Update README.md

README.md

@@ -33,13 +33,13 @@ This model is a latent diffusion model finetuned on [Waifu Diffusion v1.4 epoch
 You can use Danbooru tags to generate images.
 
 
-## Model links
+## Downloads
 
-- [**pvc-v2.safetensors**](https://huggingface.co/p1atdev/pvc/resolve/main/pvc-v2.safetensors)
-- [**pvc-v2.ckpt**](https://huggingface.co/p1atdev/pvc/resolve/main/pvc-v2.ckpt)
-- [pvc-v2.yaml](https://huggingface.co/p1atdev/pvc/blob/main/pvc-v2.yaml) (needed if you want to use the model in AUTOMATIC1111's Web UI)
-- [pvc-v2-lora.pt](https://huggingface.co/p1atdev/pvc/resolve/main/pvc-v2-lora.pt) (maybe only works with kohya's [sd-scripts](https://huggingface.co/p1atdev/pvc/resolve/main/pvc-v1-lora.pt) or [webui extension](https://github.com/kohya-ss/sd-webui-additional-networks))
-- [Waifu Diffusion 1.4 VAE](https://huggingface.co/hakurei/waifu-diffusion-v1-4/blob/main/vae/kl-f8-anime2.ckpt) (Recommended)
+<div class="flex gap-4 my-4">
+<a class="btn" href="https://huggingface.co/p1atdev/pvc/resolve/9771e2970e9574efe89910801861234a5f276aa2/pvc-v2.ckpt">Download model (ckpt)</a>
+<a class="btn" href="https://huggingface.co/p1atdev/pvc/resolve/9771e2970e9574efe89910801861234a5f276aa2/pvc-v2.safetensors">Download model (safetensors)</a>
+<a class="btn" href="https://huggingface.co/p1atdev/pvc/resolve/9771e2970e9574efe89910801861234a5f276aa2/pvc-v2-lora.pt">Download LoRA (pt)</a>
+</div>
 
 ## Prompt guide
 
@@ -52,55 +52,88 @@ nsfw, worst quality, low quality, medium quality, deleted, lowres, bad anatomy,
 
 ## Samples
 
-![v2 sample1](https://huggingface.co/p1atdev/pvc/resolve/main/samples/v2-sample1.png)
-
-
-```
+<div class="grid grid-cols-3 gap-4">
+<div class="rounded-lg overflow-hidden gap-0">
+<img class="m-0 p-0 w-full" src="https://huggingface.co/p1atdev/pvc/resolve/main/samples/v2-sample1.png"/>
+<div class="w-full h-full from-gray-100-to-white bg-gradient-to-t">
+<p class="px-4 pt-4 my-0">
 masterpiece, best quality, 1girl, green hair, sweater, beanie, turtleneck, looking at viewer, night,
 Negative prompt: nsfw, worst quality, low quality, medium quality, deleted, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts, signature, watermark, username, blurry
 Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7
-```
-
----
-
-![v2 sample2](https://huggingface.co/p1atdev/pvc/resolve/main/samples/v2-sample2.png)
-
-```
+</p>
+</div>
+</div>
+
+<div class="rounded-lg overflow-hidden bg-clip-border">
+<img class="m-0 p-0 w-full" src="https://huggingface.co/p1atdev/pvc/resolve/main/samples/v2-sample2.png"/>
+<div class="w-full h-full from-gray-100-to-white bg-gradient-to-t">
+<p class="px-4 pt-4 my-0">
 masterpiece, best quality, high quality, 1girl, bob cut, cape, belt, gloves, looking at viewer, ultra detailed
 Negative prompt: nsfw, worst quality, low quality, medium quality, deleted, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts, signature, watermark, username, blurry
 Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7
-```
-
----
-
-![v2 sample3](https://huggingface.co/p1atdev/pvc/resolve/main/samples/v2-sample3.png)
-
-```
+</p>
+</div>
+</div>
+
+<div class="rounded-lg overflow-hidden bg-clip-border">
+<img class="m-0 p-0 w-full" src="https://huggingface.co/p1atdev/pvc/resolve/main/samples/v2-sample3.png"/>
+<div class="w-full h-full from-gray-100-to-white bg-gradient-to-t">
+<p class="px-4 pt-4 my-0">
 masterpiece, best quality, 1girl, annoyed, black hair, floating hair, gothic lolita, scythe, looking at viewer,
 Negative prompt: nsfw, worst quality, low quality, medium quality, deleted, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts, signature, watermark, username, blurry
 Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7
-```
+</p>
+</div>
+</div>
+  
+</div>
+
+## License
+
+This version (v2) model is open access and available to all, with a CreativeML OpenRAIL-M license further specifying rights and usage.
+The CreativeML OpenRAIL License specifies: 
+
+1. You can't use the model to deliberately produce nor share illegal or harmful outputs or content 
+2. The authors claims no rights on the outputs you generate, you are free to use them and are accountable for their use which must not go against the provisions set in the license
+3. You may re-distribute the weights and use the model commercially and/or as a service. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the CreativeML OpenRAIL-M to all your users (please read the license entirely and carefully)
+[Please read the full license here](https://huggingface.co/spaces/CompVis/stable-diffusion-license)
+
 
 ## Training information
 
-```
---train_batch_size 1 --resolution 768,768 --learning_rate 1e-4 --max_train_steps 21100 --use_8bit_adam --xformers --mixed_precision fp16 --save_every_n_epochs 1 --network_module networks.lora --v2 --enable_bucket --max_token_length 225 --network_dim 32 --text_encoder_lr 5e-5 --lr_scheduler cosine_with_restarts --lr_warmup_steps 200
-```
+- Training script: kohya_ss/sd-scripts
+- Number of images: 2110
+- Traning steps: 21100
+- Resolution: 768,768
+- Learning rate:
+  - U-Net: 1e-4
+  - Text encoder: 5e-5
+- Optimizer: AdamW8bit
+- Max token length: 225
+- Mixed precision: fp16
+- Network dim: 16
+- Xformers: enabled
+- Bucketing: enabled
+
+# Old versions
+
+<details>
+<summary class="text-xl">
+PVC v1 (based on WDv1.4 e2)
+</summary>
+
+<div class="flex gap-4 my-4">
+<a class="btn" href="https://huggingface.co/p1atdev/pvc/resolve/9771e2970e9574efe89910801861234a5f276aa2/pvc-v1.ckpt">Download model (ckpt)</a>
+<a class="btn" href="https://huggingface.co/p1atdev/pvc/resolve/9771e2970e9574efe89910801861234a5f276aa2/pvc-v1.safetensors">Download model (safetensors)</a>
+<a class="btn" href="https://huggingface.co/p1atdev/pvc/resolve/9771e2970e9574efe89910801861234a5f276aa2/pvc-v1-lora.pt">Download LoRA (pt)</a>
+</div>
 
-# PVC v1
 
 ![eyecatch](https://huggingface.co/p1atdev/pvc/resolve/main/samples/miku.png)
 
 This model is a latent diffusion model finetuned on [Waifu Diffusion v1.4 epoch 2](https://huggingface.co/hakurei/waifu-diffusion-v1-4) with [PVC figure images](https://huggingface.co/datasets/p1atdev/pvc) using LoRA method.
 You can use Danbooru tags to generate images.
 
-## Model links
-
-- [**pvc-v1.safetensors**](https://huggingface.co/p1atdev/pvc/resolve/main/pvc-v1.safetensors)
-- [**pvc-v1.ckpt**](https://huggingface.co/p1atdev/pvc/resolve/main/pvc-v1.ckpt)
-- [pvc-v1.yaml](https://huggingface.co/p1atdev/pvc/blob/main/pvc-v1.yaml) (needed if you want to use the model in AUTOMATIC1111's Web UI)
-- [pvc-v1-lora.pt](https://huggingface.co/p1atdev/pvc/resolve/main/pvc-v1-lora.pt) (maybe only works with kohya's [sd-scripts](https://huggingface.co/p1atdev/pvc/resolve/main/pvc-v1-lora.pt) or [webui extension](https://github.com/kohya-ss/sd-webui-additional-networks))
-
 ## Samples
 
 ![sample1](https://huggingface.co/p1atdev/pvc/resolve/main/samples/sample1.png)
@@ -133,8 +166,31 @@ Negative prompt: worst quality, low quality, medium quality, deleted, lowres, ba
 Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 7
 ```
 
+## License
+
+This version (v1) model is open access and available to all, with a CreativeML OpenRAIL-M license further specifying rights and usage.
+The CreativeML OpenRAIL License specifies: 
+
+1. You can't use the model to deliberately produce nor share illegal or harmful outputs or content 
+2. The authors claims no rights on the outputs you generate, you are free to use them and are accountable for their use which must not go against the provisions set in the license
+3. You may re-distribute the weights and use the model commercially and/or as a service. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the CreativeML OpenRAIL-M to all your users (please read the license entirely and carefully)
+[Please read the full license here](https://huggingface.co/spaces/CompVis/stable-diffusion-license)
+
+
 ## Training information
 
-```
---train_batch_size 1 --resolution 768,768 --learning_rate 1e-4 --max_train_steps 17740 --use_8bit_adam --xformers --mixed_precision fp16 --save_every_n_epochs 1 --network_module networks.lora --v2 --enable_bucket --max_token_length 225 --network_dim 16 --text_encoder_lr 5e-5
-```
+- Training script: kohya_ss/sd-scripts
+- Number of images: 887
+- Traning steps: 17740
+- Resolution: 768,768
+- Learning rate:
+  - U-Net: 1e-4
+  - Text encoder: 5e-5
+- Optimizer: AdamW8bit
+- Max token length: 225
+- Mixed precision: fp16
+- Network dim: 16
+- Xformers: enabled
+- Bucketing: enabled
+
+</details>