Kernel crashes and Segmentation fault (core dumped)

[dadaaichifan]

When I run antomatic_mask_generator_example.ipynb, the kernel crashed When I run the command line in the terminal , will appear Segmentation fault (core dumped)

I think these two problems are the same kind of problem, this problem has been bothering me all day, could you please point out the problem, thank you.

[timonpalm]

Facing the same issue :/

[timonpalm]

In my case, it works sometimes. I figured out that it happens when I run specifically this code

sam = sam_model_registry[model_type](checkpoint=sam_checkpoint)

mask_generator = SamAutomaticMaskGenerator(model=sam)

masks = mask_generator.generate(image)

or

predictor.set_image(image)

The error even persists when I run it with smaller images, it runs a bit more reliable though. Restarting the kernel does not help, also restarting vscode is not a remedy. Furthermore, creating a new environment with a fresh installation of SAM either via git or pip also did not solve the problem

The only that that helped so far, was to reboot the system. Then, I was able to run the aforementioned cells a few times before it crashes again.

I suppose it has to be some kind of memory error/ garbage collection, that the memory of these computation are not being emptied after run and it thus overflows.

[dadaaichifan]

就我而言，它有时有效。我发现，当我专门运行此代码时，会发生这种情况

sam = sam_model_registry[model_type](checkpoint=sam_checkpoint)

mask_generator = SamAutomaticMaskGenerator(model=sam)

masks = mask_generator.generate(image)

或

predictor.set_image(image)

当我使用较小的图像运行它时，错误仍然存在，不过它运行得更可靠一些。重新启动内核没有帮助，重新启动 vscode 也不是一种补救措施。此外，通过 git 或 pip 使用全新安装的 SAM 创建新环境也无法解决问题

到目前为止，唯一有帮助的是重新启动系统。然后，我能够在它再次崩溃之前运行上述单元格几次。

我想这一定是某种内存错误/垃圾回收，这些计算的内存在运行后没有被清空，因此溢出

Do you have a more effective solution now