|
|
--- |
|
|
license: mit |
|
|
library_name: diffusers |
|
|
--- |
|
|
|
|
|
# Mann-E 4.2 Merged |
|
|
|
|
|
## Technical Information about the model |
|
|
|
|
|
* Base Model : [runwayml/stable-diffusion-v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5) |
|
|
* Merge : [mann-e/mann-e_4_rev-1-3](https://huggingface.co/mann-e/mann-e_4_rev-1-3) |
|
|
* Merge amount : %70 fine-tuned SD 1.5 (or _Mann-E version 4.2 base_) and %30 of Mann-E 4.1.3 in order to get the old styles such as _Model Shoot_, _Elden Ring_, _Arcane_, _Analog Style_ and _GTA V Style_. Also this merge can be helpful for _Midjourney version 4_ style artwork as well. |
|
|
|
|
|
### Training process |
|
|
|
|
|
The code for pre-processing data and fine-tuning the model is available in [this repository](https://github.com/prp-e/mann-e_training) and you can run it on your own as well. |
|
|
|
|
|
* Text encoder iterations : 1440 (number of pics times two in order to understand `mstyle` which can give the user a _Midjourney version 5_ vibe). |
|
|
* Stable Diffusion iterations : 16000 iterations for one epoch |
|
|
* Time: around 4 hours on a single T4 GPU. |
|
|
|
|
|
|
