leezu commented on issue #18314:
URL:
https://github.com/apache/incubator-mxnet/issues/18314#issuecomment-628969701
Let's close this issue as it's specific to the use of gc.DEBUG_SAVEALL
This is an automated message from the
leezu commented on issue #18314:
URL:
https://github.com/apache/incubator-mxnet/issues/18314#issuecomment-628833322
That's because I set gc.set_debug(gc.DEBUG_SAVEALL). I want to preserve all
objects found by gc, because I need to figure out where the reference cycle
that prevented them f
leezu commented on issue #18314:
URL:
https://github.com/apache/incubator-mxnet/issues/18314#issuecomment-628829305
I can confirm that __del__ is called on the object, but it's still kept
around in Python afterwards.
In practice that may only make problems when debugging memory leaks (a
leezu commented on issue #18314:
URL:
https://github.com/apache/incubator-mxnet/issues/18314#issuecomment-628821443
The mutex in above core dump is actually non existant. The problem here is
that the ffi call to MXNDArraySyncCopyToCPU passes a handle whose Chunk was
cleaned up. So the seg
leezu commented on issue #18314:
URL:
https://github.com/apache/incubator-mxnet/issues/18314#issuecomment-628407634
To get the backtrace in gdb, use the example without the for loop in the end
```
import gc
gc.set_debug(gc.DEBUG_SAVEALL)
import mxnet as mx
net = mx.gluon.
leezu commented on issue #18314:
URL:
https://github.com/apache/incubator-mxnet/issues/18314#issuecomment-628398197
On 1.6 it also triggers segfault with NaiveEngine. But not on master.
This is an automated message from the
