blob: 9bda033b30bcd51a02835523cf429bb6c3f3ec53 (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
|
graphic: 0.872
device: 0.854
instruction: 0.822
semantic: 0.649
network: 0.604
vnc: 0.387
socket: 0.379
mistranslation: 0.277
boot: 0.260
other: 0.249
assembly: 0.159
KVM: 0.018
import tensorflow causes qemu: uncaught target signal 6 (Aborted) - core dumped
Description of problem:
Crashes when importing tensorflow in Docker container under --platorm linux/amd64 on M1 Mac
```
2021-09-06 13:35:24.435613: F tensorflow/core/lib/monitoring/sampler.cc:42] Check failed: bucket_limits_[i] > bucket_limits_[i - 1] (0 vs. 10)
qemu: uncaught target signal 6 (Aborted) - core dumped
```
Steps to reproduce:
See https://gitlab.com/ryan-feather/docker-tensorflow-qemu-bug/ for Dockerfile and description of steps repeating here.
1. Using the dockerfile
```
FROM python:3.9-buster
RUN pip install tensorflow==2.6.0
```
2. `docker buildx build --iidfile build.id --platform linux/amd64 . --progress=plain`
3. ``` docker run --platform linux/amd64 `cat build.id` python -c "import tensorflow"```
Additional information:
See
https://github.com/docker/for-mac/issues/5342 where the Docker team suggests this is a qemu bug. I couldn't find where anyone had opened one of these here, so hopefully this isn't a duplicate.
|