Theoretically: yes Has it been done? No clue.

Hello guys! I need your help. I'm trying to make a model with talknet 2 (first time doing it) and I have a problem on step 6. I get this error: RuntimeError: CUDA error: no kernel image is available for execution on the device I am using colab pro, gpu: A100-SXM4-40GB Could you help me with that? I have no idea what I am doing. I am a content creator so I have no clue about python.

Okay, so I fixed it with using a different gpu. Now I'm having problems with controllable talknet. On step 3 I get error 403. On step 3B I get ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject I am using chrome without adblock.

Anyone know how to fix this? I wanna test my model.

I switched to microsoft edge and it works!

Questino

Every single time I attempt to load a model i just made into the synthesis notebook, I get the following small error under the generate button; "Model download failed"

I've tried to do different things. I'm not understanding why it's giving this as I should've trained the model fully and it's all there in the drive folder

As seen here

Help someone please. I know that it seems normally this channel goes unnoticed

I think there's something wrong with Google Drive today or something I dunno, ask hecko

Everyone is having issues?

@hecko Hecko. We have a problem.

yep

2 of my latest checkpoints didn't save

So is all of the drives for everyone not working today?

Oh so it is

I made a model that i trained all night with over 700 epoch

@Gosmokeless28 please make an announcement

and nearly 3,000 steps

What do you want me to say?

It seems my work saved but everything else to load it just refused to work

google drive is struggling currently, please wait before training any models as they might not save properly

I noticed that for some reason some files were giving "access denied" stuff and it was in fact ok

Yeah it's everyone

Has this happened before?

not to my knowledge

hold on gonna rephrase