TheLastBen
RunPod not showing Jupyter Lab option
The runpod template isn't providing the option connect to jupyter lab anymore.
Runpd decided to discontinue support for fast stable diffusion notebooks, it's out of my hands
@TheLastBen oh I see. That's unfortunate.
Is colab the only option remaining then? Since paperspace seems to have issues with this as well.
Traceback (most recent call last):
File /usr/local/lib/python3.10/dist-packages/IPython/core/interactiveshell.py:3508 in run_code exec(code_obj, self.user_global_ns, self.user_ns)
Cell In[1], line 8 with open('/dev/null', 'w') as devnull:import requests, os, time, importlib;open('/workspace/mainrunpodA1111.py', 'wb').write(requests.get('https://huggingface.co/datasets/TheLastBen/RNPD/raw/main/Scripts/mainrunpodA1111.py').content);os.chdir('/workspace');time.sleep(2);import mainrunpodA1111;importlib.reload(mainrunpodA1111);from mainrunpodA1111 import *;Deps(force_reinstall)
File /workspace/mainrunpodA1111.py:1 Invalid username or password. ^ SyntaxError: invalid syntax
I pay Runpod only to use this template. If they discontinue supprt for it, I will contact them try to work out why.
I pay Runpod only to use this template. If they discontinue supprt for it, I will contact them try to work out why.
same here. I just finished fully migrating from paperspace and settling in, depending on runpod+fast-stable-diffusion for the foreseeable future. I just wrote to runpod and let them know all my business with them depends on support for this template. it sucks cause last week just rented a 4000 pod for a month
@TheLastBen oh I see. That's unfortunate.
Is colab the only option remaining then? Since paperspace seems to have issues with this as well.
Paperspace is up and running
@TheLastBen oh I see. That's unfortunate. Is colab the only option remaining then? Since paperspace seems to have issues with this as well.
Paperspace is up and running
say whaaa. ok now theyre just fuckin with me lol. i literally wiped my disks and canceled my account last week
@TheLastBen
Is there any chance that Runpod will start it again ? Should I ask them for a refund and start with Paperspace ?
Anyone knows about how pricing works on Paperspace for using SD ?
I mean template would still work if it would not relay on external hugging face repo. In that case it's more like breaking it on purpose.
@TheLastBen I need RNPD mainrunpodDreamboothv1.py. Can I use dream booth nothbook the same way in paperspace?
So who is going to refund me for the shitty code, the broken disconnects, the terrible training results, and now the random disconnect of the service I was using?
I highly suggest everyone on this thread put pressure on Runpod to restore the service, or demand refunds. I also highly suggest Ben improves his code and provides detailed settings information for better results. This is absolute nonsense, and we've all been scammed.
@TheLastBen I need RNPD mainrunpodDreamboothv1.py. Can I use dream booth nothbook the same way in paperspace?
Yes
I highly suggest everyone on this thread put pressure on Runpod to restore the service, or demand refunds. I also highly suggest Ben improves his code and provides detailed settings information for better results. This is absolute nonsense, and we've all been scammed.
- RunPod did not remove the service, TheLastBen decided to break it to be spiteful after RunPod decided to stop sponsoring him, but I guess that was a poor call by RunPod considering that they just recently received a significant amount of investment.
- Dreambooth is unpredictable and mostly depends on the quality of your training dataset. Nobody can give you settings that will guarantee better results. You have to run multiple experiments and compare them to find the best results.
I highly suggest everyone on this thread put pressure on Runpod to restore the service, or demand refunds. I also highly suggest Ben improves his code and provides detailed settings information for better results. This is absolute nonsense, and we've all been scammed.
1. RunPod did not remove the service, TheLastBen decided to break it to be spiteful after RunPod decided to stop sponsoring him, but I guess that was a poor call by RunPod considering that they just recently received a significant amount of investment. 2. Dreambooth is unpredictable and mostly depends on the quality of your training dataset. Nobody can give you settings that will guarantee better results. You have to run multiple experiments and compare them to find the best results.
I had a sneaking suspicion it was something like that. As for #2, I'm intimately aware how Dreambooth works, and this is, in fact, false. Assuming a properly prepared dataset, there are a range of settings that are good practice, and many that are not worth trying, objectively. If Ben's repo was as flexible and amazing as he claims, there would be a consistent range of settings that would work with it, but there really aren't - unless you are only training faces or celebrities, with few photos. He's been asked directly about these details, especially regarding how the text encoder works, and he's dodged the question every time. This is a feature of him, not Dreambooth. With other repos, a specific range of settings could be given to make any properly selected and edited dataset work almost every time. With this one? Not so much.
Both Ben and Runpod are responsible for this mess. Ben's (Runpod) notebook has been broken or subpar for a long time and it never got fixed. He should have responded to users and fixed it, and engaged with the community more about how to use it. On the flip side, Runpod is idiotic for cancelling his sponsorship, as Ben's notebook has tons of potential and is very popular, so in that case, I can see why he'd break it for Runpod too. If they are going to break ties with him and not give him what is due for his work, why allow the notebook to work on their service? As with most things, both sides are at fault here, IMO.
I am really confused that Ben said he removed the Colab support because it was unstable. IMO Colab is one of the most stable options out there lol. Hope he can re-consider his decisions - ppl will move to other devs
I highly suggest everyone on this thread put pressure on Runpod to restore the service, or demand refunds. I also highly suggest Ben improves his code and provides detailed settings information for better results. This is absolute nonsense, and we've all been scammed.
- RunPod did not remove the service, TheLastBen decided to break it to be spiteful after RunPod decided to stop sponsoring him, but I guess that was a poor call by RunPod considering that they just recently received a significant amount of investment.
- Dreambooth is unpredictable and mostly depends on the quality of your training dataset. Nobody can give you settings that will guarantee better results. You have to run multiple experiments and compare them to find the best results.
The notebooks were under a contract since the beginning, not sponsorship, the contract ended, the service stops, no spite whatsoever, what happens when your runpod balance goes to zero ? does their service stop out of spite ? do you expect the service to continue for free ? and save 25$/day ?
I highly suggest everyone on this thread put pressure on Runpod to restore the service, or demand refunds. I also highly suggest Ben improves his code and provides detailed settings information for better results. This is absolute nonsense, and we've all been scammed.
- RunPod did not remove the service, TheLastBen decided to break it to be spiteful after RunPod decided to stop sponsoring him, but I guess that was a poor call by RunPod considering that they just recently received a significant amount of investment.
- Dreambooth is unpredictable and mostly depends on the quality of your training dataset. Nobody can give you settings that will guarantee better results. You have to run multiple experiments and compare them to find the best results.
The notebooks were under a contract since the beginning, not sponsorship, the contract ended, the service stops, no spite whatsoever, what happens when your runpod balance goes to zero ? does their service stop out of spite ? do you expect the service to continue for free ? and save 25$/day ?
Okay that makes some sense, but doesn't make sense as to why they wouldn't renew the contract since so many people depend on the template. Using your anaology, if your balance runs out, you add more, you don't just shrug it off and basically say screw you to the users.
Okay that makes some sense, but doesn't make sense as to why they wouldn't renew the contract since so many people depend on the template. Using your anaology, if your balance runs out, you add more, you don't just shrug it off and basically say screw you to the users.
I explicitly told them that it would be a great pleasure continuing to work with them, but eventually, they felt it could be replaced and the users will eventually find a way to copy it and use it for free as a community template.
Ben's (Runpod) notebook has been broken or subpar for a long time and it never got fixed.
what was broken about the notebooks ?
I explicitly told them that it would be a great pleasure continuing to work with them, but eventually, they felt it could be replaced and the users will eventually find a way to copy it and use it for free as a community template.
Good luck with that, the template is/was extremely complex pulling in files and resources from various different sources.
I tried to take some time to understand the layout about a year ago, but just gave up with it. I probably have the most Community templates on RunPod, so I honestly don't see how someone else is going to figure it out either.
There was a code: Args="--listen --port 3000 --preview-method auto"
#-------------------- sd() !python /workspace/ComfyUI/main.py $Args
Could you please write what was inside sd()?
- Dreambooth is unpredictable and mostly depends on the quality of your training dataset. Nobody can give you settings that will guarantee better results. You have to run multiple experiments and compare them to find the best results.
TheLastBen’s trainer is awesome and not unpredictable at all. I have used it for all my finetunings including Aether Cloud.
@TheLastBen oh I see. That's unfortunate. Is colab the only option remaining then? Since paperspace seems to have issues with this as well.
Paperspace is up and running
No idea how to get this working even on paid alternatives.
@jockisen did you try a different machine (gpu) ?
I don’t even know how to do that. The readme your provided seems outdated. You need a paid sub + pay like 3x the amount per gpu hour?
Ben's (Runpod) notebook has been broken or subpar for a long time and it never got fixed.
what was broken about the notebooks ?
I commented about this and sent you messages on various forums multiple times. Other people have created issues up to a year ago. One issue out of many was that it would constantly give you the "port" error. Adding the port to the pod does nothing to resolve this. The other issue was any settings beyond about 20 images resulted in garbage, and specific questions about text encoder values were never answered. In general, it gives worse results than other repos, but got no response from you. As far as the "contract" issue, I don't think anyone really cares. You could have continued to provide it on Runpod for free.
@TheLastBen yeah man, please fix this and provide the trainer for free. I can’t work without it.
Ben's (Runpod) notebook has been broken or subpar for a long time and it never got fixed.
what was broken about the notebooks ?
I commented about this and sent you messages on various forums multiple times. Other people have created issues up to a year ago. One issue out of many was that it would constantly give you the "port" error. Adding the port to the pod does nothing to resolve this. The other issue was any settings beyond about 20 images resulted in garbage, and specific questions about text encoder values were never answered. In general, it gives worse results than other repos, but got no response from you. As far as the "contract" issue, I don't think anyone really cares. You could have continued to provide it on Runpod for free.
Have you figured out how to run it on Paperspace?
Ben's (Runpod) notebook has been broken or subpar for a long time and it never got fixed.
what was broken about the notebooks ?
I commented about this and sent you messages on various forums multiple times. Other people have created issues up to a year ago. One issue out of many was that it would constantly give you the "port" error. Adding the port to the pod does nothing to resolve this. The other issue was any settings beyond about 20 images resulted in garbage, and specific questions about text encoder values were never answered. In general, it gives worse results than other repos, but got no response from you. As far as the "contract" issue, I don't think anyone really cares. You could have continued to provide it on Runpod for free.
Have you figured out how to run it on Paperspace?
No, I gave up on it and am working with Everydream 2 now to try to fix my model.