diff --git a/.gitattributes b/.gitattributes index a6344aac8c09253b3b630fb776ae94478aa0275b..eea1d8e4da6083525243bbdb899f6bae1ce3d57c 100644 --- a/.gitattributes +++ b/.gitattributes @@ -33,3 +33,5 @@ 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 +images_0.png filter=lfs diff=lfs merge=lfs -text +images_1.png filter=lfs diff=lfs merge=lfs -text diff --git a/README.md b/README.md new file mode 100644 index 0000000000000000000000000000000000000000..1dfdcf7870fd59be0e8a0dfda789af31d65eab9d --- /dev/null +++ b/README.md @@ -0,0 +1,45 @@ +--- +base_model: runwayml/stable-diffusion-v1-5 +library_name: diffusers +license: creativeml-openrail-m +tags: +- stable-diffusion +- stable-diffusion-diffusers +- text-to-image +- diffusers +- controlnet +- diffusers-training +inference: true +--- + + + + +# controlnet-licb-k/model0114 + +These are controlnet weights trained on runwayml/stable-diffusion-v1-5 with new type of conditioning. +You can find some example images below. + +prompt: thermal infrared image of rock surface with crack +![images_0)](./images_0.png) +prompt: thermal infrared image of rock surface with crack +![images_1)](./images_1.png) + + + +## Intended uses & limitations + +#### How to use + +```python +# TODO: add an example code snippet for running this diffusion pipeline +``` + +#### Limitations and bias + +[TODO: provide examples of latent issues and potential remediations] + +## Training details + +[TODO: describe the data used to train the model] \ No newline at end of file diff --git a/checkpoint-1000/controlnet/config.json b/checkpoint-1000/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-1000/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-1000/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-1000/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5578fdbeaa997c86919fd7561fd2eaa78097d0e8 --- /dev/null +++ b/checkpoint-1000/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3a52c694e1bfcc45903f82d21222939fc83b7982b3bc1c0d8639b66334b11817 +size 1445157120 diff --git a/checkpoint-1000/optimizer.bin b/checkpoint-1000/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..09a9fd6225d3c33ea625d6c9c34ca78b6b1181de --- /dev/null +++ b/checkpoint-1000/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ba6b56cd0b12d3b9eca987403a2dfd0500af7b02557dc4222f4f8635d24aaef +size 2890518478 diff --git a/checkpoint-1000/random_states_0.pkl b/checkpoint-1000/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..97322e62e871d0aa097b6de42525abfccbc003a5 --- /dev/null +++ b/checkpoint-1000/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d96d564433aa619318f7d14d15a5ed89505105af57004629d2ee43bc535d5b07 +size 14604 diff --git a/checkpoint-1000/scheduler.bin b/checkpoint-1000/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..3436b9f7a51f9dda3e9355776bd247d4b50ac61d --- /dev/null +++ b/checkpoint-1000/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6ed9b0f338d07743081c4e6e899c96d5ed17b6e9266b82e7b10064abf9e0f341 +size 1000 diff --git a/checkpoint-1500/controlnet/config.json b/checkpoint-1500/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-1500/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-1500/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-1500/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..02d641fbd8a253d98432de5ad63b70b39d09fef8 --- /dev/null +++ b/checkpoint-1500/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8aba9000c9cfe2897d8976bea76d9a703efbbe9164eaaf801ea7c194351ad1ce +size 1445157120 diff --git a/checkpoint-1500/optimizer.bin b/checkpoint-1500/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..9865e1e3e932ace2455afc702961329a12c9ff1f --- /dev/null +++ b/checkpoint-1500/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:759af96658f405f1f6c643b6a6499f91d95e566ab4daa71075838b36ec21e584 +size 2890518478 diff --git a/checkpoint-1500/random_states_0.pkl b/checkpoint-1500/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..b272e629dfa08daedd0ff6a4c2ecb8b08f5c9055 --- /dev/null +++ b/checkpoint-1500/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b182cae06e443e4578bb56ea80ab117111ca36fe037891975eb30d8914efcda2 +size 14604 diff --git a/checkpoint-1500/scheduler.bin b/checkpoint-1500/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..bbb1a2f26f112b1ae0eb33ad904a3cbccdc58931 --- /dev/null +++ b/checkpoint-1500/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:af8bad064664d2f3a8a575280b309f6ec56c271a06b732f04fd44b854ccb7fd2 +size 1000 diff --git a/checkpoint-2000/controlnet/config.json b/checkpoint-2000/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-2000/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-2000/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-2000/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..96232361a6ca6b8f4fed5e375e32dac16eb256e7 --- /dev/null +++ b/checkpoint-2000/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:795a2f6aa47e13ee9fffdec0ea74f2bf69c513f944917ab1a9d22e09821d241f +size 1445157120 diff --git a/checkpoint-2000/optimizer.bin b/checkpoint-2000/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..6effc348c1fb5c9a5a316a4a1bc2ff68fea5ffec --- /dev/null +++ b/checkpoint-2000/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b9925a7272981c40b3d5eba857b4b19f78cdd4ad2706607570ea05f9f7d8f8a3 +size 2890518478 diff --git a/checkpoint-2000/random_states_0.pkl b/checkpoint-2000/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..26b350882ee0b6dd87c4ab74acbd4dad16d9527a --- /dev/null +++ b/checkpoint-2000/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1a51568240e3efc6c41c24962aa8aa81c08ce9127fa705f33179a87f3fa8e37d +size 14604 diff --git a/checkpoint-2000/scheduler.bin b/checkpoint-2000/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..72b3b0e5a295f0068de95431ca9c4ac4fd598301 --- /dev/null +++ b/checkpoint-2000/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cbbc673b298785e9b2bc10ee6cfc3d2ffbd0ed2cb99da01556bd134cc6ac2f2f +size 1000 diff --git a/checkpoint-2500/controlnet/config.json b/checkpoint-2500/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-2500/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-2500/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-2500/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5a811334d8f800c0cdfa3fa514c8f59029dd2053 --- /dev/null +++ b/checkpoint-2500/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c43c3b0cb45fe11ce89e09c54aa61fe0a24aee7ea105e3be5a5e55d016f75fb +size 1445157120 diff --git a/checkpoint-2500/optimizer.bin b/checkpoint-2500/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..cef11bae4bbdd6cfe796c65231698d714d547a6c --- /dev/null +++ b/checkpoint-2500/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:66e35ac64aca241f276e239cfce17a82ced09a7a32776c80766f6262e02c2da0 +size 2890518478 diff --git a/checkpoint-2500/random_states_0.pkl b/checkpoint-2500/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..ca53014a194d6ded5018fc6e0f39c207f091e3d8 --- /dev/null +++ b/checkpoint-2500/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d1f63c5217ac2772f4585fc1088e6b2fab4c91e649cc134e69db8b090656a9a6 +size 14604 diff --git a/checkpoint-2500/scheduler.bin b/checkpoint-2500/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..e5bdbc6bf81a2d7333570a3df444b6d1eecfb6ac --- /dev/null +++ b/checkpoint-2500/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6353762aaa1c52881aa1d6234c85280b3c734c301ea46eb27bbf7f54a6daf124 +size 1000 diff --git a/checkpoint-3000/controlnet/config.json b/checkpoint-3000/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-3000/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-3000/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-3000/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fc3f7ae7c9a800e8aa66eeeea221b2acebb06ed1 --- /dev/null +++ b/checkpoint-3000/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6e8a97999d24b22fbad9da9ab9c2553eda376c87a7c8c5c35b1282f4b0d315b6 +size 1445157120 diff --git a/checkpoint-3000/optimizer.bin b/checkpoint-3000/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..72222d2a393a475bde644dccdf7f21f2be65538b --- /dev/null +++ b/checkpoint-3000/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5ce4f64e3721cee82e785573d2e2d0dc5bb5d4e473258292f28bdd5b7604547 +size 2890518478 diff --git a/checkpoint-3000/random_states_0.pkl b/checkpoint-3000/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..f683187699aa845b8dfb2638d9fd04d0fcffd8fc --- /dev/null +++ b/checkpoint-3000/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3bf127c824e39c6fece163173775330b09a828044085efc8822331118a02733f +size 14604 diff --git a/checkpoint-3000/scheduler.bin b/checkpoint-3000/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..ba2cc3b01b2417e557da8786431c98c8a94f48d2 --- /dev/null +++ b/checkpoint-3000/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:910e7b81b91dda8cf27d9a8f5ab0f27fcae595c1765ce971c4296041bcfbe551 +size 1000 diff --git a/checkpoint-3500/controlnet/config.json b/checkpoint-3500/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-3500/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-3500/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-3500/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..aaad40e0cd313a3711bafa0c78b77923e68407b4 --- /dev/null +++ b/checkpoint-3500/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:655aa823dd6a9f27d78eb648c997e6f24b72a08569a1e4b688efbb07436b340f +size 1445157120 diff --git a/checkpoint-3500/optimizer.bin b/checkpoint-3500/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..b63f119a66115f5b20a9dd29546ef0ba6eb38f06 --- /dev/null +++ b/checkpoint-3500/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5a9f69054b2cab8cff0130b381ad1320e71fc6c62a2564a4070e2a09cc0268c9 +size 2890518478 diff --git a/checkpoint-3500/random_states_0.pkl b/checkpoint-3500/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..dade7b75fa1abb8c81a21bfa5f6ef277ed134b3b --- /dev/null +++ b/checkpoint-3500/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5da3b486d7f5331ab2944d0053db132fcba758d2b50b98b835a177d491db538b +size 14604 diff --git a/checkpoint-3500/scheduler.bin b/checkpoint-3500/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..02ea9d6b2ac679ac729f671ff2fe620f63f00d81 --- /dev/null +++ b/checkpoint-3500/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:028c42f5139fef439208dd82620055df43c50636b0df3f5c77562cdc762631c4 +size 1000 diff --git a/checkpoint-4000/controlnet/config.json b/checkpoint-4000/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-4000/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-4000/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-4000/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..73cf846e9717614d2a58c161fa137cf957878089 --- /dev/null +++ b/checkpoint-4000/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:455514ff6f1264541a922c628ac15d3af84d80b069130dfd6708ea129b334e2b +size 1445157120 diff --git a/checkpoint-4000/optimizer.bin b/checkpoint-4000/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..5007dad9b5748effec5782704f104bce169e2f4e --- /dev/null +++ b/checkpoint-4000/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:69c9a9e132124cb8300cf6ea1cb08f3b6a3b95ec8f46e32b5530a63196b64f8d +size 2890518478 diff --git a/checkpoint-4000/random_states_0.pkl b/checkpoint-4000/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..dd7ef3c074732f87f20a35f39b5eacd837c3d10d --- /dev/null +++ b/checkpoint-4000/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b0270fda8b9ee10334f199c8fad52c0242e7b8e6f819fef43d0f2f67dc3fb297 +size 14604 diff --git a/checkpoint-4000/scheduler.bin b/checkpoint-4000/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..0b9724f9e172cb1291efb21d869a895d016ab566 --- /dev/null +++ b/checkpoint-4000/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b0da7609303685e021f64c9a7aa3dfe8810cac29fd2800358a6bfc3b05aaa4d +size 1000 diff --git a/checkpoint-4500/controlnet/config.json b/checkpoint-4500/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-4500/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-4500/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-4500/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..87b209679cde30b05887050ba8ca59ff960a07e4 --- /dev/null +++ b/checkpoint-4500/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:40c5786f9d407fcccca7c93b8929330374fe677160db27687262455f650d6be3 +size 1445157120 diff --git a/checkpoint-4500/optimizer.bin b/checkpoint-4500/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..1cbdd5c57f01b5fce91fa2586cb1f1d9355f7158 --- /dev/null +++ b/checkpoint-4500/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:788d3da61f5646cb381cbaa6feb5d384269a9065a06a21bcb6226ab34d19b9cf +size 2890518478 diff --git a/checkpoint-4500/random_states_0.pkl b/checkpoint-4500/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..4335dfc13bf2a08625a49620255658ec924e8735 --- /dev/null +++ b/checkpoint-4500/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7293cc46e0f70f83a7e8702781cffacebc852bb069fce00adf7278ed876daeb7 +size 14604 diff --git a/checkpoint-4500/scheduler.bin b/checkpoint-4500/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..9488c6d847961fbdce1e2a2c36ee042181a1d14a --- /dev/null +++ b/checkpoint-4500/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:57e3a3b44f386aba6f6ab46d1f33b36dde213b39026a4201509d429fb9b55fc4 +size 1000 diff --git a/checkpoint-500/controlnet/config.json b/checkpoint-500/controlnet/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/checkpoint-500/controlnet/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/checkpoint-500/controlnet/diffusion_pytorch_model.safetensors b/checkpoint-500/controlnet/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3781be730440abf778208131c838e6d96398588f --- /dev/null +++ b/checkpoint-500/controlnet/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fa22d0c91d7f1411e73a5dab0457d34ed2d26560d8d697d42cad93f89f72ee49 +size 1445157120 diff --git a/checkpoint-500/optimizer.bin b/checkpoint-500/optimizer.bin new file mode 100644 index 0000000000000000000000000000000000000000..e9add97c1b30e68918da41e5f148bee86089ebba --- /dev/null +++ b/checkpoint-500/optimizer.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6986037e378a27a4f759e533fbd3576d14bd2666f107e855b6a933ebea290934 +size 2890518478 diff --git a/checkpoint-500/random_states_0.pkl b/checkpoint-500/random_states_0.pkl new file mode 100644 index 0000000000000000000000000000000000000000..3abe4596ece74c2bad4bfc6ef1bba079ef41e09a --- /dev/null +++ b/checkpoint-500/random_states_0.pkl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad2e5c8cc4d8cef3410fd78a25d373b1a53964e3fb8d7edd53fc6baa73a5a52c +size 14604 diff --git a/checkpoint-500/scheduler.bin b/checkpoint-500/scheduler.bin new file mode 100644 index 0000000000000000000000000000000000000000..99f1ee7a9acad5a9c583be5e925340f00dfbfa9f --- /dev/null +++ b/checkpoint-500/scheduler.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:78924d09c0c344f579f288fc951eb08178192554f924151b7413fee247ff5804 +size 1000 diff --git a/config.json b/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c89d099d56823dcca9362abe9bf057e11de8a732 --- /dev/null +++ b/config.json @@ -0,0 +1,51 @@ +{ + "_class_name": "ControlNetModel", + "_diffusers_version": "0.32.0", + "act_fn": "silu", + "addition_embed_type": null, + "addition_embed_type_num_heads": 64, + "addition_time_embed_dim": null, + "attention_head_dim": 8, + "block_out_channels": [ + 320, + 640, + 1280, + 1280 + ], + "class_embed_type": null, + "conditioning_channels": 3, + "conditioning_embedding_out_channels": [ + 16, + 32, + 96, + 256 + ], + "controlnet_conditioning_channel_order": "rgb", + "cross_attention_dim": 768, + "down_block_types": [ + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "CrossAttnDownBlock2D", + "DownBlock2D" + ], + "downsample_padding": 1, + "encoder_hid_dim": null, + "encoder_hid_dim_type": null, + "flip_sin_to_cos": true, + "freq_shift": 0, + "global_pool_conditions": false, + "in_channels": 4, + "layers_per_block": 2, + "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, + "projection_class_embeddings_input_dim": null, + "resnet_time_scale_shift": "default", + "transformer_layers_per_block": 1, + "upcast_attention": false, + "use_linear_projection": false +} diff --git a/diffusion_pytorch_model.safetensors b/diffusion_pytorch_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ffdb42f95bdddeb2af58e2d025a9d780bcbad146 --- /dev/null +++ b/diffusion_pytorch_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:54b1848514585ab833a18d9cf2fc67ac2c620780cc268eccfbb0ce29812d274d +size 1445157120 diff --git a/image_control.png b/image_control.png new file mode 100644 index 0000000000000000000000000000000000000000..c9379f32621fc0b1f69bc6980db1e02eb359969d Binary files /dev/null and b/image_control.png differ diff --git a/images_0.png b/images_0.png new file mode 100644 index 0000000000000000000000000000000000000000..1746e600fe507cd82d855e5831a013dc4b7007f7 --- /dev/null +++ b/images_0.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0cb9a386aaa16ab08ed42ab3e114d83ec6c360f04a80aa87d0ae623012b928b5 +size 2907781 diff --git a/images_1.png b/images_1.png new file mode 100644 index 0000000000000000000000000000000000000000..3c620d683e2f4038e77d0a2868d49c25c5d0caee --- /dev/null +++ b/images_1.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:da83b906111b4a07619ae2458eeb9788673dd0722203be5d93e02dc92ddbdda2 +size 3011009 diff --git a/logs/train_controlnet/1736827839.273136/events.out.tfevents.1736827839.DESKTOP-III6ILP.19208.1 b/logs/train_controlnet/1736827839.273136/events.out.tfevents.1736827839.DESKTOP-III6ILP.19208.1 new file mode 100644 index 0000000000000000000000000000000000000000..b1a4292726c24e2d7f480b2652ebb7ce766761fe --- /dev/null +++ b/logs/train_controlnet/1736827839.273136/events.out.tfevents.1736827839.DESKTOP-III6ILP.19208.1 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:224214e731b779ba6cec983b523424c600ff8778c37fa69325344040956d67cd +size 2304 diff --git a/logs/train_controlnet/1736827839.3725805/hparams.yml b/logs/train_controlnet/1736827839.3725805/hparams.yml new file mode 100644 index 0000000000000000000000000000000000000000..2b09413ecdf4917ba7e73846b96bc23947dd233f --- /dev/null +++ b/logs/train_controlnet/1736827839.3725805/hparams.yml @@ -0,0 +1,50 @@ +adam_beta1: 0.9 +adam_beta2: 0.999 +adam_epsilon: 1.0e-08 +adam_weight_decay: 0.01 +allow_tf32: false +cache_dir: null +caption_column: text +checkpointing_steps: 500 +checkpoints_total_limit: null +conditioning_image_column: conditioning_image +controlnet_model_name_or_path: null +dataloader_num_workers: 0 +dataset_config_name: null +dataset_name: D:/lcb/diffusion/diffu_data/ +enable_xformers_memory_efficient_attention: false +gradient_accumulation_steps: 1 +gradient_checkpointing: false +hub_model_id: null +hub_token: null +image_column: image +learning_rate: 2.0e-06 +logging_dir: logs +lr_num_cycles: 1 +lr_power: 1.0 +lr_scheduler: constant +lr_warmup_steps: 500 +max_grad_norm: 1.0 +max_train_samples: null +max_train_steps: 4550 +mixed_precision: null +num_train_epochs: 5 +num_validation_images: 4 +output_dir: D:/lcb/diffusion/model0114 +pretrained_model_name_or_path: runwayml/stable-diffusion-v1-5 +proportion_empty_prompts: 0 +push_to_hub: true +report_to: tensorboard +resolution: 640 +resume_from_checkpoint: null +revision: null +scale_lr: false +seed: null +set_grads_to_none: false +tokenizer_name: null +tracker_project_name: train_controlnet +train_batch_size: 4 +train_data_dir: null +use_8bit_adam: false +validation_steps: 100 +variant: null diff --git a/logs/train_controlnet/events.out.tfevents.1736827839.DESKTOP-III6ILP.19208.0 b/logs/train_controlnet/events.out.tfevents.1736827839.DESKTOP-III6ILP.19208.0 new file mode 100644 index 0000000000000000000000000000000000000000..cd9e963213f12973a0afc807aefb6bb6c0401593 --- /dev/null +++ b/logs/train_controlnet/events.out.tfevents.1736827839.DESKTOP-III6ILP.19208.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ac93af16a7099079744f5ae938da743e27960062e02ce7544571e32cfb8df47 +size 269378305