Rjacobucci / regsem

Regularized Structural Equation Modeling
14 stars 8 forks source link

error thrown in det_range_par #18

Open IsobelRidler opened 2 years ago

IsobelRidler commented 2 years ago

Hi Ross,

I've been running stability selection on a cluster, and I get this error from the det_range_par function above a certain number of bootstraps: "(list) object cannot be coerced to type 'double'... Execution halted"

Is this a known issue? At first I thought it was my script, but it runs without issue for other values of the argument (e.g. times = 30), and I've had this on a couple of scripts, when very similar alternative models/scripts are fine running for times = 50. I'm not so concerned when I can still run for 30, but with one example, I can't even get it to work for 25, and since you only test for 50 in your paper and mention 30 in the recent regsem tutorial paper, I'd ideally like to be able to use 30 at least.

Many thanks in advance for your help with this.

NB: I'm using 5 cores on one node, with 16G per core

Rjacobucci commented 2 years ago

Hi Isobel,

We just checked on example code and it was working fine. Could you pass along the script you are using for it? Ideally something that we could reproduce. Feel free to email me directly and include Xiaobei in the message.

Best, Ross

On Sun, Jun 12, 2022 at 4:57 PM IsobelRidler @.***> wrote:

Hi Ross,

I've been running stability selection on a cluster, and I get this error from the det_range_par function above a certain number of bootstraps: "(list) object cannot be coerced to type 'double'... Execution halted"

Is this a known issue? At first I thought it was my script, but it runs without issue for other values of the argument (e.g. times = 30), and I've had this on a couple of scripts, when very similar alternative models/scripts are fine running for times = 50. I'm not so concerned when I can still run for 30, but with one example, I can't even get it to work for 25, and since you only test for 50 in your paper and mention 30 in the recent regsem tutorial paper, I'd ideally like to be able to use 30 at least.

Many thanks in advance for your help with this.

NB: I'm using 5 cores on one node, with 16G per core

— Reply to this email directly, view it on GitHub https://github.com/Rjacobucci/regsem/issues/18, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABRVLHLXY53CYUJ6KPSBIMDVOZFJZANCNFSM5YSJPRUA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

IsobelRidler commented 2 years ago

Hi Ross, it works for some and not others for me.. I have edited your function just to make it reproducible and include the std.lv=TRUE option, so I will email you and Xiaobei directly as you suggest. Many thanks, Isobel