Upload tokenizer
Browse files- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -1
special_tokens_map.json
CHANGED
|
@@ -6,6 +6,7 @@
|
|
| 6 |
"rstrip": false,
|
| 7 |
"single_word": false
|
| 8 |
},
|
|
|
|
| 9 |
"pad_token": {
|
| 10 |
"content": "<PAD>",
|
| 11 |
"lstrip": false,
|
|
|
|
| 6 |
"rstrip": false,
|
| 7 |
"single_word": false
|
| 8 |
},
|
| 9 |
+
"eos_token": "<|endoftext|>",
|
| 10 |
"pad_token": {
|
| 11 |
"content": "<PAD>",
|
| 12 |
"lstrip": false,
|
tokenizer_config.json
CHANGED
|
@@ -324,7 +324,7 @@
|
|
| 324 |
},
|
| 325 |
"bos_token": "<|endoftext|>",
|
| 326 |
"clean_up_tokenization_spaces": true,
|
| 327 |
-
"eos_token":
|
| 328 |
"model_max_length": 2048,
|
| 329 |
"pad_token": "<PAD>",
|
| 330 |
"tokenizer_class": "CodeGenTokenizer",
|
|
|
|
| 324 |
},
|
| 325 |
"bos_token": "<|endoftext|>",
|
| 326 |
"clean_up_tokenization_spaces": true,
|
| 327 |
+
"eos_token": "<|endoftext|>",
|
| 328 |
"model_max_length": 2048,
|
| 329 |
"pad_token": "<PAD>",
|
| 330 |
"tokenizer_class": "CodeGenTokenizer",
|