duzx16 commited on
Commit ·
0d8b08d
1
Parent(s): 0101afd
Revert "Update quantization"
Browse filesThis reverts commit 0101afdfa7f279e082623ee57d7ad6f84e1e82dc.
- quantization.py +1 -5
quantization.py
CHANGED
|
@@ -7,11 +7,7 @@ import base64
|
|
| 7 |
import ctypes
|
| 8 |
|
| 9 |
from typing import List
|
| 10 |
-
|
| 11 |
-
try:
|
| 12 |
-
from cpm_kernels.kernels.base import LazyKernelCModule, KernelFunction, round_up
|
| 13 |
-
except:
|
| 14 |
-
pass
|
| 15 |
|
| 16 |
|
| 17 |
class W8A16Linear(torch.autograd.Function):
|
|
|
|
| 7 |
import ctypes
|
| 8 |
|
| 9 |
from typing import List
|
| 10 |
+
from cpm_kernels.kernels.base import LazyKernelCModule, KernelFunction, round_up
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
|
| 12 |
|
| 13 |
class W8A16Linear(torch.autograd.Function):
|