fast-stable-diffusion: Attempt to free invalid pointer

hello, I commented out this line (in Requirements cell) :

%env LD_PRELOAD=libtcmalloc.so --> # %env LD_PRELOAD=libtcmalloc.so

I also did (EDIT: this one is not needed):

• !pip install open-clip-torch==2.20.0 -qq --no-deps

• !pip install --no-cache-dir open-clip-torch==2.20.0 -qq --no-deps

It worked again, does it works for you ? I did the two changes at once so cannot tell if this is the first, the second or the both. The changes have to be done in the ‘Requirements’ cell

I’ll check it out, probably another google’s attempt to block SD, I really suggest migrating to other hosts such as runpod and paperspace, they are cheaper and faster

I understand mdlieber. so, what you can do here, with the colab of ben, is go to the File menu, “save a copy into Drive”

Once it done it should open a new version of the Colab project that is your version. so now, you can bring change to your own google colab (note that it will not update anymore too, that’s the annoying stuff).

Click one Show Code here :

Then modify the lines like I said before :

use the latest notebook, but you need to set the runtime to High-RAM

I’ll check it out, probably another google’s attempt to block SD, I really suggest migrating to other hosts such as runpod and paperspace, they are cheaper and faster

Ben, thanks a lot for your works and help, I tried other services, but for me Google is more comfortable. But if there is a block again, I will return to other services.

hello, I commented out this line (in Requirements cell) : %env LD_PRELOAD=libtcmalloc.so --> # %env LD_PRELOAD=libtcmalloc.so It worked again, those it works for you ? I also did :

• !pip install open-clip-torch==2.20.0 -qq --no-deps
• !pip install --no-cache-dir open-clip-torch==2.20.0 -qq --no-deps

I did the two changes at once so cannot tell if this is the first, the second or the both. The changes have to be done in the ‘Requirements’ cell

I did this and it worked for me. All I had to do was comment out that first line, I didn’t add the other two.

ha, cool : )

hello, I commented out this line (in Requirements cell) :

%env LD_PRELOAD=libtcmalloc.so --> # %env LD_PRELOAD=libtcmalloc.so

It worked again, those it works for you ? I also did :

• !pip install open-clip-torch==2.20.0 -qq --no-deps

• !pip install --no-cache-dir open-clip-torch==2.20.0 -qq --no-deps

I did the two changes at once so cannot tell if this is the first, the second or the both. The changes have to be done in the ‘Requirements’ cell

I did this and it worked for me. All I had to do was comment out that first line, I didn’t add the other two.

I’m a PRO user in Google and I got the same error just today, I was working with it yesterday until 1 am. So let’s figure out how to do it in Runpod or Paperspace and cancel Google’s subs.🤷🏻‍♂️

I honestly have never been able to figure out how to get the other services to work and I assume they require moving all my lora’s and models out of google drive and into some other online drive?

I also find it easier to use google colab for SD. Hopefully it’s just a simple bug and doesn’t restrict us from using it because for PRO or free users, the same error occurs.

Same error here. It was working an hour ago.

I honestly have never been able to figure out how to get the other services to work and I assume they require moving all my lora’s and models out of google drive and into some other online drive?

Same error. Colab PRO user. Was working at ~1:30am est… Today NG across multiple browsers.

• the same

Same error src/tcmalloc.cc:283] Attempt to free invalid pointer 0x58376413f9d0

I’ve switch over to runpod. Best decision of my life. Selectable GPU and its fast as hell.

I’m very confused on how to upload custom models, loras, embeddings ect, with the other services. Do you have to purchase google drive-like storage with them?

I’ve switch over to runpod. Best decision of my life. Selectable GPU and its fast as hell.

Adding my error report too. Was working this morning before work.

Adding my error too. Same one. Was using it, shut down for for 20 minutes, tried to fire it back up and the error shows up.

same here

same error