Object Detection - Yolov7 Training issue on Google Colab - deep-learning

I am training the Yolov7 Object detection model but facing an error in the training, Previously it was working correctly but now facing an error.
I have attached the error screenshot.

Related

Autodesk forge - Error while loading Model

I have an IFC file that I can open with Forge viewer version 7.71 and earlier but not with versions 7.72 to 7.79. The view remains with the loading spinner.
I try with both svf and svf2 options when requesting conversion job.
Here is the error log:
Uncaught TypeError: Cannot read properties of undefined (reading 'length')
at O.postLoadBubbleConsolidation (f42fd425-48e8-4395-9483-8669758069f0:2:164771)
at O.postLoad (f42fd425-48e8-4395-9483-8669758069f0:2:166089)
at Function.<anonymous> (f42fd425-48e8-4395-9483-8669758069f0:2:149625)
at u (f42fd425-48e8-4395-9483-8669758069f0:2:59301)
at XMLHttpRequest.v (f42fd425-48e8-4395-9483-8669758069f0:2:62363)
Regards.
Our engineering team fixed this bug and released a new viewer version, v7.80, yesterday night of my time (UTC+8). Please have a try! Again, we apologize for this problem.
https://forge.autodesk.com/en/docs/viewer/v7/change_history/changelog_v7/#section-1
Note. IFC file is currently processed by the Navisworks technique. See https://forge.autodesk.com/blog/faq-and-tips-ifc-translation-model-derivative-api

Error: Processing is failing in Autodesk-Forge when converting the image to 3D

I am using https://github.com/Autodesk-Forge/recap-walkthrough-photo.to.3d for converting image to 3D, but at the end of the execution its throwing an error when the image is getting processed.
Any Help on this would be highly appreciated.
data: {
Usage: '0.44537496566772',
Resource: '/photoscene/jsjshdfjsdhgjfsdjf/progress',
Photoscene: {
photosceneid: 'jsjshdfjsdhgjfsdjf',
progressmsg: 'ERROR',
progress: '100'
}
}
is the error. Reason for this error is not coming.
Please contact the support via forge (dot) help (at) autodesk (dot) com, providing as much information as possible, e.g.:
how many images you uploaded, and what was their resolution?
how did you upload these files? If you used a custom code or script, could you share it?
when did you run these requests, and what was the Forge Client ID you used for them?
EDIT:
While the response isn't very descriptive, the engineering team confirms that this error is typically caused by insufficient input data. In this particular case only 4 images were uploaded, and the algorithm wasn't able to reconstruct the 3D screen from those.

2D viewables from Revit 2022 models do not work in the Forge Viewer

We have a Forge Viewer (GuiViewer3D, v. 7) implementation in our Nuxt app, using a few extensions (MarkupsCore, Measure, FullScreen) and custom docking panels such as a list of viewables to select from. Ever since we upgraded from Revit version 2021 to 2022, we haven't been able to display the 2D viewables from our svf files in the viewer. When we select any 2D viewable, the viewer freezes in this state:
With a few warnings thrown in the console:
Deprecated API usage: No “GlobalWorkerOptions.workerSrc” specified.
and
Warning: getOperatorList - ignoring errors during "GetOperatorList: page 0" task: "r: Cannot read properties of undefined (reading 'X')"
of which the former lead to this stackoverflow post, to which Augusto Goncalves suggested we tried uploading our revit file in your online viewer at https://viewer.autodesk.com/designviews. This works, and so we tried to create a new Nuxt app with a simple Forge Viewer implementation following the v. 7 example guide. This also works.
What we did next was to re-implement the v. 7 viewer in our production app by following the official documentation once again, thus removing any customization and leaving out everything else than the bare minimum. With this v. 7 viewer, running in a node environment on either node version 10.0 or 14.9, the bug persists, ruling out the possibility of any custom code being the cause.
Upon further debugging of scripts (loaded in via Autodesk) such as pdf.js, CompGeom.min.js, PDFLoader.js, path2d.js etc., we’ve gotten to a point where this message appears, and this is where our current trail ends:
TypeError: Cannot read properties of undefined (reading 'X') at c.strokeClipperContour (https://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/CompGeom/CompGeom.min.js:17:173979) at https://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/CompGeom/CompGeom.min.js:17:177686 at https://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/CompGeom/CompGeom.min.js:17:177811 at Array.map (<anonymous>) at c.fill (https://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/CompGeom/CompGeom.min.js:17:177413) at t.value https://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/CompGeom/CompGeom.min.js:17:16067) at t.value (https://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/CompGeom/CompGeom.min.js:17:9554) at e.beginDrawing (https://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/PDF/PDF.min.js:39:211570) at t.value (https://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/PDF/PDF.min.js:39:185323) at ttps://developer.api.autodesk.com/modelderivative/v2/viewers/7.*/extensions/PDF/PDF.min.js:39:158419
We cannot effectively debug these minified scripts, and are hoping you can provide some insight as to what might be causing this issue.

WebGL context lost error on Autodesk Forge

While loading a product into the Forge viewer using latest version of Chrome, user receives a message that states "WebGL context lost". They had a model loaded into the viewer which was working fine, clicked on another model to view it and that is when they received the error message (see attachment).
Is there anything that can be done to mitigate or resolve this issue?
Click for Picture of Error Message
Here are two possibilities for this issue:
The graphic card driver is out-date, please try to upgrade it: https://knowledge.autodesk.com/support/bim-360/troubleshooting/caas/sfdcarticles/sfdcarticles/Error-WebGL-context-lost-when-viewing-files-in-BIM-360.html
There might be some viewer instances that haven't been destroyed properly. you can type NOP_VIEWERS in your browser console to see if the array length is greater than two. To fix this, you can take advantage of viewer.finish() before loading other models.

Autodesk Forge Design automation - Revit IO: unable to upload application package to design automation app

I am trying to explore design automation for Revit by exploring the code at https://github.com/Autodesk-Forge/design.automation-nodejs-revit.window.family.create.
When I run the application, I get an error message
"Failed to create the family:( "
and the following log -
jstree.min.js:5 [Deprecation] document.registerElement is deprecated
and will be removed in M73, around March 2019. Please use
window.customElements.define instead. See
https://www.chromestatus.com/features/4642138092470272 for more
details. (anonymous) # jstree.min.js:5 ForgeTree.js:348 Uncaught
TypeError: Cannot read property 'className' of undefined
at refreshTypeList (ForgeTree.js:348)
at HTMLSpanElement.spanRemove.onclick (ForgeTree.js:339)
Also while following the instruction in the provided code, most of the links give in README.md for Postman collection (https://github.com/Developer-Autodesk/design-automation-for-revit/blob/master/Docs/AppBundle.md ) was in-accessible to me.
The error you mentioned is on the browser side and it's related to the jsTree, which needs updating due to a browser change. We're fixing it on our samples.
About the overall error, it probably due to a missing setup on the sample. As a suggestion, I would suggest a basic tutorial for Design Automation. Back to the error, you probably need to check the reports.txt to see what happened during the Workitem execution.
We're also fixing a few broken links on the documentation, sorry about that. Are you looking for anything specific?
Update
Design Automation requires code:all scope.