- tokenization_xgen.py +6 -2
tokenization_xgen.py
CHANGED
|
@@ -6,11 +6,15 @@
|
|
| 6 |
|
| 7 |
from typing import List, Optional
|
| 8 |
|
| 9 |
-
import tiktoken
|
| 10 |
-
|
| 11 |
from transformers.tokenization_utils import AddedToken, PreTrainedTokenizer
|
| 12 |
from transformers.utils import logging
|
| 13 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
logger = logging.get_logger(__name__)
|
| 15 |
|
| 16 |
MAX_MODEL_INPUT_SIZES = {
|
|
|
|
| 6 |
|
| 7 |
from typing import List, Optional
|
| 8 |
|
|
|
|
|
|
|
| 9 |
from transformers.tokenization_utils import AddedToken, PreTrainedTokenizer
|
| 10 |
from transformers.utils import logging
|
| 11 |
|
| 12 |
+
try:
|
| 13 |
+
import tiktoken
|
| 14 |
+
except ModuleNotFoundError as e:
|
| 15 |
+
raise ModuleNotFoundError("XGen requires the installation of tiktoken. Please install it via `pip install tiktoken`.") from e
|
| 16 |
+
|
| 17 |
+
|
| 18 |
logger = logging.get_logger(__name__)
|
| 19 |
|
| 20 |
MAX_MODEL_INPUT_SIZES = {
|