seonjeongh
commited on
Commit
•
9cecef8
1
Parent(s):
e8fa983
commit files to HF hub
Browse files
README.md
CHANGED
@@ -1,3 +1,18 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Vocabulary Trimmed [google/mt5-base](https://huggingface.co/google/mt5-base): `seonjeongh/mt5-base-ko`
|
2 |
+
This model is a trimmed version of [google/mt5-base](https://huggingface.co/google/mt5-base) by [`vocabtrimmer`](https://github.com/asahi417/lm-vocab-trimmer), a tool for trimming vocabulary of language models to compress the model size.
|
3 |
+
Following table shows a summary of the trimming process.
|
4 |
+
|
5 |
+
| | google/mt5-base | seonjeongh/mt5-base-ko |
|
6 |
+
|:---------------------------|:------------------|:-------------------------|
|
7 |
+
| parameter_size_full | 582,401,280 | 310,904,064 |
|
8 |
+
| parameter_size_embedding | 384,172,032 | 112,674,816 |
|
9 |
+
| vocab_size | 250,112 | 73,356 |
|
10 |
+
| compression_rate_full | 100.0 | 53.38 |
|
11 |
+
| compression_rate_embedding | 100.0 | 29.33 |
|
12 |
+
|
13 |
+
|
14 |
+
Following table shows the parameter used to trim vocabulary.
|
15 |
+
|
16 |
+
| language | dataset | dataset_column | dataset_name | dataset_split | target_vocab_size | min_frequency |
|
17 |
+
|:-----------|:----------------------------|:-----------------|:---------------|:----------------|:--------------------|----------------:|
|
18 |
+
| ko | vocabtrimmer/mc4_validation | text | ko | validation | | 2 |
|