Upload 6 files
Browse files- special_tokens_map.json +7 -7
- tokenizer_config.json +5 -3
special_tokens_map.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
-
"bos_token": "<|
|
| 3 |
-
"eos_token": "<|
|
| 4 |
-
"pad_token":
|
| 5 |
-
|
|
|
|
| 6 |
"lstrip": false,
|
| 7 |
-
"normalized":
|
| 8 |
"rstrip": false,
|
| 9 |
"single_word": false
|
| 10 |
-
}
|
| 11 |
-
"unk_token": "<|endoftext|>"
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"bos_token": "<|bos|>",
|
| 3 |
+
"eos_token": "<|eos|>",
|
| 4 |
+
"pad_token": "<|pad|>",
|
| 5 |
+
"unk_token": {
|
| 6 |
+
"content": "<|endoftext|>",
|
| 7 |
"lstrip": false,
|
| 8 |
+
"normalized": true,
|
| 9 |
"rstrip": false,
|
| 10 |
"single_word": false
|
| 11 |
+
}
|
|
|
|
| 12 |
}
|
tokenizer_config.json
CHANGED
|
@@ -1,4 +1,6 @@
|
|
| 1 |
{
|
|
|
|
|
|
|
| 2 |
"add_prefix_space": false,
|
| 3 |
"added_tokens_decoder": {
|
| 4 |
"0": {
|
|
@@ -18,12 +20,12 @@
|
|
| 18 |
"special": true
|
| 19 |
}
|
| 20 |
},
|
| 21 |
-
"bos_token": "<|
|
| 22 |
"clean_up_tokenization_spaces": false,
|
| 23 |
-
"eos_token": "<|
|
| 24 |
"extra_special_tokens": {},
|
| 25 |
"model_max_length": 1024,
|
| 26 |
-
"pad_token": "
|
| 27 |
"tokenizer_class": "GPT2Tokenizer",
|
| 28 |
"unk_token": "<|endoftext|>"
|
| 29 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": true,
|
| 4 |
"add_prefix_space": false,
|
| 5 |
"added_tokens_decoder": {
|
| 6 |
"0": {
|
|
|
|
| 20 |
"special": true
|
| 21 |
}
|
| 22 |
},
|
| 23 |
+
"bos_token": "<|bos|>",
|
| 24 |
"clean_up_tokenization_spaces": false,
|
| 25 |
+
"eos_token": "<|eos|>",
|
| 26 |
"extra_special_tokens": {},
|
| 27 |
"model_max_length": 1024,
|
| 28 |
+
"pad_token": "<|pad|>",
|
| 29 |
"tokenizer_class": "GPT2Tokenizer",
|
| 30 |
"unk_token": "<|endoftext|>"
|
| 31 |
}
|