Software Engineering Courses (SECourses)•3y ago

For the sake of reporting, this was the error I was getting. Exception initializing accelerator: Acc

For the sake of reporting, this was the error I was getting. Exception initializing accelerator: Accelerator.init() got an unexpected keyword argument 'project_dir'

I AOP•3/16/23, 6:45 PM

Maybe it can be useful. In my case after many hours online, I got easily distracted and didn't see that there are two files: requirements & requirements versions

I AOP•3/16/23, 6:46 PM

If the training starts and finishes correctly, I will share my experience with you Dr Furkan.

I AOP•3/16/23, 7:07 PM

First update: In Dreambooth optimizer 8bit Adam is not there anymore. I have 1. Torch AdamW 2. Lion

ZZono50 may train with 0.0005 next time or somethin

Furkan Gözükara SECourses•3/16/23, 7:16 PM

yes this requires experimentation

II A For the sake of reporting, this was the error I was getting. Exception initializ...

Furkan Gözükara SECourses•3/16/23, 7:17 PM

yep this is version error

Furkan Gözükara SECourses•3/16/23, 7:17 PM

i spent like 5 hours first time i was trying to fix it

I AOP•3/16/23, 7:19 PM

OK, the training started, so I can confirm that error above was my mistake.

I AOP•3/16/23, 7:19 PM

Now with the exact same settings, as in your video, I'm getting this error:

I AOP•3/16/23, 7:19 PM

Exception training model: ''NoneType' object is not subscriptable'.

I AOP•3/16/23, 7:20 PM

I will try to figure out what to change

II A Exception training model: ''NoneType' object is not subscriptable'.

Furkan Gözükara SECourses•3/16/23, 7:22 PM

can you show entire message

I AOP•3/16/23, 7:23 PM

Do you want me to copy/paste message from terminal here? Or in private?

I AOP•3/16/23, 7:23 PM

It's long.

I AOP•3/16/23, 7:29 PM

Please check DM

II A Please check DM

Furkan Gözükara SECourses•3/16/23, 7:29 PM

you can paste here

I AOP•3/16/23, 7:32 PM

Running training
Num batches each epoch = 0
Num Epochs = 200
Batch Size Per Device = 1
Gradient Accumulation steps = 1
Total train batch size (w. parallel, distributed & accumulation) = 1
Text Encoder Epochs: 150
Total optimization steps = 0
Total training steps = 0
Resuming from checkpoint: False
First resume epoch: 0
First resume step: 0
Lora: False, Optimizer: Torch AdamW, Prec: fp16
Gradient Checkpointing: False
EMA: True
UNET: True
Freeze CLIP Normalization Layers: False
LR: 1e-06
V2: False
Steps: : 0it [00:00, ?it/s]Traceback (most recent call last):
File "/workspace/stable-diffusion-webui/extensions/sd_dreambooth_extension/dreambooth/ui_functions.py", line 727, in start_training
result = main(class_gen_method=class_gen_method)
File "/workspace/stable-diffusion-webui/extensions/sd_dreambooth_extension/dreambooth/train_dreambooth.py", line 1448, in main
return inner_loop()
File "/workspace/stable-diffusion-webui/extensions/sd_dreambooth_extension/dreambooth/memory.py", line 119, in decorator
return function(batch_size, grad_size, prof, *args, **kwargs)
File "/workspace/stable-diffusion-webui/extensions/sd_dreambooth_extension/dreambooth/train_dreambooth.py", line 1200, in inner_loop
latents = batch["images"].to(accelerator.device)
TypeError: 'NoneType' object is not subscriptable
Steps: : 0it [00:00, ?it/s]
Restored system models.
Duration: 00:00:08

II A ***** Running training ***** Num batches each epoch = 0 Num Epochs = 200 B...

Furkan Gözükara SECourses•3/16/23, 7:49 PM

can you

Furkan Gözükara SECourses•3/16/23, 7:49 PM

do pip reinstall -r requirements.txt again?

Furkan Gözükara SECourses•3/16/23, 7:49 PM

also you fixed versions right?

Furkan Gözükara SECourses•3/16/23, 7:50 PM

you are using filewords?

I AOP•3/16/23, 7:50 PM

Yes, versions fixed. Accelerate updated to 0.17 this morning was 0.16

Furkan Gözükara SECourses•3/16/23, 7:50 PM

"I confirm, there is issue. I had the same error when some of "filewords" had the UTF16 LE encoding. With UTF8 it works properly."

Furkan Gözükara SECourses•3/16/23, 7:50 PM

https://github.com/d8ahazard/sd_dreambooth_extension/issues/893

GitHub

TypeError: 'NoneType' object is not subscriptable · Issue #893 · d8...

File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/extensions/sd_dreambooth_extension/scripts/dreambooth.py", line 455, in start_training result = main(config, use_txt2img=us...

I AOP•3/16/23, 7:51 PM

I replicated your tutorial, just using my own token.

I AOP•3/16/23, 7:51 PM

I will try the reinstall and update you

II A I will try the reinstall and update you

Furkan Gözükara SECourses•3/16/23, 7:52 PM

do you use file captions? filewords?

Furkan Gözükara SECourses•3/16/23, 7:54 PM

i asked to extension developer

I AOP•3/16/23, 7:59 PM

No I don't think so. I'm only using images, renamed with a little description.

I AOP•3/16/23, 7:59 PM

Now after the I reinstalled the requirements, I got a new error, so I can't launch web ui:

I AOP•3/16/23, 8:00 PM

I gues the GPU is not avilable on Runpod, after I stopped it. GPUs have high availability, but my pod is set to 0 GPU.

FFurkan Gözükara SECourses i asked to extension developer

I AOP•3/16/23, 8:03 PM

Thank you!

I AOP•3/16/23, 8:20 PM

I'm still waiting, lost GPU spot, now redoing everything again on a new pod.

I AOP•3/16/23, 8:20 PM

Will update you in a few minutes

Zono50•3/16/23, 8:32 PM

with the inspect emedding training program, do we have to check magnitude for overtraining or just strength?

BRKO•3/16/23, 8:40 PM

2.1 modeli çalıştırınca siyah box çıktısı alıyor sadece nasıl çözebilirim?

ZZono50 with the inspect emedding training program, do we have to check magnitude for ov...

Furkan Gözükara SECourses•3/16/23, 8:43 PM

you dont have to. best way is always comparing checkpoints

I AOP•3/16/23, 8:58 PM

OK, I reinstalled everything on a new machine, and I get the same erorr. But I'm not using filewords.

II A OK, I reinstalled everything on a new machine, and I get the same erorr. But I'm...

Furkan Gözükara SECourses•3/16/23, 8:59 PM

your classification directory is invalid

Furkan Gözükara SECourses•3/16/23, 9:00 PM

put / beginning @I A

I AOP•3/16/23, 9:00 PM

Yes I did..white space?

II A Yes I did..white space?

Furkan Gözükara SECourses•3/16/23, 9:02 PM

by the way you written them with incorrect order. above one is training below one is classification

Furkan Gözükara SECourses•3/16/23, 9:02 PM

you noticed that right?

I AOP•3/16/23, 9:03 PM

NO! Let me check...thanks

I AOP•3/16/23, 9:12 PM

Oh yeah, it's working now. Your the man! Thanks

I AOP•3/16/23, 9:12 PM

I will update you later.

Furkan Gözükara SECourses•3/16/23, 9:13 PM

you are welcome

I AOP•3/16/23, 9:14 PM

Meanwhile I used some extra settings, based on your conclusions, like offset noise to 0.07, image genertion to DEISMultistep, and testing to deterministic.

Zono50•3/16/23, 9:34 PM

training lower rate for longer seems to raise the magnitude which makes the embedding not play well with other prompts, so it seems like 0.005 is about as good as it's gonna get. Wonder if i make the backgrounds transparent if that will allow the ai to focus on the model better

I AOP•3/16/23, 9:43 PM

So far so good, but gradio froze, so I cannot track the progress. I assume if I close the window, the training will stop? I can see it running on the terminal, tough.

II A So far so good, but gradio froze, so I cannot track the progress. I assume if I ...

Furkan Gözükara SECourses•3/16/23, 9:46 PM

follow the terminal. yes unfortunately gradio freezes

For the sake of reporting, this was the error I was getting. Exception initializing accelerator: Acc

Similar Threads