Commit
·
a511285
1
Parent(s):
fa26583
fix: separate source and target vocab
Browse files- target_vocab.json +0 -0
- tokenizer_config.json +13 -1
- vocab.json +0 -0
target_vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -1 +1,13 @@
|
|
| 1 |
-
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"source_lang": "ko",
|
| 3 |
+
"target_lang": "en",
|
| 4 |
+
"unk_token": "<unk>",
|
| 5 |
+
"eos_token": "</s>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"model_max_length": 512,
|
| 8 |
+
"sp_model_kwargs": {},
|
| 9 |
+
"separate_vocabs": true,
|
| 10 |
+
"special_tokens_map_file": null,
|
| 11 |
+
"name_or_path": "marian-models/opusTCv20210807-sepvoc_transformer-big_2022-07-28/ko-en",
|
| 12 |
+
"tokenizer_class": "MarianTokenizer"
|
| 13 |
+
}
|
vocab.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|