first

.gitattributes

@@ -32,3 +32,42 @@ 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
+control_picasso11_openpose.ckpt filter=lfs diff=lfs merge=lfs -text
+pose_1.png filter=lfs diff=lfs merge=lfs -text
+girl_1.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_17.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_30.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_0.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_21.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_2.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_34.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_6.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_12.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_14.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_19.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_23.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_33.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_5.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_13.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_16.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_18.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_24.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_27.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_29.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_31.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_10.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_15.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_1.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_20.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_32.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_3.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_7.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_11.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_25.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_26.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_28.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_4.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_9.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_22.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_35.png filter=lfs diff=lfs merge=lfs -text
+sample_poses/bone_8.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,29 @@
 ---
 license: openrail++
 ---
+
+# ControlNet for Stable Diffusion 2.1
+
+## Picasso Diffusion 1.1 (OpenPose)
+
+**Note: This model is a proof of concept. You can test the model with sample poses only. Give me more training data.**
+
+1. Install [Web UI](https://github.com/AUTOMATIC1111/stable-diffusion-webui).
+1. Install [ControlNet extention](https://github.com/Mikubill/sd-webui-controlnet).
+1. Download [Picasso Diffusion 1.1](https://huggingface.co/alfredplpl/picasso-diffusion-1-1/blob/main/v1-1.safetensors)
+1. Move it into the folder: models -> Stable-diffusion .
+1. Download the [model](control_picasso11_openpose.ckpt) and the [config](control_picasso11_openpose.yaml).
+1. Move them into the folder: extentions -> sd-webui-controlnet -> models.
+1. Use it in the web ui with the sample pose.
+
+## Examples
+
+![girl_1](girl_1.png)
+```
+anime, a girl
+Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 3038848424, Size: 768x768, Model hash: 5eb33121a0, Model: picasso_diffusion, ControlNet Enabled: True, ControlNet Module: none, ControlNet Model: control_picasso11_openpose [80042ed5], ControlNet Weight: 1, ControlNet Guidance Strength: 1
+```
+![pose_1](pose_1.png)
+
+
+

control_picasso11_openpose.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4845d4d1faea6d754beb110514f8e615d5ca0e6b4812eeb3b3bd475139d6ce75
+size 3336262761

control_picasso11_openpose.yaml

@@ -0,0 +1,85 @@
+model:
+  target: cldm.cldm.ControlLDM
+  params:
+    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"
+    control_key: "hint"
+    image_size: 64
+    channels: 4
+    cond_stage_trainable: false
+    conditioning_key: crossattn
+    monitor: val/loss_simple_ema
+    scale_factor: 0.18215
+    use_ema: False
+    only_mid_control: False
+
+    control_stage_config:
+      target: cldm.cldm.ControlNet
+      params:
+        use_checkpoint: True
+        image_size: 32 # unused
+        in_channels: 4
+        hint_channels: 3
+        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
+
+    unet_config:
+      target: cldm.cldm.ControlledUnetModel
+      params:
+        use_checkpoint: 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"