Upload 4 files

.gitattributes

@@ -25,7 +25,6 @@
 *.safetensors filter=lfs diff=lfs merge=lfs -text
 saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.tar.* filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
 *.tflite filter=lfs diff=lfs merge=lfs -text
 *.tgz filter=lfs diff=lfs merge=lfs -text
 *.wasm filter=lfs diff=lfs merge=lfs -text
@@ -33,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+demo.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,124 @@
+---
+license: other
+---
+
+# Model Card for Model ID
+
+hololive-diffusion, a stable diffusion 2.1 768x768 model.
+
+Trained on ~73k hololive fanart images
+
+![](https://huggingface.co/double-negative/hololive-diffusion/resolve/main/demo.png)
+
+## License
+This model is a fine-tune of Waifu diffusion 1.5.
+
+hololive-diffusion is released under the Fair AI Public License 1.0-SD (https://freedevproject.org/faipl-1.0-sd/). If any derivative of this model is made, please share your changes accordingly. Special thanks to ronsor/undeleted (https://undeleted.ronsor.com/) for help with the license.
+
+When creating characters owned and copyrighted by cover corp, you may not use this model or any of its outputs in a way that is counter to Hololive's fanart guidelines: https://hololivepro.com/en/terms/
+
+
+
+## Prompting
+
+The start of the prompt should have all the following:
+
+"hololive, anime, waifu"
+
+Add one of danbooru's content ratings
+* general
+* sensitive
+* questionable
+* explicit
+ 
+and the character name
+
+### names
+The model was trained using the following names. These aren't necessarily hololive vtubers, and the characters with less popular artworks will be harder to get correct.
+
+* a-chan (hololive)
+* airani iofifteen
+* akai haato
+* aki rosenthal
+* allegro (hoshimachi suisei)
+* amane kanata
+* anemachi
+* anya melfissa
+* artia
+* avatar (holoearth)
+* ayunda risu
+* azki (hololive)
+* ceres fauna
+* civia
+* coco kaine
+* deadbeat (calliope mori)
+* doris (hololive)
+* enma-chan
+* gawr gura
+* haaton (akai haato)
+* hakos baelz
+* hakui koyori
+* harusaki nodoka
+* himemori luna
+* hoshimachi suisei
+* houshou marine
+* inugami korone
+* irys (hololive)
+* j-chad
+* kaela kovalskia
+* kagura nana
+* kazama iroha
+* kiryu coco
+* kobo kanaeru
+* kureiji ollie
+* kurokami fubuki
+* la+ darknesss
+* mama lillie
+* mano aloe
+* matsurisu
+* minato aqua
+* moku seiko
+* momosuzu nene
+* moona hoshinova
+* mori calliope
+* murasaki shion
+* nakiri ayame
+* nanashi mumei
+* natsuiro matsuri
+* nekomata okayu
+* ninomae ina'nis
+* nousagi (usada pekora)
+* omaru polka
+* omega alpha
+* ookami mio
+* oozora subaru
+* ouro kronii
+* pavolia reine
+* pekomama
+* roboco-san
+* sakamata chloe
+* sakura miko
+* shigure ui (vtuber)
+* shirakami fubuki
+* shiranui flare
+* shirogane noel
+* shishiro botan
+* spade echo
+* sukonbu (shirakami fubuki)
+* takanashi kiara
+* takane lui
+* takodachi (ninomae ina'nis)
+* tokino sora
+* tokoyami towa
+* tsukumo sana
+* tsunomaki watame
+* uruha rushia
+* usada pekora
+* vestia zeta
+* wahtcher (ninomae ina'nis)
+* watoto (mythbreakers)
+* watson amelia
+* yozora mel
+* yukihana lamy
+* yuul b alwright (mythbreakers)
+* yuzuki choco

hololive_wd15.yaml

@@ -0,0 +1,68 @@
+model:
+  base_learning_rate: 1.0e-4
+  target: ldm.models.diffusion.ddpm.LatentDiffusion
+  params:
+    parameterization: "v"
+    linear_start: 0.00085
+    linear_end: 0.0120
+    num_timesteps_cond: 1
+    log_every_t: 200
+    timesteps: 1000
+    first_stage_key: "jpg"
+    cond_stage_key: "txt"
+    image_size: 64
+    channels: 4
+    cond_stage_trainable: false
+    conditioning_key: crossattn
+    monitor: val/loss_simple_ema
+    scale_factor: 0.18215
+    use_ema: False # we set this to false because this is an inference only config
+
+    unet_config:
+      target: ldm.modules.diffusionmodules.openaimodel.UNetModel
+      params:
+        use_checkpoint: True
+        use_fp16: True
+        image_size: 32 # unused
+        in_channels: 4
+        out_channels: 4
+        model_channels: 320
+        attention_resolutions: [ 4, 2, 1 ]
+        num_res_blocks: 2
+        channel_mult: [ 1, 2, 4, 4 ]
+        num_head_channels: 64 # need to fix for flash-attn
+        use_spatial_transformer: True
+        use_linear_in_transformer: True
+        transformer_depth: 1
+        context_dim: 1024
+        legacy: False
+
+    first_stage_config:
+      target: ldm.models.autoencoder.AutoencoderKL
+      params:
+        embed_dim: 4
+        monitor: val/rec_loss
+        ddconfig:
+          #attn_type: "vanilla-xformers"
+          double_z: true
+          z_channels: 4
+          resolution: 256
+          in_channels: 3
+          out_ch: 3
+          ch: 128
+          ch_mult:
+          - 1
+          - 2
+          - 4
+          - 4
+          num_res_blocks: 2
+          attn_resolutions: []
+          dropout: 0.0
+        lossconfig:
+          target: torch.nn.Identity
+
+    cond_stage_config:
+      target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
+      params:
+        freeze: True
+        layer: "penultimate"