commit files to HF hub
Browse files
README.md
ADDED
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Vocabulary Trimmed [lmqg/mbart-large-cc25-squad-qa](https://huggingface.co/lmqg/mbart-large-cc25-squad-qa): `vocabtrimmer/mbart-large-cc25-squad-qa-trimmed-en`
|
2 |
+
This model is a trimmed version of [lmqg/mbart-large-cc25-squad-qa](https://huggingface.co/lmqg/mbart-large-cc25-squad-qa) 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 |
+
| | lmqg/mbart-large-cc25-squad-qa | vocabtrimmer/mbart-large-cc25-squad-qa-trimmed-en |
|
6 |
+
|:---------------------------|:---------------------------------|:----------------------------------------------------|
|
7 |
+
| parameter_size_full | 610,852,864 | 532,235,264 |
|
8 |
+
| parameter_size_embedding | 512,057,344 | 354,822,144 |
|
9 |
+
| vocab_size | 250,028 | 173,253 |
|
10 |
+
| compression_rate_full | 100.0 | 87.13 |
|
11 |
+
| compression_rate_embedding | 100.0 | 69.29 |
|
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 |
+
| en | vocabtrimmer/mc4_validation | text | en | validation | | 2 |
|