[question] How to save Meshroom files of Google Colab, so you can run it afterwards (the 12 hour time limit) #767
Comments
|
Just Now I got this error,
7/12 DepthMap
7/12 DepthMap
[17:57:43.150186][info] CUDA-Enabled GPU. [17:57:43.745351][info] Supported CUDA-Enabled GPU detected. WARNING: downgrade status on node "DepthMap_1(26)" from Status.SUBMITTED to Status.NONE
[17:57:43.150186][info] CUDA-Enabled GPU. [17:57:43.745351][info] Supported CUDA-Enabled GPU detected. |
|
I am downloading the Cache File (/tmp/MeshroomCache), and file do nothing until I get a response from someone. Please help this noob! |
Just download the cache folder and the mg file and load the mg file from the meshroom gui. You might need to open the mg file in a text editor and modify the "path": "..." to match your desktop computer path for the images
The image dataset might not be optimal. What are you trying to reconstruct? |
|
Solved! I encountered the same issue as you while running with 451 images on Colab. The status is saved after each successfully computed node and not in between. This means that for example if the computation was halted at DepthMap (21/23), the whole DepthMap node is reset and restarts from the last stable/success node. So, in order to utilize the free Google Colab and save your progress, the best thing to do is,
This will start saving your cache folder alongside your input files in google drive which will not be erased when you restart your runtime.. I am yet to figure out how this actually works but I guess the cache folder is also presumably changed while you change your file paths in iterative mode.. Use Notepad++ or similar.. |
So this is what I did, I used to create a template pipeline of the settings and then re-use it for different projects. Ran that in Meshroom first, just to create a pipeline for my dataset. Then use the output (Render.mg) and run that using these parameters This will be the plan from now on if I rename my Gdrive as mydrive: Put meshroom cache and Render.mg directly to Google Drive, and --forceStatus the way in, so could you explain how you renamed My Drive as mydrive? |
|
@ironman26820 put double quotes around the path before using it or create a symlink |
@ironman26820 I did not rename it actually.. But what you can do is you can either put double quotes as @natowi said or if not you can simply add a '\' before each space like this /content/drive/My\ Drive/Folder\ with\ space/File\ with\ space.mg I used the backslash method in my project.. Since '\' is an escape character, any character next to it will be considered literally. |
Describe the problem
I have taken 177 images of an object (I know its a lot of photos) and tried to make a model about 6 times on my PC. It would always crash on either Feature Extraction or DepthMap. I then saw about Google Colab and so I uploaded all the images on Google Colab and started the meshing. I am now at DepthMap 24/59, and around 6 hours after which Colab will terminate it. Taking average times, I can determine that the DepthMap will take 4-5 hours to complete. I can download all the data (as soon as Depth is completed) on my Computer and then continue it afterwards. I will paste all the MeshroomCache in the same file path (/tmp/MeshroomCache), and the images in the path I have right now. But how will I continue the after DepthMap, so I do not lose any progress.
Also in other issue #251, I saw that I should reduce the maxPoints, and how should I do that on GoogleColab. I am sorry, but I am a NOOB at CLI and python.
Desktop (please complete the following and other pertinent information):
The text was updated successfully, but these errors were encountered: