-
Notifications
You must be signed in to change notification settings - Fork 4.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Rasa1.7.1 segmentation fault(core dumped) #5228
Comments
@gitpranjal is this thread helpful at all? #5196 |
Got it. Yeah, the segmentation fault seems to only happen on Ubuntu based on the issues i've seen. The rest seems like just a general memory error, how much memory does your machine have? |
I have the same problem when using the Docker images. With this image, training works fine: With these, core training fails: (Segmentation fault right at the beginning of training core). Tried boosting the RAM of Docker to 6GB with a 4GB swap but to no avail. Also, the failure happens at the exact same point regardless of the amount of RAM. |
Thanks @benos That made it really easy to reproduce for me. I'll have a look |
Using
Getting this in 1.7.2-full:
|
|
I suggest pinning this to <=0.15.4. Research is removing the |
i'm fixing the segmentation fault with #5248 . @gitpranjal How much memory has your machine? |
I have got 16 GB ram, Ubuntu 18.04 OS
…On Mon, 17 Feb, 2020, 7:41 PM Tobias Wochinger, ***@***.***> wrote:
i'm fixing the segmentation fault with #5248
<#5248> .
@gitpranjal <https://github.com/gitpranjal> How much memory has your
machine?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#5228?email_source=notifications&email_token=AGPDPDXXLHFZLXLKTD3H2QTRDKLJZA5CNFSM4KTWIQR2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEL6RPCY#issuecomment-587011979>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AGPDPDSO7L5UEF5WSM533VDRDKLJZANCNFSM4KTWIQRQ>
.
|
And how much memory is consumed? Can you please monitor the memory consumption during the training with a tool like |
[image: Screenshot from 2020-02-18 11-31-26.png]
[image: Screenshot from 2020-02-18 11-31-58.png]
[image: Screenshot from 2020-02-18 11-32-00.png]
[image: Screenshot from 2020-02-18 11-31-47.png]
[image: Screenshot from 2020-02-18 11-29-57.png]
In my config.yml file, max history is set to 251. It is getting killed at
even 181
…On Mon, Feb 17, 2020 at 7:44 PM Tobias Wochinger ***@***.***> wrote:
And how much memory is consumed? Can you please monitor the memory
consumption during the training with a tool like htop or similar
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#5228?email_source=notifications&email_token=AGPDPDXUKTCO6Y54LGFG5BDRDKLT7A5CNFSM4KTWIQR2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEL6RYDA#issuecomment-587013132>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AGPDPDWN5MTP3QAQYW4I723RDKLT7ANCNFSM4KTWIQRQ>
.
|
This is a very very very large |
Yes, the stories are very long with a lot of slot filling and buttons being
clicked giving multiple cases. On reducing the max_history below 180,
stories are not followed properly.
…On Wed, Feb 19, 2020 at 2:26 PM Tobias Wochinger ***@***.***> wrote:
In my config.yml file, max history is set to 251. It is getting killed at
This is a very very very large max_history. Any specific reason to set it
that hight? Can you maybe give it a try with something in the range 3-10?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#5228?email_source=notifications&email_token=AGPDPDW3IHUQKQLCCMCNLCDRDTX2NA5CNFSM4KTWIQR2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEMG45SQ#issuecomment-588107466>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AGPDPDVNZOXOEYIJ4DHE2BTRDTX2NANCNFSM4KTWIQRQ>
.
|
I don't know your use case / story structure, but I would argue that this seems like an issue with the stories. I talked to a couple of colleagues and nobody has ever seen a |
Rasa version:
Rasa SDK version (if used & relevant):
Rasa X version (if used & relevant):
1.7.1
Python version:
3.6.8
Operating system (windows, osx, ...):
Ubuntu18.04LTS
Issue:
Segmentation fault(core dumped) upon attempting to training the rasa core model
Error (including full traceback):
Command or request that led to error:
Content of configuration file (config.yml) (if relevant):
Content of domain file (domain.yml) (if relevant):
The text was updated successfully, but these errors were encountered: