Upload folder using huggingface_hub
Browse files- config.json +1 -2
- tokenizer_config.json +1 -6
config.json
CHANGED
@@ -8,9 +8,8 @@
|
|
8 |
"n_layer": 6,
|
9 |
"block_size": 256,
|
10 |
"dropout": 0.2,
|
11 |
-
"tokenizer_class": "
|
12 |
"tokenizer_file": "spm_model.model",
|
13 |
-
"sentencepiece_model_file": "spm_model.model",
|
14 |
"_name_or_path": "Duino/Darija-LM",
|
15 |
"model_type": "gpt2"
|
16 |
}
|
|
|
8 |
"n_layer": 6,
|
9 |
"block_size": 256,
|
10 |
"dropout": 0.2,
|
11 |
+
"tokenizer_class": "SentencePieceTokenizerFast",
|
12 |
"tokenizer_file": "spm_model.model",
|
|
|
13 |
"_name_or_path": "Duino/Darija-LM",
|
14 |
"model_type": "gpt2"
|
15 |
}
|
tokenizer_config.json
CHANGED
@@ -1,9 +1,4 @@
|
|
1 |
{
|
2 |
-
"tokenizer_class": "
|
3 |
-
"unk_token": "<unk>",
|
4 |
-
"sep_token": "</s>",
|
5 |
-
"pad_token": "<pad>",
|
6 |
-
"cls_token": "[CLS]",
|
7 |
-
"mask_token": "[MASK]",
|
8 |
"model_file": "spm_model.model"
|
9 |
}
|
|
|
1 |
{
|
2 |
+
"tokenizer_class": "SentencePieceTokenizerFast",
|
|
|
|
|
|
|
|
|
|
|
3 |
"model_file": "spm_model.model"
|
4 |
}
|