Software Engineering Courses (SECourses)•3y ago

What’s new?

Vlad•12/20/23, 12:23 AM

Hello,
Does anyone know how to access tensorboard from a pod on runpod.io as it doesn't generate a public url?

Tried this method here https://blog.yyliu.net/remote-tensorboard/ but I can't make it work, receiving this on my local:
channel 4: open failed: unknown channel type: unsupported channel type

(the connection ssh between my machine and the pod works)

Yaoyao's Blog

How to use the tensorboard on a remote server

Tensorboard is a very convenient tools for tensorflow, whihc allows us to view the summaries of the running experiments in the browser. However, for the people who study the DNNs, it is very common to

VVlad Hello, Does anyone know how to access tensorboard from a pod on runpod.io as it ...

Vlad•12/20/23, 1:42 AM

[SOLVED in case someone else needs this] First expose a port (e.g. 6001) from pod's settings and then add --bind all to the command: tensorboard --logdir "/workspace/something" --port 6001 --bind_all

AIGambino•12/20/23, 3:01 AM

Did you ever make that tutorial on what constitutes a good dataset

eyepatch_2022•12/20/23, 10:18 AM

Hey can anyone guide me on this error on Kaggle?

eyepatch_2022•12/20/23, 10:19 AM

eyepatch_2022•12/20/23, 10:25 AM

while installing Pyngrok it is giving this error:

ERROR: Could not install packages due to an OSError: [Errno 2] No such file or directory: '/opt/conda/lib/python3.10/site-packages/PyYAML-6.0.1.dist-info/METADATA'

Arshia•12/20/23, 10:44 AM

hey im new to this, how can i enable the control net like in webui, but in code Enabled on the .safetensor i want not default using replicate or any Gpu provider?

like using ControlNet on the checkpoint(blalala.safetensors) i want not the default checkpoint, using the the replicate? exactly how webui allows you to decide the checkpoint and then enabling control net on selected model. whats the process of that should i change the controlnet model change the base checkpoint and cog it to replicate or is there any better way to do that?

i have tried running the model i want on replicate by giving the controlnet seg map got some familier result but still far from the combination in webui, and in that way don't know how should change the settings like

starting control steps: , ending control steps:

Imogen Fox•12/20/23, 10:53 AM

Hello - I am trying to create some animations with animatedDiff. In the stable diffusion checkpoint I am using a model i used to train safetensor files with on my own style (sdxl 1.0) but it does not want to output any sort of gif animation with this model - can anyone tell me how i can do this? OR how can i train an animated SD model (like the ones you find on CivitAI) to use for text to MP4 output with my own safetensor files etc?

Deleted User•12/20/23, 1:09 PM

vv•12/20/23, 2:04 PM

guys does training my photos as 1024 * 1024 model better than 512 * 512 model;
also I want to use controlnet extensively and sdxl doesnt support openpose; should I train using sd 1.5 model?

vv•12/20/23, 2:04 PM

@Dr. Furkan Gözükara

vv•12/20/23, 2:05 PM

will really apprciate the help; am trying to get something where my model is helping other people; but my output is very very underwhelming

vv•12/20/23, 2:07 PM

https://www.youtube.com/watch?v=EEV8RPohsbw&ab_channel=SECourses
I used the above video and config 24gb_notextencoder config and I have used a variety of prompt but none is working..

vv•12/20/23, 2:08 PM

i got 4090; will really appreciate the help

Motoko_Shi•12/20/23, 2:20 PM

Struggling with Rerender A Video.

So far I've been unable to get the implementation running after trying multiple methods, including Software Engineering Course's one-click installer. Whether following the repo instructions or the installer, running WebUI.py or Run.bat always results in:

Traceback (most recent call last):
File "C:\RRAV\Rerender_A_Video\webUI.py", line 5, in
import cv2
ModuleNotFoundError: No module named 'cv2'
Press any key to continue . . .

The culprit seems to be this error. (see attached image.png). I'm not a programmer and I'm afraid I just don't know what I'm doing wrong. I've tried both with the venv active and without it, same error.

Motoko_Shi•12/20/23, 2:21 PM

I've tried all the suggestions on the Patreon discussion page as well.

Motoko_Shi•12/20/23, 2:21 PM

reinstalled cuda, in Visual Studio have win 10 and 11 packages in place

StarBuffer•12/20/23, 2:48 PM

Hello, I am trying inpainting and struggles in the result.

StarBuffer•12/20/23, 2:49 PM

StarBuffer•12/20/23, 2:49 PM

The red area is the original image

StarBuffer•12/20/23, 2:51 PM

And the result is really bad.

StarBuffer•12/20/23, 2:51 PM

How can I avoid this fade effect?

JJMPA Javier Milei president of Argentina 🫡💎

Xylber•12/20/23, 3:52 PM

Nice job. Hope he can destroy the Central Bank, every country should be doing the same.

DDeleted User Click to see attachment

Chip•12/20/23, 7:28 PM

where can I find the exact workflow for this? It says workflow included but I did not find it

JJMPA Javier Milei president of Argentina 🫡💎

Chip•12/20/23, 7:28 PM

these are really nice. How did you make these? What was your process? Would love to create these images with photos of friends of mine

Rodlyman•12/20/23, 7:32 PM

Hello! Suddenly I cannot run fooocus anymore. I'm getting this error: "warnings.warn("loaded more than 1 DLL from .libs:""

Rodlyman•12/20/23, 7:33 PM

Any thoughts on what might be happening?

I'm going to re-download and reinstall

Resolved the issue**

SStarBuffer Hello, I am trying inpainting and struggles in the result.

Rodlyman•12/20/23, 7:34 PM

Hi - do you mean Outpainting? As in you're trying to extend an image?

୧୧ʕ•̀ᴥ•́ʔ୨ I hope Kaggle doesn't get so popular so that they reduce the allowances

Furkan Gözükara SECourses•12/20/23, 9:18 PM