Upload 8 files
Browse files- merges.txt +0 -0
- tokenizer_config.json +2 -1
- vocab.json +0 -0
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -1,11 +1,12 @@
|
|
| 1 |
{
|
| 2 |
-
"tokenizer_class": "
|
| 3 |
"tokenizer_file": "tokenizer.json",
|
| 4 |
"model_max_length": 1024,
|
| 5 |
"add_prefix_space": true,
|
| 6 |
"clean_up_tokenization_spaces": false,
|
| 7 |
"bos_token": "<|begin_of_text|>",
|
| 8 |
"eos_token": "<|eot_id|>",
|
|
|
|
| 9 |
"pad_token": "<|pad|>",
|
| 10 |
"unk_token": "<|unk|>",
|
| 11 |
"additional_special_tokens": [
|
|
|
|
| 1 |
{
|
| 2 |
+
"tokenizer_class": "GPT2TokenizerFast",
|
| 3 |
"tokenizer_file": "tokenizer.json",
|
| 4 |
"model_max_length": 1024,
|
| 5 |
"add_prefix_space": true,
|
| 6 |
"clean_up_tokenization_spaces": false,
|
| 7 |
"bos_token": "<|begin_of_text|>",
|
| 8 |
"eos_token": "<|eot_id|>",
|
| 9 |
+
"eot_token": "<|eot_id|>",
|
| 10 |
"pad_token": "<|pad|>",
|
| 11 |
"unk_token": "<|unk|>",
|
| 12 |
"additional_special_tokens": [
|
vocab.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|