.

.tarkan•2/11/25, 8:06 PM

Does it make sense in certain cases to use a different guidance scale for flux dev dedistilled training?

AAiInfluence @Furkan Gözükara SECourses

F

Furkan Gözükara SECoursesOP•2/11/25, 8:06 PM

yolo face

..tarkan Does it make sense in certain cases to use a different guidance scale for flux d...

F

Furkan Gözükara SECoursesOP•2/11/25, 8:06 PM

3.5 is best

F

Furkan Gözükara SECoursesOP•2/11/25, 8:06 PM

i compared all :d

F

Furkan Gözükara SECoursesOP•2/11/25, 8:06 PM

patreon has grids you can check

MMido anyone here having issue when training a lora for flux using kohyass?

F

Furkan Gözükara SECoursesOP•2/11/25, 8:06 PM

what kind of error? i am training no issues

Ccyleap @Furkan Gözükara SECourses hey, i installed invoke ai on my linux machine locall...

F

Furkan Gözükara SECoursesOP•2/11/25, 8:06 PM

you need to put into accurate folders

FFurkan Gözükara SECourses what kind of error? i am training no issues

M

Mido•2/11/25, 8:06 PM

gimme a sec

VVioletAnt @Furkan Gözükara SECourses I did training of flux model. And get very nice looki...

F

Furkan Gözükara SECoursesOP•2/11/25, 8:07 PM

100%

F

Furkan Gözükara SECoursesOP•2/11/25, 8:07 PM

according to the prompt it will use the expression your dataset has

F

Furkan Gözükara SECoursesOP•2/11/25, 8:07 PM

automatically

MMido Whenever i train kohyass lora on a flux model, i get error "returned non-zero e...

M

Mido•2/11/25, 8:07 PM

 File "/data/app/sd-scripts/library/model_util.py", line 267, in convert_ldm_unet_checkpoint
    new_checkpoint["time_embedding.linear_1.weight"] = unet_state_dict["time_embed.0.weight"]
KeyError: 'time_embed.0.weight'
Traceback (most recent call last):
  File "/data/app/venv/bin/accelerate", line 8, in <module>
    sys.exit(main())
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/accelerate_cli.py", line 47, in main
    args.func(args)
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 1017, in launch_command
    simple_launcher(args)
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 637, in simple_launcher
    raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)
subprocess.CalledProcessError: Command '['/data/app/venv/bin/python', '/data/app/sd-scripts/train_network.py', '--config_file', '/data/app/input/dest-flux/model/config_lora-20250208-100919.toml']' returned non-zero exit status 1.
10:10:40-131591 INFO     Training has ended.

 File "/data/app/sd-scripts/library/model_util.py", line 267, in convert_ldm_unet_checkpoint
    new_checkpoint["time_embedding.linear_1.weight"] = unet_state_dict["time_embed.0.weight"]
KeyError: 'time_embed.0.weight'
Traceback (most recent call last):
  File "/data/app/venv/bin/accelerate", line 8, in <module>
    sys.exit(main())
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/accelerate_cli.py", line 47, in main
    args.func(args)
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 1017, in launch_command
    simple_launcher(args)
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 637, in simple_launcher
    raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)
subprocess.CalledProcessError: Command '['/data/app/venv/bin/python', '/data/app/sd-scripts/train_network.py', '--config_file', '/data/app/input/dest-flux/model/config_lora-20250208-100919.toml']' returned non-zero exit status 1.
10:10:40-131591 INFO     Training has ended.

 File "/data/app/sd-scripts/library/model_util.py", line 267, in convert_ldm_unet_checkpoint
    new_checkpoint["time_embedding.linear_1.weight"] = unet_state_dict["time_embed.0.weight"]
KeyError: 'time_embed.0.weight'
Traceback (most recent call last):
  File "/data/app/venv/bin/accelerate", line 8, in <module>
    sys.exit(main())
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/accelerate_cli.py", line 47, in main
    args.func(args)
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 1017, in launch_command
    simple_launcher(args)
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 637, in simple_launcher
    raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)
subprocess.CalledProcessError: Command '['/data/app/venv/bin/python', '/data/app/sd-scripts/train_network.py', '--config_file', '/data/app/input/dest-flux/model/config_lora-20250208-100919.toml']' returned non-zero exit status 1.
10:10:40-131591 INFO     Training has ended.

 File "/data/app/sd-scripts/library/model_util.py", line 267, in convert_ldm_unet_checkpoint
    new_checkpoint["time_embedding.linear_1.weight"] = unet_state_dict["time_embed.0.weight"]
KeyError: 'time_embed.0.weight'
Traceback (most recent call last):
  File "/data/app/venv/bin/accelerate", line 8, in <module>
    sys.exit(main())
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/accelerate_cli.py", line 47, in main
    args.func(args)
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 1017, in launch_command
    simple_launcher(args)
  File "/data/app/venv/lib/python3.10/site-packages/accelerate/commands/launch.py", line 637, in simple_launcher
    raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)
subprocess.CalledProcessError: Command '['/data/app/venv/bin/python', '/data/app/sd-scripts/train_network.py', '--config_file', '/data/app/input/dest-flux/model/config_lora-20250208-100919.toml']' returned non-zero exit status 1.
10:10:40-131591 INFO     Training has ended.

MMido Click to see attachment

M

Mido•2/11/25, 8:07 PM

here's everything related to it

MMido ```cmd File "/data/app/sd-scripts/library/model_util.py", line 267, in convert_...

F

Furkan Gözükara SECoursesOP•2/11/25, 8:07 PM

inaccurate model

F

Furkan Gözükara SECoursesOP•2/11/25, 8:08 PM

could be corrupt as well

FFurkan Gözükara SECourses could be corrupt as well

M

Mido•2/11/25, 8:08 PM

had that in mind, reinstalled it multiple times

M

Mido•2/11/25, 8:08 PM

and tried to built in version of it in mimicpc

M

Mido•2/11/25, 8:08 PM

same issue

MMido had that in mind, reinstalled it multiple times

F

Furkan Gözükara SECoursesOP•2/11/25, 8:08 PM

well that is the error reasson

F

Furkan Gözükara SECoursesOP•2/11/25, 8:08 PM

so still you have that issue

F

Furkan Gözükara SECoursesOP•2/11/25, 8:08 PM

hard to know without seeing everything

F

Furkan Gözükara SECoursesOP•2/11/25, 8:08 PM

are you gold member?

F

Furkan Gözükara SECoursesOP•2/11/25, 8:09 PM

i can connect pc and check

M

Mido•2/11/25, 8:09 PM

i am not using it on PC, using it on runpod and mimicpc

MMido i am not using it on PC, using it on runpod and mimicpc

F

Furkan Gözükara SECoursesOP•2/11/25, 8:10 PM

in that case how you download model onto there the error

FFurkan Gözükara SECourses in that case how you download model onto there the error

M

Mido•2/11/25, 8:11 PM

curl -L -o model.safetensors -H "Authorization: Bearer XXX" "https://civitai.com/api/download/models/691639?type=Model&format=SafeTensor&size=full&fp=fp32"

M

Mido•2/11/25, 8:11 PM

removed the bearer tho

FFurkan Gözükara SECourses you need to put into accurate folders

C

cyleap•2/11/25, 8:18 PM

invoke dosent have models directory like comfyui or a1111, do you know which folder?

FFurkan Gözükara SECourses did it fix?

K

kareem9000•2/11/25, 8:22 PM

About to try will update u

Ccyleap invoke dosent have models directory like comfyui or a1111, do you know which fol...

F

Furkan Gözükara SECoursesOP•2/11/25, 8:24 PM

sadly i dont know

F

Furkan Gözükara SECoursesOP•2/11/25, 8:24 PM

download few and see where it saves

MMido curl -L -o model.safetensors -H "Authorization: Bearer XXX" "https://civitai.com...

F

Furkan Gözükara SECoursesOP•2/11/25, 8:25 PM

upload hugging face

F

Furkan Gözükara SECoursesOP•2/11/25, 8:25 PM

and download from there

F

Furkan Gözükara SECoursesOP•2/11/25, 8:25 PM

but verify on your pc it works or not

M

Mido•2/11/25, 8:25 PM

ok ill give it a shot and keep you updated, thank you

FFurkan Gözükara SECourses download few and see where it saves

C

cyleap•2/11/25, 8:27 PM

ohh ok, thanks ill try.

VVioletAnt @Furkan Gözükara SECourses I did training of flux model. And get very nice looki...

J

julius•2/11/25, 10:40 PM

Generated images will by default look similar to what’s most common in your in your dataset. Sometimes you can’t even prompt you way to what you want if your dataset does not include it, or at least the generation might only follow your prompt sometimes. But I mean for the most part it’s possible to generate many different styles of photos no matter what. But yea will need explicit prompting for smiles most the time for you if your dataset has very little of it

A

Archi_teknik101•2/12/25, 12:44 AM

hey there i'm trying to run the kijai node for hunyuan
https://github.com/kijai/ComfyUI-Hunyuan3DWrapper
the first part builds correctly then i gte this error when generating the multivbiews i guess ?

GitHub

GitHub - kijai/ComfyUI-Hunyuan3DWrapper

Contribute to kijai/ComfyUI-Hunyuan3DWrapper development by creating an account on GitHub.

A

Archi_teknik101•2/12/25, 12:44 AM

ComfyUI Error Report

Error Details

Node ID: 79
Node Type: Hy3DRenderMultiView
Exception Type: ModuleNotFoundError
**Exception Message:** No module named 'custom_rasterizer'
Stack Trace
```File "C:\Users\Utilisateur\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\execution.py", line 327, in execute output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^File "C:\Users\Utilisateur\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\execution.py", line 202, in get_output_data return_values = _map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^File "C:\Users\Utilisateur\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\execution.py", line 174, in _map_node_over_list process_inputs(input_dict, i)File "C:\Users\Utilisateur\AppData\Local\Programs\@comfyorgcomfyui-electron\resources\ComfyUI\execution.py", line 163, in process_inputs results.append(getattr(obj, func)(**inputs)) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^File "C:\Users\Utilisateur\Documents\ComfyUI\custom_nodes/comfyui-hunyan3dwrapper\nodes.py", line 433, in process self.render = MeshRender( ^^^^^^^^^^^File "C:\Users\Utilisateur\Documents\ComfyUI\custom_nodes/comfyui-hunyan3dwrapper\hy3dgen\texgen\differentiable_renderer\mesh_render.py", line 158, in __init__ import custom_rasterizer as cr

```

A

Archi_teknik101•2/12/25, 12:45 AM

i'v seen on the repo that they are similar errors but i'm quite clueless on how to proceed, any idea ?

A

AIGambino•2/12/25, 3:52 AM

https://tobias-kirschstein.github.io/nersemble/

NeRSemble: Multi-view Radiance Field Reconstruction of Human Heads

AArchi_teknik101 hey there i'm trying to run the kijai node for hunyuan https://github.com/kijai...

F

Furkan Gözükara SECoursesOP•2/12/25, 11:40 AM

i didnt try comfyui yet but gradio tutorial coming soon hopefully

AAIGambino https://tobias-kirschstein.github.io/nersemble/

F

Furkan Gözükara SECoursesOP•2/12/25, 11:41 AM

what it takes to generate?

F

Furkan Gözükara SECoursesOP•2/12/25, 11:48 AM

working on upgrade torch to swarmui so that rtx 5090 series can use

F

Furkan Gözükara SECoursesOP•2/12/25, 12:45 PM

RTX 5090 getting 2.15 it / second at FP16 for FLUX dev

F

Furkan Gözükara SECoursesOP•2/12/25, 12:52 PM

rtx 5090 vs 3090 ti at fp8

FFurkan Gözükara SECourses rtx 5090 vs 3090 ti at fp8

M

Mido•2/12/25, 1:07 PM

yo, i managed to fix the issue earlier

MMido ok ill give it a shot and keep you updated, thank you ❤️

M

Mido•2/12/25, 1:07 PM

this one

M

Mido•2/12/25, 1:08 PM

but i have a question

M

Mido•2/12/25, 1:08 PM

i am making a model for instagram, do you suggest using reg images? because in a video tutorial i watched for you you did not use reg images, but instead you trained it on 200 epochs, so which one is better for my case?

did it fix?

ComfyUI Error Report

Error Details

Stack Trace

Similar Threads

ComfyUI Error Report

Error Details

Stack Trace