You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2020/01/27 18:09:53 UTC
[GitHub] [incubator-tvm] comaniac commented on a change in pull request
#4779: [AUTOTVM] Fix a bug in generating the search space
comaniac commented on a change in pull request #4779: [AUTOTVM] Fix a bug in generating the search space
URL: https://github.com/apache/incubator-tvm/pull/4779#discussion_r371397521
##########
File path: python/tvm/autotvm/task/space.py
##########
@@ -226,7 +226,13 @@ def __init__(self, axes, policy, **kwargs):
def _generate_space(self, now, tmp_stack, enforce_no_tail=False):
"""Generate space by DFS"""
if now == self.num_output - 1:
- prod = np.prod(tmp_stack, dtype=np.int64)
+ prod = 1
Review comment:
It seems to me that manually implementing a classic array production is not necessary in any case. Since limited types are only enforced in numpy and Python's types are unlimited, it would be more concise to use Python builtins to calculate the product:
```python
import functools
import operator
prod = functools.reduce(operator.mul, tmp_stack, 1)
```
Note that the length of `tmp_stack` is always small (currently 4 at most, and I don't think it would be longer than 10), so this won't hurt the performance.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
With regards,
Apache Git Services