Upload DeepSeekV2Lite DenseMixer model
Browse filesDeepSeekV2Lite model with DenseMixer architecture
- modeling_deepseek.py +1 -1
modeling_deepseek.py
CHANGED
|
@@ -54,7 +54,7 @@ from transformers.utils import (
|
|
| 54 |
replace_return_docstrings,
|
| 55 |
)
|
| 56 |
from transformers.utils.import_utils import is_torch_fx_available
|
| 57 |
-
from configuration_deepseek import DeepseekV2Config
|
| 58 |
import torch.distributed as dist
|
| 59 |
import numpy as np
|
| 60 |
|
|
|
|
| 54 |
replace_return_docstrings,
|
| 55 |
)
|
| 56 |
from transformers.utils.import_utils import is_torch_fx_available
|
| 57 |
+
from .configuration_deepseek import DeepseekV2Config
|
| 58 |
import torch.distributed as dist
|
| 59 |
import numpy as np
|
| 60 |
|