diff --git a/RealCustom/configs/realcustom_sigdino_highres - back.json b/RealCustom/configs/realcustom_sigdino_highres - back.json new file mode 100644 index 0000000..735534c --- /dev/null +++ b/RealCustom/configs/realcustom_sigdino_highres - back.json @@ -0,0 +1,119 @@ +{ + "act_fn": "silu", + "addition_embed_type": "text_time", + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": 256, + "attention_head_dim": [ + 5, + 10, + 20 + ], + "block_out_channels": [ + 320, + 640, + 1280 + ], + "center_input_sample": false, + "class_embed_type": null, + "class_embeddings_concat": false, + "conv_in_kernel": 3, + "conv_out_kernel": 3, + "cross_attention_dim": 2048, + "cross_attention_norm": null, + "down_block_types": [ + "DownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D" + ], + "downsample_padding": 1, + "dual_cross_attention": false, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "in_channels": 4, + "layers_per_block": 2, + "mid_block_only_cross_attention": null, + "mid_block_scale_factor": 1, + "mid_block_type": "UNetMidBlock2DCrossAttn", + "norm_eps": 1e-05, + "norm_num_groups": 32, + "num_attention_heads": null, + "num_class_embeds": null, + "only_cross_attention": false, + "out_channels": 4, + "projection_class_embeddings_input_dim": 2816, + "resnet_out_scale_factor": 1.0, + "resnet_skip_time_act": false, + "resnet_time_scale_shift": "default", + "sample_size": 128, + "time_cond_proj_dim": null, + "time_embedding_act_fn": null, + "time_embedding_dim": null, + "time_embedding_type": "positional", + "timestep_post_act": null, + "transformer_layers_per_block": [ + 1, + 2, + 10 + ], + "up_block_types": [ + "CrossAttnUpBlock2D", + "CrossAttnUpBlock2D", + "UpBlock2D" + ], + "upcast_attention": false, + "use_linear_projection": true, + "image_ref_processor_config": { + "target": "utils.image_ref_processor.default.NaiveResizeProcessor", + "params": { + "target_image_size": 768, + "resize_mode": "resize", + "crop_min_ratio": 1.0, + "crop_max_ratio": 1.0, + "image_dropout": 0.1 + } + }, + "image_ref_processor_input_keys": [ + "image_ref" + ], + "vision_model_config": { + "vision_model_config": { + "target": "models.image_encoder_siglipdino_shallowdeep.ShallowDeepPatchfySiglipDinoEncoder", + "params": { + "siglip_config": { + "backbone_name_or_path": "vit_so400m_patch14_siglip_384", + "image_resize_strategy": "resize-naive", + "default_image_size": 384, + "feature_index": [ + 25 + ] + }, + "dino_config": { + "backbone_name_or_path": "vit_large_patch14_reg4_dinov2.lvd142m", + "image_resize_strategy": "resize-naive", + "default_image_size": 384, + "feature_index": [ + 22 + ] + }, + "patchfy_scale": 2, + "default_image_size": 384 + } + } + }, + "image_prompt_settings": { + "vision_projection_type": "custom", + "vision_projection_config": { + "target": "models.projectors.ProjectorHighResMinAttn", + "params": { + "vision_dim": 2176, + "out_dim": 2048, + "dim_head": 64, + "adaptive_scale": false + } + }, + "image_prompt_mode": "naive", + "cross_attention_id": 70 + } +} \ No newline at end of file diff --git a/RealCustom/configs/realcustom_sigdino_highres.json b/RealCustom/configs/realcustom_sigdino_highres.json new file mode 100644 index 0000000..d356731 --- /dev/null +++ b/RealCustom/configs/realcustom_sigdino_highres.json @@ -0,0 +1,122 @@ +{ + "act_fn": "silu", + "addition_embed_type": "text_time", + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": 256, + "attention_head_dim": [ + 5, + 10, + 20 + ], + "block_out_channels": [ + 320, + 640, + 1280 + ], + "center_input_sample": false, + "class_embed_type": null, + "class_embeddings_concat": false, + "conv_in_kernel": 3, + "conv_out_kernel": 3, + "cross_attention_dim": 2048, + "cross_attention_norm": null, + "down_block_types": [ + "DownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D" + ], + "downsample_padding": 1, + "dual_cross_attention": false, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "in_channels": 4, + "layers_per_block": 2, + "mid_block_only_cross_attention": null, + "mid_block_scale_factor": 1, + "mid_block_type": "UNetMidBlock2DCrossAttn", + "norm_eps": 1e-05, + "norm_num_groups": 32, + "num_attention_heads": null, + "num_class_embeds": null, + "only_cross_attention": false, + "out_channels": 4, + "projection_class_embeddings_input_dim": 2816, + "resnet_out_scale_factor": 1.0, + "resnet_skip_time_act": false, + "resnet_time_scale_shift": "default", + "sample_size": 128, + "time_cond_proj_dim": null, + "time_embedding_act_fn": null, + "time_embedding_dim": null, + "time_embedding_type": "positional", + "timestep_post_act": null, + "transformer_layers_per_block": [ + 1, + 2, + 10 + ], + "up_block_types": [ + "CrossAttnUpBlock2D", + "CrossAttnUpBlock2D", + "UpBlock2D" + ], + "upcast_attention": false, + "use_linear_projection": true, + "image_ref_processor_config": { + "target": ".utils.image_ref_processor.default.NaiveResizeProcessor", + "params": { + "target_image_size": 768, + "resize_mode": "resize", + "crop_min_ratio": 1.0, + "crop_max_ratio": 1.0, + "image_dropout": 0.1 + } + }, + "image_ref_processor_input_keys": [ + "image_ref" + ], + "vision_model_config": { + "vision_model_config": { + "target": "ComfyUI.custom_nodes.ComfyUI_StoryDiffusion.RealCustom.models.image_encoder_siglipdino_shallowdeep.ShallowDeepPatchfySiglipDinoEncoder", + "params": { + "siglip_config": { + "backbone_name_or_path": "vit_so400m_patch14_siglip_384", + "image_resize_strategy": "resize-naive", + "default_image_size": 384, + "feature_index": [ + 25 + ] + + }, + "dino_config": { + "backbone_name_or_path": "vit_large_patch14_reg4_dinov2.lvd142m", + "image_resize_strategy": "resize-naive", + "default_image_size": 384, + "feature_index": [ + 22 + ] + }, + "patchfy_scale": 2, + "default_image_size": 384, + "siglip_path": " ", + "dino_path": " " + } + } + }, + "image_prompt_settings": { + "vision_projection_type": "custom", + "vision_projection_config": { + "target": "ComfyUI.custom_nodes.ComfyUI_StoryDiffusion.RealCustom.models.projectors.ProjectorHighResMinAttn", + "params": { + "vision_dim": 2176, + "out_dim": 2048, + "dim_head": 64, + "adaptive_scale": false + } + }, + "image_prompt_mode": "naive", + "cross_attention_id": 70 + } +} \ No newline at end of file diff --git a/RealCustom/configs/realcustom_sigdino_highres_shallow- back.json b/RealCustom/configs/realcustom_sigdino_highres_shallow- back.json new file mode 100644 index 0000000..3f0e100 --- /dev/null +++ b/RealCustom/configs/realcustom_sigdino_highres_shallow- back.json @@ -0,0 +1,114 @@ +{ + "act_fn": "silu", + "addition_embed_type": "text_time", + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": 256, + "attention_head_dim": [ + 5, + 10, + 20 + ], + "block_out_channels": [ + 320, + 640, + 1280 + ], + "center_input_sample": false, + "class_embed_type": null, + "class_embeddings_concat": false, + "conv_in_kernel": 3, + "conv_out_kernel": 3, + "cross_attention_dim": 2048, + "cross_attention_norm": null, + "down_block_types": [ + "DownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D" + ], + "downsample_padding": 1, + "dual_cross_attention": false, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "in_channels": 4, + "layers_per_block": 2, + "mid_block_only_cross_attention": null, + "mid_block_scale_factor": 1, + "mid_block_type": "UNetMidBlock2DCrossAttn", + "norm_eps": 1e-05, + "norm_num_groups": 32, + "num_attention_heads": null, + "num_class_embeds": null, + "only_cross_attention": false, + "out_channels": 4, + "projection_class_embeddings_input_dim": 2816, + "resnet_out_scale_factor": 1.0, + "resnet_skip_time_act": false, + "resnet_time_scale_shift": "default", + "sample_size": 128, + "time_cond_proj_dim": null, + "time_embedding_act_fn": null, + "time_embedding_dim": null, + "time_embedding_type": "positional", + "timestep_post_act": null, + "transformer_layers_per_block": [ + 1, + 2, + 10 + ], + "up_block_types": [ + "CrossAttnUpBlock2D", + "CrossAttnUpBlock2D", + "UpBlock2D" + ], + "upcast_attention": false, + "use_linear_projection": true, + + "image_ref_processor_input_keys": ["image_ref"], + "vision_model_config": { + "vision_model_config": { + "target": "models.image_encoder_siglipdino_shallowdeep.ShallowDeepPatchfySiglipDinoEncoder_v2", + "params": { + "siglip_config": { + "backbone_name_or_path": "vit_so400m_patch14_siglip_384", + "image_resize_strategy": "resize-naive", + "default_image_size": 384, + "feature_index": [ + 7, + 13, + 19, + 25 + ] + }, + "dino_config": { + "backbone_name_or_path": "vit_large_patch14_reg4_dinov2.lvd142m", + "image_resize_strategy": "resize-naive", + "default_image_size": 384, + "feature_index": [ + 4, + 10, + 16, + 22 + ] + }, + "patchfy_scale": 2, + "default_image_size": 384 + } + } + }, + + "image_prompt_settings": { + "vision_projection_type": "custom", + "vision_projection_config": { + "target": "models.projectors.ProjectorHighResShallowMinAttnV1", + "params": { + "vision_dim": 2176, + "out_dim": 2048, + "dim_head": 64 + } + }, + + "image_prompt_mode": "naive" + } +} \ No newline at end of file diff --git a/RealCustom/configs/realcustom_sigdino_highres_shallow.json b/RealCustom/configs/realcustom_sigdino_highres_shallow.json new file mode 100644 index 0000000..b99f79f --- /dev/null +++ b/RealCustom/configs/realcustom_sigdino_highres_shallow.json @@ -0,0 +1,116 @@ +{ + "act_fn": "silu", + "addition_embed_type": "text_time", + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": 256, + "attention_head_dim": [ + 5, + 10, + 20 + ], + "block_out_channels": [ + 320, + 640, + 1280 + ], + "center_input_sample": false, + "class_embed_type": null, + "class_embeddings_concat": false, + "conv_in_kernel": 3, + "conv_out_kernel": 3, + "cross_attention_dim": 2048, + "cross_attention_norm": null, + "down_block_types": [ + "DownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D" + ], + "downsample_padding": 1, + "dual_cross_attention": false, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "in_channels": 4, + "layers_per_block": 2, + "mid_block_only_cross_attention": null, + "mid_block_scale_factor": 1, + "mid_block_type": "UNetMidBlock2DCrossAttn", + "norm_eps": 1e-05, + "norm_num_groups": 32, + "num_attention_heads": null, + "num_class_embeds": null, + "only_cross_attention": false, + "out_channels": 4, + "projection_class_embeddings_input_dim": 2816, + "resnet_out_scale_factor": 1.0, + "resnet_skip_time_act": false, + "resnet_time_scale_shift": "default", + "sample_size": 128, + "time_cond_proj_dim": null, + "time_embedding_act_fn": null, + "time_embedding_dim": null, + "time_embedding_type": "positional", + "timestep_post_act": null, + "transformer_layers_per_block": [ + 1, + 2, + 10 + ], + "up_block_types": [ + "CrossAttnUpBlock2D", + "CrossAttnUpBlock2D", + "UpBlock2D" + ], + "upcast_attention": false, + "use_linear_projection": true, + + "image_ref_processor_input_keys": ["image_ref"], + "vision_model_config": { + "vision_model_config": { + "target": "ComfyUI.custom_nodes.ComfyUI_StoryDiffusion.RealCustom.models.image_encoder_siglipdino_shallowdeep.ShallowDeepPatchfySiglipDinoEncoder_v2", + "params": { + "siglip_config": { + "backbone_name_or_path": "vit_so400m_patch14_siglip_384", + "image_resize_strategy": "resize-naive", + "default_image_size": 384, + "feature_index": [ + 7, + 13, + 19, + 25 + ] + }, + "dino_config": { + "backbone_name_or_path": "vit_large_patch14_reg4_dinov2.lvd142m", + "image_resize_strategy": "resize-naive", + "default_image_size": 384, + "feature_index": [ + 4, + 10, + 16, + 22 + ] + }, + "patchfy_scale": 2, + "default_image_size": 384, + "siglip_path": " ", + "dino_path": " " + } + } + }, + + "image_prompt_settings": { + "vision_projection_type": "custom", + "vision_projection_config": { + "target": "ComfyUI.custom_nodes.ComfyUI_StoryDiffusion.RealCustom.models.projectors.ProjectorHighResShallowMinAttnV1", + "params": { + "vision_dim": 2176, + "out_dim": 2048, + "dim_head": 64 + } + }, + + "image_prompt_mode": "naive" + } +} \ No newline at end of file