 72 million |93.89 |71.43 |
 
107
  |Finnish-NLP/t5-small-nl24-finnish | 260 million |**94.68** |74.90 |
108
  |Finnish-NLP/byt5-base-finnish | 582 million |92.33 |73.13 |
109
  |Finnish-NLP/t5-base-nl36-finnish | 814 million |94.40 |**75.97** |
110
- |Finnish-NLP/t5-large-nl36-finnish | 1425 million |TBA |TBA |
111
 
112
 
113
  Fine-tuning Google's multilingual mT5 models on the same datasets we can clearly see that our monolingual Finnish T5 models achieve much better results on Finnish text classification:
 
98
 
99
  Evaluation was done by fine-tuning the model on a downstream text classification task with two different labeled Finnish datasets: [Yle News](https://github.com/spyysalo/yle-corpus) and [Eduskunta](https://github.com/aajanki/eduskunta-vkk). Classification fine-tuning was done with a sequence length of 128 tokens.
100
 
101
+ When fine-tuned on those datasets, this model (the fourth row of the table) achieves the following accuracy results compared to our other T5 models and their parameter counts:
102
 
103
  | | Model parameters | Yle News accuracy | Eduskunta accuracy |
104
  |-------------------------------------------------------|------------------|---------------------|----------------------|
105
  |Finnish-NLP/t5-tiny-nl6-finnish | 31 million |92.80 |69.07 |
106
  |Finnish-NLP/t5-mini-nl8-finnish | 72 million |93.89 |71.43 |
107
+ |Finnish-NLP/t5-small-nl16-finnish | 184 million |94.46 |74.00 |
108
  |Finnish-NLP/t5-small-nl24-finnish | 260 million |**94.68** |74.90 |
109
  |Finnish-NLP/byt5-base-finnish | 582 million |92.33 |73.13 |
110
  |Finnish-NLP/t5-base-nl36-finnish | 814 million |94.40 |**75.97** |
111
+ |Finnish-NLP/t5-large-nl36-finnish | 1425 million |94.17 |73.50 |
112
 
113
 
114
  Fine-tuning Google's multilingual mT5 models on the same datasets we can clearly see that our monolingual Finnish T5 models achieve much better results on Finnish text classification: