Add tokenizer_config.json
Browse files- tokenizer_config.json +34 -0
tokenizer_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|endoftext|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|endoftext|>",
|
| 10 |
+
"<|im_start|>",
|
| 11 |
+
"<|im_end|>",
|
| 12 |
+
"<repo_name>",
|
| 13 |
+
"<reponame>",
|
| 14 |
+
"<file_sep>",
|
| 15 |
+
"<filename>",
|
| 16 |
+
"<gh_stars>",
|
| 17 |
+
"<issue_start>",
|
| 18 |
+
"<issue_comment>",
|
| 19 |
+
"<issue_closed>",
|
| 20 |
+
"<jupyter_start>",
|
| 21 |
+
"<jupyter_text>",
|
| 22 |
+
"<jupyter_code>",
|
| 23 |
+
"<jupyter_output>",
|
| 24 |
+
"<jupyter_script>",
|
| 25 |
+
"<empty_output>"
|
| 26 |
+
],
|
| 27 |
+
"is_local": false,
|
| 28 |
+
"local_files_only": false,
|
| 29 |
+
"model_max_length": 8192,
|
| 30 |
+
"pad_token": "<|endoftext|>",
|
| 31 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 32 |
+
"unk_token": "<|endoftext|>",
|
| 33 |
+
"vocab_size": 49152
|
| 34 |
+
}
|