Etched/oasis-500m
Updated
•
5.33k
•
415
Sergio K PRO
Sergidev
AI & ML interests
LLM & Deep Learning
Recent Activity
liked
a model
19 days ago
Etched/oasis-500m
updated
a Space
20 days ago
Sergidev/360PanoImage
liked
a Space
20 days ago
mrfakename/E2-F5-TTS
Organizations
Sergidev's activity
liked
a
model
19 days ago
updated
a
Space
20 days ago
liked
a
Space
20 days ago
updated
a
Space
20 days ago
liked
2
Spaces
about 1 month ago
updated
a
Space
about 1 month ago
liked
a
Space
about 2 months ago
liked
a
model
about 2 months ago
updated
a
Space
about 2 months ago
Reacted to
MonsterMMORPG's
post with 🔥
2 months ago
Post
3304
Full Fine Tuning of FLUX yields way better results than LoRA training as expected, overfitting and bleeding reduced a lot
Configs and Full Experiments
Full configs and grid files shared here : https://www.patreon.com/posts/kohya-flux-fine-112099700
Details
I am still rigorously testing different hyperparameters and comparing impact of each one to find the best workflow
So far done 16 different full trainings and completing 8 more at the moment
I am using my poor overfit 15 images dataset for experimentation (4th image)
I have already proven that when I use a better dataset it becomes many times betters and generate expressions perfectly
Here example case : https://www.reddit.com/r/FluxAI/comments/1ffz9uc/tried_expressions_with_flux_lora_training_with_my/
Conclusions
When the results are analyzed, Fine Tuning is way lesser overfit and more generalized and better quality
In first 2 images, it is able to change hair color and add beard much better, means lesser overfit
In the third image, you will notice that the armor is much better, thus lesser overfit
I noticed that the environment and clothings are much lesser overfit and better quality
Disadvantages
Kohya still doesn’t have FP8 training, thus 24 GB GPUs gets a huge speed drop
Moreover, 48 GB GPUs has to use Fused Back Pass optimization, thus have some speed drop
16 GB GPUs gets way more aggressive speed drop due to lack of FP8
Clip-L and T5 trainings still not supported
Speeds
Rank 1 Fast Config — uses 27.5 GB VRAM, 6.28 second / it (LoRA is 4.85 second / it)
Rank 1 Slower Config — uses 23.1 GB VRAM, 14.12 second / it (LoRA is 4.85 second / it)
Rank 1 Slowest Config — uses 15.5 GB VRAM, 39 second / it (LoRA is 6.05 second / it)
Final Info
Saved checkpoints are FP16 and thus 23.8 GB (no Clip-L or T5 trained)
According to the Kohya, applied optimizations doesn’t change quality so all configs are ranked as Rank 1 at the moment
I am still testing whether these optimizations make any impact on quality or not
Configs and Full Experiments
Full configs and grid files shared here : https://www.patreon.com/posts/kohya-flux-fine-112099700
Details
I am still rigorously testing different hyperparameters and comparing impact of each one to find the best workflow
So far done 16 different full trainings and completing 8 more at the moment
I am using my poor overfit 15 images dataset for experimentation (4th image)
I have already proven that when I use a better dataset it becomes many times betters and generate expressions perfectly
Here example case : https://www.reddit.com/r/FluxAI/comments/1ffz9uc/tried_expressions_with_flux_lora_training_with_my/
Conclusions
When the results are analyzed, Fine Tuning is way lesser overfit and more generalized and better quality
In first 2 images, it is able to change hair color and add beard much better, means lesser overfit
In the third image, you will notice that the armor is much better, thus lesser overfit
I noticed that the environment and clothings are much lesser overfit and better quality
Disadvantages
Kohya still doesn’t have FP8 training, thus 24 GB GPUs gets a huge speed drop
Moreover, 48 GB GPUs has to use Fused Back Pass optimization, thus have some speed drop
16 GB GPUs gets way more aggressive speed drop due to lack of FP8
Clip-L and T5 trainings still not supported
Speeds
Rank 1 Fast Config — uses 27.5 GB VRAM, 6.28 second / it (LoRA is 4.85 second / it)
Rank 1 Slower Config — uses 23.1 GB VRAM, 14.12 second / it (LoRA is 4.85 second / it)
Rank 1 Slowest Config — uses 15.5 GB VRAM, 39 second / it (LoRA is 6.05 second / it)
Final Info
Saved checkpoints are FP16 and thus 23.8 GB (no Clip-L or T5 trained)
According to the Kohya, applied optimizations doesn’t change quality so all configs are ranked as Rank 1 at the moment
I am still testing whether these optimizations make any impact on quality or not
- 2 replies
updated
2
Spaces
2 months ago
updated
2
Spaces
3 months ago
liked
a
Space
3 months ago
updated
2
Spaces
3 months ago
Reacted to
clem's
post with 🔥
3 months ago
Post
4126
Just crossed 200,000 free public AI datasets shared by the community on Hugging Face! Text, image, video, audio, time-series & many more... Thanks everyone!
http://hf.co/datasets
http://hf.co/datasets
liked
a
Space
4 months ago