idlebg's picture
img temp fix
c3aee77
---
license: creativeml-openrail-m
language:
- en
tags:
- di.ffusion.ai
- stable-diffusion
- LyCORIS
- LoRA
---
# Model Card for di.FFUSION.ai Text Encoder - SD 2.1 LyCORIS
![image.png](https://cdn-uploads.huggingface.co/production/uploads/6380cf05f496d57325c12194/zcw9AUCSbanb61xe6pIUc.png)
<!-- Provide a quick summary of what the model is/does. [Optional] -->
di.FFUSION.ai-tXe-FXAA
Trained on &#34;121361&#34; images.
- **DOWNLOAD:** https://huggingface.co/FFusion/FFUSION.ai-Text-Encoder-LyCORIS-SD-2.1/blob/main/di.FFUSION.ai-tXe-FXAA.safetensors
Enhance your model&#39;s quality and sharpness using your own pre-trained Unet.
The text encoder (without UNET) is wrapped in LyCORIS. Optimizer: torch.optim.adamw.AdamW(weight_decay=0.01, betas=(0.9, 0.99))
Network dimension/rank: 768.0 Alpha: 768.0 Module: lycoris.kohya {&#39;conv_dim&#39;: &#39;256&#39;, &#39;conv_alpha&#39;: &#39;256&#39;, &#39;algo&#39;: &#39;loha&#39;}
Large size due to Lyco CONV 256
![image.png](https://cdn-uploads.huggingface.co/production/uploads/6380cf05f496d57325c12194/Ig1IOYZAyUrhpWIhdC6U-.png)
![image.png](https://cdn-uploads.huggingface.co/production/uploads/6380cf05f496d57325c12194/66eAHPc501sbQx35-B0Oo.png)
This is a heavy experimental version we used to test even with sloppy captions (quick WD tags and terrible clip), yet the results were satisfying.
Note: This is not the text encoder used in the official FFUSION AI model.
# SAMPLES
**Available also at https://civitai.com/models/83622**
![image.png](https://cdn-uploads.huggingface.co/production/uploads/6380cf05f496d57325c12194/agjJ--YR_k_Pbn8tOMsqr.png)
For a1111
Install https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris
Download di.FFUSION.ai-tXe-FXAA to /models/Lycoris
Option1:
Insert <lyco:di.FFUSION.ai-tXe-FXAA:1.0> to prompt
No need to split Unet and Text Enc as its only TX encoder there.
You can go up to 2x weights
Option2: If you need it always ON (ex run a batch from txt file) then you can go to settings / Quicksettings list
![image.png](https://cdn-uploads.huggingface.co/production/uploads/6380cf05f496d57325c12194/N6M4-9eIkvi3nn3koh1fA.png)
add sd_lyco
restart and you should have a drop-down now 🤟 🥃
![image.png](https://cdn-uploads.huggingface.co/production/uploads/6380cf05f496d57325c12194/e8ROXaN8jIaT9lu7tNRjD.png)
# Table of Contents
- [Model Card for di.FFUSION.ai Text Encoder - SD 2.1 LyCORIS](#model-card-for--model_id-)
- [Table of Contents](#table-of-contents)
- [Table of Contents](#table-of-contents-1)
- [Model Details](#model-details)
- [Model Description](#model-description)
- [Uses](#uses)
- [Direct Use](#direct-use)
- [Downstream Use [Optional]](#downstream-use-optional)
- [Out-of-Scope Use](#out-of-scope-use)
- [Bias, Risks, and Limitations](#bias-risks-and-limitations)
- [Recommendations](#recommendations)
- [Training Details](#training-details)
- [Training Data](#training-data)
- [Training Procedure](#training-procedure)
- [Preprocessing](#preprocessing)
- [Speeds, Sizes, Times](#speeds-sizes-times)
- [Evaluation](#evaluation)
- [Testing Data, Factors & Metrics](#testing-data-factors--metrics)
- [Testing Data](#testing-data)
- [Factors](#factors)
- [Metrics](#metrics)
- [Results](#results)
- [Model Examination](#model-examination)
- [Environmental Impact](#environmental-impact)
- [Technical Specifications [optional]](#technical-specifications-optional)
- [Model Architecture and Objective](#model-architecture-and-objective)
- [Compute Infrastructure](#compute-infrastructure)
- [Hardware](#hardware)
- [Software](#software)
- [Citation](#citation)
- [Glossary [optional]](#glossary-optional)
- [More Information [optional]](#more-information-optional)
- [Model Card Authors [optional]](#model-card-authors-optional)
- [Model Card Contact](#model-card-contact)
- [How to Get Started with the Model](#how-to-get-started-with-the-model)
# Model Details
## Model Description
<!-- Provide a longer summary of what this model is/does. -->
di.FFUSION.ai-tXe-FXAA
Trained on &#34;121361&#34; images.
Enhance your model&#39;s quality and sharpness using your own pre-trained Unet.
The text encoder (without UNET) is wrapped in LyCORIS. Optimizer: torch.optim.adamw.AdamW(weight_decay=0.01, betas=(0.9, 0.99))
Network dimension/rank: 768.0 Alpha: 768.0 Module: lycoris.kohya {&#39;conv_dim&#39;: &#39;256&#39;, &#39;conv_alpha&#39;: &#39;256&#39;, &#39;algo&#39;: &#39;loha&#39;}
Large size due to Lyco CONV 256
This is a heavy experimental version we used to test even with sloppy captions (quick WD tags and terrible clip), yet the results were satisfying.
Note: This is not the text encoder used in the official FFUSION AI model.
- **Developed by:** FFusion.ai
- **Shared by [Optional]:** idle stoev
- **Model type:** Language model
- **Language(s) (NLP):** en
- **License:** creativeml-openrail-m
- **Parent Model:** More information needed
- **Resources for more information:** More information needed
# Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
## Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
<!-- If the user enters content, print that. If not, but they enter a task in the list, use that. If neither, say "more info needed." -->
The text encoder (without UNET) is wrapped in LyCORIS. Optimizer: torch.optim.adamw.AdamW(weight_decay=0.01, betas=(0.9, 0.99))
Network dimension/rank: 768.0 Alpha: 768.0 Module: lycoris.kohya {&#39;conv_dim&#39;: &#39;256&#39;, &#39;conv_alpha&#39;: &#39;256&#39;, &#39;algo&#39;: &#39;loha&#39;}
Large size due to Lyco CONV 256
# Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). Predictions generated by the model may include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups.
## Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
# Training Details
## Training Data
<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
Trained on &#34;121361&#34; images.
ss_caption_tag_dropout_rate: &#34;0.0&#34;,
ss_multires_noise_discount: &#34;0.3&#34;,
ss_mixed_precision: &#34;bf16&#34;,
ss_text_encoder_lr: &#34;1e-07&#34;,
ss_keep_tokens: &#34;3&#34;,
ss_network_args: &#34;{&#34;conv_dim&#34;: &#34;256&#34;, &#34;conv_alpha&#34;: &#34;256&#34;, &#34;algo&#34;: &#34;loha&#34;}&#34;,
ss_caption_dropout_rate: &#34;0.02&#34;,
ss_flip_aug: &#34;False&#34;,
ss_learning_rate: &#34;2e-07&#34;,
ss_sd_model_name: &#34;stabilityai/stable-diffusion-2-1-base&#34;,
ss_max_grad_norm: &#34;1.0&#34;,
ss_num_epochs: &#34;2&#34;,
ss_gradient_checkpointing: &#34;False&#34;,
ss_face_crop_aug_range: &#34;None&#34;,
ss_epoch: &#34;2&#34;,
ss_num_train_images: &#34;121361&#34;,
ss_color_aug: &#34;False&#34;,
ss_gradient_accumulation_steps: &#34;1&#34;,
ss_total_batch_size: &#34;100&#34;,
ss_prior_loss_weight: &#34;1.0&#34;,
ss_training_comment: &#34;None&#34;,
ss_network_dim: &#34;768&#34;,
ss_output_name: &#34;FusionaMEGA1tX&#34;,
ss_max_bucket_reso: &#34;1024&#34;,
ss_network_alpha: &#34;768.0&#34;,
ss_steps: &#34;2444&#34;,
ss_shuffle_caption: &#34;True&#34;,
ss_training_finished_at: &#34;1684158038.0763328&#34;,
ss_min_bucket_reso: &#34;256&#34;,
ss_noise_offset: &#34;0.09&#34;,
ss_enable_bucket: &#34;True&#34;,
ss_batch_size_per_device: &#34;20&#34;,
ss_max_train_steps: &#34;2444&#34;,
ss_network_module: &#34;lycoris.kohya&#34;,
## Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
### Preprocessing
&#34;{&#34;buckets&#34;: {&#34;0&#34;: {&#34;resolution&#34;: [192, 256], &#34;count&#34;: 1}, &#34;1&#34;: {&#34;resolution&#34;: [192, 320], &#34;count&#34;: 1}, &#34;2&#34;: {&#34;resolution&#34;: [256, 384], &#34;count&#34;: 1}, &#34;3&#34;: {&#34;resolution&#34;: [256, 512], &#34;count&#34;: 1}, &#34;4&#34;: {&#34;resolution&#34;: [384, 576], &#34;count&#34;: 2}, &#34;5&#34;: {&#34;resolution&#34;: [384, 640], &#34;count&#34;: 2}, &#34;6&#34;: {&#34;resolution&#34;: [384, 704], &#34;count&#34;: 1}, &#34;7&#34;: {&#34;resolution&#34;: [384, 1088], &#34;count&#34;: 15}, &#34;8&#34;: {&#34;resolution&#34;: [448, 448], &#34;count&#34;: 5}, &#34;9&#34;: {&#34;resolution&#34;: [448, 576], &#34;count&#34;: 1}, &#34;10&#34;: {&#34;resolution&#34;: [448, 640], &#34;count&#34;: 1}, &#34;11&#34;: {&#34;resolution&#34;: [448, 768], &#34;count&#34;: 1}, &#34;12&#34;: {&#34;resolution&#34;: [448, 832], &#34;count&#34;: 1}, &#34;13&#34;: {&#34;resolution&#34;: [448, 1088], &#34;count&#34;: 25}, &#34;14&#34;: {&#34;resolution&#34;: [448, 1216], &#34;count&#34;: 1}, &#34;15&#34;: {&#34;resolution&#34;: [512, 640], &#34;count&#34;: 2}, &#34;16&#34;: {&#34;resolution&#34;: [512, 768], &#34;count&#34;: 10}, &#34;17&#34;: {&#34;resolution&#34;: [512, 832], &#34;count&#34;: 3}, &#34;18&#34;: {&#34;resolution&#34;: [512, 896], &#34;count&#34;: 1525}, &#34;19&#34;: {&#34;resolution&#34;: [512, 960], &#34;count&#34;: 2}, &#34;20&#34;: {&#34;resolution&#34;: [512, 1024], &#34;count&#34;: 665}, &#34;21&#34;: {&#34;resolution&#34;: [512, 1088], &#34;count&#34;: 8}, &#34;22&#34;: {&#34;resolution&#34;: [576, 576], &#34;count&#34;: 5}, &#34;23&#34;: {&#34;resolution&#34;: [576, 768], &#34;count&#34;: 1}, &#34;24&#34;: {&#34;resolution&#34;: [576, 832], &#34;count&#34;: 667}, &#34;25&#34;: {&#34;resolution&#34;: [576, 896], &#34;count&#34;: 9601}, &#34;26&#34;: {&#34;resolution&#34;: [576, 960], &#34;count&#34;: 872}, &#34;27&#34;: {&#34;resolution&#34;: [576, 1024], &#34;count&#34;: 17}, &#34;28&#34;: {&#34;resolution&#34;: [640, 640], &#34;count&#34;: 3}, &#34;29&#34;: {&#34;resolution&#34;: [640, 768], &#34;count&#34;: 7}, &#34;30&#34;: {&#34;resolution&#34;: [640, 832], &#34;count&#34;: 608}, &#34;31&#34;: {&#34;resolution&#34;: [640, 896], &#34;count&#34;: 90}, &#34;32&#34;: {&#34;resolution&#34;: [704, 640], &#34;count&#34;: 1}, &#34;33&#34;: {&#34;resolution&#34;: [704, 704], &#34;count&#34;: 11}, &#34;34&#34;: {&#34;resolution&#34;: [704, 768], &#34;count&#34;: 1}, &#34;35&#34;: {&#34;resolution&#34;: [704, 832], &#34;count&#34;: 1}, &#34;36&#34;: {&#34;resolution&#34;: [768, 640], &#34;count&#34;: 225}, &#34;37&#34;: {&#34;resolution&#34;: [768, 704], &#34;count&#34;: 6}, &#34;38&#34;: {&#34;resolution&#34;: [768, 768], &#34;count&#34;: 74442}, &#34;39&#34;: {&#34;resolution&#34;: [832, 576], &#34;count&#34;: 23784}, &#34;40&#34;: {&#34;resolution&#34;: [832, 640], &#34;count&#34;: 554}, &#34;41&#34;: {&#34;resolution&#34;: [896, 512], &#34;count&#34;: 1235}, &#34;42&#34;: {&#34;resolution&#34;: [896, 576], &#34;count&#34;: 50}, &#34;43&#34;: {&#34;resolution&#34;: [896, 640], &#34;count&#34;: 88}, &#34;44&#34;: {&#34;resolution&#34;: [960, 512], &#34;count&#34;: 165}, &#34;45&#34;: {&#34;resolution&#34;: [960, 576], &#34;count&#34;: 5246}, &#34;46&#34;: {&#34;resolution&#34;: [1024, 448], &#34;count&#34;: 5}, &#34;47&#34;: {&#34;resolution&#34;: [1024, 512], &#34;count&#34;: 1187}, &#34;48&#34;: {&#34;resolution&#34;: [1024, 576], &#34;count&#34;: 40}, &#34;49&#34;: {&#34;resolution&#34;: [1088, 384], &#34;count&#34;: 70}, &#34;50&#34;: {&#34;resolution&#34;: [1088, 448], &#34;count&#34;: 36}, &#34;51&#34;: {&#34;resolution&#34;: [1088, 512], &#34;count&#34;: 3}, &#34;52&#34;: {&#34;resolution&#34;: [1216, 448], &#34;count&#34;: 36}, &#34;53&#34;: {&#34;resolution&#34;: [1344, 320], &#34;count&#34;: 29}, &#34;54&#34;: {&#34;resolution&#34;: [1536, 384], &#34;count&#34;: 1}}, &#34;mean_img_ar_error&#34;: 0.01693107810697896}&#34;,
### Speeds, Sizes, Times
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
ss_resolution: &#34;(768, 768)&#34;,
ss_v2: &#34;True&#34;,
ss_cache_latents: &#34;False&#34;,
ss_unet_lr: &#34;2e-07&#34;,
ss_num_reg_images: &#34;0&#34;,
ss_max_token_length: &#34;225&#34;,
ss_lr_scheduler: &#34;linear&#34;,
ss_reg_dataset_dirs: &#34;{}&#34;,
ss_lr_warmup_steps: &#34;303&#34;,
ss_num_batches_per_epoch: &#34;1222&#34;,
ss_lowram: &#34;False&#34;,
ss_multires_noise_iterations: &#34;None&#34;,
ss_optimizer: &#34;torch.optim.adamw.AdamW(weight_decay=0.01,betas=(0.9, 0.99))&#34;,
# Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
## Testing Data, Factors & Metrics
### Testing Data
<!-- This should link to a Data Card if possible. -->
More information needed
### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
More information needed
### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
More information needed
## Results
More information needed
# Model Examination
More information needed
# Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** 8xA100
- **Hours used:** 64
- **Cloud Provider:** CoreWeave
- **Compute Region:** US Main
- **Carbon Emitted:** 6.72
# Technical Specifications [optional]
## Model Architecture and Objective
Enhance your model&#39;s quality and sharpness using your own pre-trained Unet.
## Compute Infrastructure
More information needed
### Hardware
8xA100
### Software
Fully trained only with Kohya S &amp; Shih-Ying Yeh (Kohaku-BlueLeaf)
https://arxiv.org/abs/2108.06098
# Citation
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
More information needed
**APA:**
@misc{LyCORIS,
author = &#34;Shih-Ying Yeh (Kohaku-BlueLeaf), Yu-Guan Hsieh, Zhidong Gao&#34;,
title = &#34;LyCORIS - Lora beYond Conventional methods, Other Rank adaptation Implementations for Stable diffusion&#34;,
howpublished = &#34;\url{https://github.com/KohakuBlueleaf/LyCORIS}&#34;,
month = &#34;March&#34;,
year = &#34;2023&#34;
}
# Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
More information needed
# More Information [optional]
More information needed
# Model Card Authors [optional]
<!-- This section provides another layer of transparency and accountability. Whose views is this model card representing? How many voices were included in its construction? Etc. -->
idle stoev
# Model Card Contact
di@ffusion.ai
# How to Get Started with the Model
Use the code below to get started with the model.
<details>
<summary> Click to expand </summary>
For a1111
Install https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris
Download di.FFUSION.ai-tXe-FXAA to /models/Lycoris
Option1:
Insert <lyco:di.FFUSION.ai-tXe-FXAA:1.0> to prompt
No need to split Unet and Text Enc as its only TX encoder there.
You can go up to 2x weights
Option2: If you need it always ON (ex run a batch from txt file) then you can go to settings / Quicksettings list
add sd_lyco
restart and you should have a drop-down now 🤟 🥃
</details>