Create tokenizer_config.json
Browse files- tokenizer_config.json +23 -0
tokenizer_config.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"sp_model_kwargs": {
|
| 3 |
+
"nbest_size": 6,
|
| 4 |
+
"alpha": 0.1,
|
| 5 |
+
"language": "en"
|
| 6 |
+
},
|
| 7 |
+
"src_lang": "en",
|
| 8 |
+
"tgt_lang": "ff",
|
| 9 |
+
"model_max_length": 512,
|
| 10 |
+
"tokenizer_class": "MarianTokenizer",
|
| 11 |
+
"unk_token": "<unk>",
|
| 12 |
+
"bos_token": "<s>",
|
| 13 |
+
"eos_token": "</s>",
|
| 14 |
+
"pad_token": "<pad>",
|
| 15 |
+
"mask_token": "<mask>",
|
| 16 |
+
"additional_special_tokens": [
|
| 17 |
+
"<pad>",
|
| 18 |
+
"<unk>",
|
| 19 |
+
"<s>",
|
| 20 |
+
"</s>",
|
| 21 |
+
"<mask>"
|
| 22 |
+
]
|
| 23 |
+
}
|