Upload folder using huggingface_hub (#1)
Browse files- 4276423d8d8a467d67146824ef6b3a99d968aa7604a1bda88ad6f3047784ed1d (27672e055f8d3306ed2ee2f7b6bbaf54e1a7e5d4)
- a1a451a8c5ba6d2cd744cbb4634d82d183536b899e4208744c35d417f9e6cf82 (0dc880e7a094ae8faef48edad97f6507e447cf7b)
- 840441395bb88114726ff6d28bebab51b837906f4ec88ca7a57a03f3f5f40113 (f428b39592313453691c876004a28c1ee7b335c5)
- e1b8e4635c2c688bf8dffbb14077e303bc5803b88f3d22c872b0a3901ae05036 (7c5b67ac1f7109b999cc85224cc40bd7766dc8c7)
- c041581efa2f1577fdd14ce5b55ce70682ebd16ee8b9fb9a984a3c7e9dcd3aef (b8e9d19cb09df4b871d9634ad3d3f4070c275565)
- 6a335719721f84d3645b78f43125cbbd0ba645b27a29ac39c959a1d02a168f81 (ad154ff89485b90657c59ca70215291479526248)
- 2c28e6714e15877fbbb256da4c0f7dc766e4cc0ad104c96a7439711f1b69b9a0 (977435a9a44bad59c59cbb69c30290ec4059070d)
- 7c006054dabb31a85eee349956900c2c78830e04360ac7cea484ef6625fc075e (e4dbf05ef6dd776d165cf9237f8f702fc36d6fdc)
- 4b3bde7574a7279674614d261549688889610267b06b59740e9bb0a2d507bc10 (8e309d2a782035aefad9be601945fc4e53def9d3)
- fde369fdfbb441fe19aa018665e751e74a3b47cd9f0a45a2fb4539e5931e5bd1 (33f500930e021f2ad7fefafe6d5f37c8d832bdf8)
- 439680b38c09e35222547d9b63d9110abdc05c121e60b616c066dbef46e78b2a (a9c149f4865b3faa282948de9fdabaa81177e743)
- checkpoint-2600/optimizer.bin +3 -0
- checkpoint-2600/random_states_0.pkl +3 -0
- checkpoint-2600/scaler.pt +3 -0
- checkpoint-2600/scheduler.bin +3 -0
- checkpoint-2600/text_encoder/config.json +25 -0
- checkpoint-2600/text_encoder/pytorch_model.bin +3 -0
- checkpoint-2600/unet/config.json +66 -0
- checkpoint-2600/unet/diffusion_pytorch_model.bin +3 -0
- checkpoint-2800/optimizer.bin +3 -0
- checkpoint-2800/random_states_0.pkl +3 -0
- checkpoint-2800/scaler.pt +3 -0
- checkpoint-2800/scheduler.bin +3 -0
- checkpoint-2800/text_encoder/config.json +25 -0
- checkpoint-2800/text_encoder/pytorch_model.bin +3 -0
- checkpoint-2800/unet/config.json +66 -0
- checkpoint-2800/unet/diffusion_pytorch_model.bin +3 -0
- checkpoint-3000/optimizer.bin +3 -0
- checkpoint-3000/random_states_0.pkl +3 -0
- checkpoint-3000/scaler.pt +3 -0
- checkpoint-3000/scheduler.bin +3 -0
- checkpoint-3000/text_encoder/config.json +25 -0
- checkpoint-3000/text_encoder/pytorch_model.bin +3 -0
- checkpoint-3000/unet/config.json +66 -0
- checkpoint-3000/unet/diffusion_pytorch_model.bin +3 -0
- checkpoint-3200/optimizer.bin +3 -0
- checkpoint-3200/text_encoder/config.json +25 -0
- checkpoint-3200/text_encoder/pytorch_model.bin +3 -0
- checkpoint-3200/unet/config.json +66 -0
- checkpoint-3200/unet/diffusion_pytorch_model.bin +3 -0
- logs/dreambooth/1684156168.9564312/events.out.tfevents.1684156168.3144edec4884.489.1 +3 -0
- logs/dreambooth/1684156168.9588692/.ipynb_checkpoints/hparams-checkpoint.yml +52 -0
- logs/dreambooth/1684156168.9588692/hparams.yml +52 -0
- logs/dreambooth/events.out.tfevents.1684156168.3144edec4884.489.0 +3 -0
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a36f2969caa2bcd8321024e94c9f37e71f85c0f1d71a9859ab69962bbbff0db5
|
3 |
+
size 2421618099
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3e4550019fb7193af7a7915840c56728074a14ba94a8704bec8da214c589139c
|
3 |
+
size 15691
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4a539466aff5683738c96fa14692f1a5f94bdb15cfa06056afc7125be84466d3
|
3 |
+
size 557
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c0f88d065a93974f98c2659e2b0620cffd45a1c72e3d911c12a3c5aa9ae9ae8c
|
3 |
+
size 563
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "stabilityai/stable-diffusion-2-1",
|
3 |
+
"architectures": [
|
4 |
+
"CLIPTextModel"
|
5 |
+
],
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"dropout": 0.0,
|
9 |
+
"eos_token_id": 2,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_size": 1024,
|
12 |
+
"initializer_factor": 1.0,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 4096,
|
15 |
+
"layer_norm_eps": 1e-05,
|
16 |
+
"max_position_embeddings": 77,
|
17 |
+
"model_type": "clip_text_model",
|
18 |
+
"num_attention_heads": 16,
|
19 |
+
"num_hidden_layers": 23,
|
20 |
+
"pad_token_id": 1,
|
21 |
+
"projection_dim": 512,
|
22 |
+
"torch_dtype": "float32",
|
23 |
+
"transformers_version": "4.29.1",
|
24 |
+
"vocab_size": 49408
|
25 |
+
}
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4321f9c9a64e57229f670de289f7d2ff1ad62c5f8ada1ac3a92132ce28cd629a
|
3 |
+
size 1361679905
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "UNet2DConditionModel",
|
3 |
+
"_diffusers_version": "0.17.0.dev0",
|
4 |
+
"_name_or_path": "stabilityai/stable-diffusion-2-1",
|
5 |
+
"act_fn": "silu",
|
6 |
+
"addition_embed_type": null,
|
7 |
+
"addition_embed_type_num_heads": 64,
|
8 |
+
"attention_head_dim": [
|
9 |
+
5,
|
10 |
+
10,
|
11 |
+
20,
|
12 |
+
20
|
13 |
+
],
|
14 |
+
"block_out_channels": [
|
15 |
+
320,
|
16 |
+
640,
|
17 |
+
1280,
|
18 |
+
1280
|
19 |
+
],
|
20 |
+
"center_input_sample": false,
|
21 |
+
"class_embed_type": null,
|
22 |
+
"class_embeddings_concat": false,
|
23 |
+
"conv_in_kernel": 3,
|
24 |
+
"conv_out_kernel": 3,
|
25 |
+
"cross_attention_dim": 1024,
|
26 |
+
"cross_attention_norm": null,
|
27 |
+
"down_block_types": [
|
28 |
+
"CrossAttnDownBlock2D",
|
29 |
+
"CrossAttnDownBlock2D",
|
30 |
+
"CrossAttnDownBlock2D",
|
31 |
+
"DownBlock2D"
|
32 |
+
],
|
33 |
+
"downsample_padding": 1,
|
34 |
+
"dual_cross_attention": false,
|
35 |
+
"encoder_hid_dim": null,
|
36 |
+
"flip_sin_to_cos": true,
|
37 |
+
"freq_shift": 0,
|
38 |
+
"in_channels": 4,
|
39 |
+
"layers_per_block": 2,
|
40 |
+
"mid_block_only_cross_attention": null,
|
41 |
+
"mid_block_scale_factor": 1,
|
42 |
+
"mid_block_type": "UNetMidBlock2DCrossAttn",
|
43 |
+
"norm_eps": 1e-05,
|
44 |
+
"norm_num_groups": 32,
|
45 |
+
"num_class_embeds": null,
|
46 |
+
"only_cross_attention": false,
|
47 |
+
"out_channels": 4,
|
48 |
+
"projection_class_embeddings_input_dim": null,
|
49 |
+
"resnet_out_scale_factor": 1.0,
|
50 |
+
"resnet_skip_time_act": false,
|
51 |
+
"resnet_time_scale_shift": "default",
|
52 |
+
"sample_size": 96,
|
53 |
+
"time_cond_proj_dim": null,
|
54 |
+
"time_embedding_act_fn": null,
|
55 |
+
"time_embedding_dim": null,
|
56 |
+
"time_embedding_type": "positional",
|
57 |
+
"timestep_post_act": null,
|
58 |
+
"up_block_types": [
|
59 |
+
"UpBlock2D",
|
60 |
+
"CrossAttnUpBlock2D",
|
61 |
+
"CrossAttnUpBlock2D",
|
62 |
+
"CrossAttnUpBlock2D"
|
63 |
+
],
|
64 |
+
"upcast_attention": true,
|
65 |
+
"use_linear_projection": true
|
66 |
+
}
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9967419e6041664fcb2e50e1b24c97a767014553bc85b807ab417933680db6c6
|
3 |
+
size 3463934693
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:90fe3da84bd74b88010f355046563da5ef690ee078487a190589fa5732348e83
|
3 |
+
size 2421618099
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f72e722987f95d8323d7fae5fc10d4b225614f3148a0d5cc205d2aabdedd3a10
|
3 |
+
size 15691
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:61a896139947e0a8f0817315828ec9a0e47ca7e1e7096a39eb6a7eccaf6b1764
|
3 |
+
size 557
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:70917121c12d2ac2e3f59806d8c31a5576975367ff435541dbf3b541728b7b1b
|
3 |
+
size 563
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "stabilityai/stable-diffusion-2-1",
|
3 |
+
"architectures": [
|
4 |
+
"CLIPTextModel"
|
5 |
+
],
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"dropout": 0.0,
|
9 |
+
"eos_token_id": 2,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_size": 1024,
|
12 |
+
"initializer_factor": 1.0,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 4096,
|
15 |
+
"layer_norm_eps": 1e-05,
|
16 |
+
"max_position_embeddings": 77,
|
17 |
+
"model_type": "clip_text_model",
|
18 |
+
"num_attention_heads": 16,
|
19 |
+
"num_hidden_layers": 23,
|
20 |
+
"pad_token_id": 1,
|
21 |
+
"projection_dim": 512,
|
22 |
+
"torch_dtype": "float32",
|
23 |
+
"transformers_version": "4.29.1",
|
24 |
+
"vocab_size": 49408
|
25 |
+
}
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c4fbc33bddae47a6e54afc776b54ea4040d81c339fe93fe7e2bad6a2aa1e942c
|
3 |
+
size 1361679905
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "UNet2DConditionModel",
|
3 |
+
"_diffusers_version": "0.17.0.dev0",
|
4 |
+
"_name_or_path": "stabilityai/stable-diffusion-2-1",
|
5 |
+
"act_fn": "silu",
|
6 |
+
"addition_embed_type": null,
|
7 |
+
"addition_embed_type_num_heads": 64,
|
8 |
+
"attention_head_dim": [
|
9 |
+
5,
|
10 |
+
10,
|
11 |
+
20,
|
12 |
+
20
|
13 |
+
],
|
14 |
+
"block_out_channels": [
|
15 |
+
320,
|
16 |
+
640,
|
17 |
+
1280,
|
18 |
+
1280
|
19 |
+
],
|
20 |
+
"center_input_sample": false,
|
21 |
+
"class_embed_type": null,
|
22 |
+
"class_embeddings_concat": false,
|
23 |
+
"conv_in_kernel": 3,
|
24 |
+
"conv_out_kernel": 3,
|
25 |
+
"cross_attention_dim": 1024,
|
26 |
+
"cross_attention_norm": null,
|
27 |
+
"down_block_types": [
|
28 |
+
"CrossAttnDownBlock2D",
|
29 |
+
"CrossAttnDownBlock2D",
|
30 |
+
"CrossAttnDownBlock2D",
|
31 |
+
"DownBlock2D"
|
32 |
+
],
|
33 |
+
"downsample_padding": 1,
|
34 |
+
"dual_cross_attention": false,
|
35 |
+
"encoder_hid_dim": null,
|
36 |
+
"flip_sin_to_cos": true,
|
37 |
+
"freq_shift": 0,
|
38 |
+
"in_channels": 4,
|
39 |
+
"layers_per_block": 2,
|
40 |
+
"mid_block_only_cross_attention": null,
|
41 |
+
"mid_block_scale_factor": 1,
|
42 |
+
"mid_block_type": "UNetMidBlock2DCrossAttn",
|
43 |
+
"norm_eps": 1e-05,
|
44 |
+
"norm_num_groups": 32,
|
45 |
+
"num_class_embeds": null,
|
46 |
+
"only_cross_attention": false,
|
47 |
+
"out_channels": 4,
|
48 |
+
"projection_class_embeddings_input_dim": null,
|
49 |
+
"resnet_out_scale_factor": 1.0,
|
50 |
+
"resnet_skip_time_act": false,
|
51 |
+
"resnet_time_scale_shift": "default",
|
52 |
+
"sample_size": 96,
|
53 |
+
"time_cond_proj_dim": null,
|
54 |
+
"time_embedding_act_fn": null,
|
55 |
+
"time_embedding_dim": null,
|
56 |
+
"time_embedding_type": "positional",
|
57 |
+
"timestep_post_act": null,
|
58 |
+
"up_block_types": [
|
59 |
+
"UpBlock2D",
|
60 |
+
"CrossAttnUpBlock2D",
|
61 |
+
"CrossAttnUpBlock2D",
|
62 |
+
"CrossAttnUpBlock2D"
|
63 |
+
],
|
64 |
+
"upcast_attention": true,
|
65 |
+
"use_linear_projection": true
|
66 |
+
}
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a90127fbc0d9628c0129b1fc4df59bd1eafabd232167cb894a3e73e64830a2a2
|
3 |
+
size 3463934693
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8302f9804cb0ef55e40c4ea80ac53d90036d45ca43f4029872809a63615cff6e
|
3 |
+
size 2421618099
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4de02617490feb242ab55f9b0e0744f97b0845893d38dfb750cab353208782f8
|
3 |
+
size 15691
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9b379633b32f4e5b6d3a8b5862ea536f97221101d1391f1d8595d83bd7844a22
|
3 |
+
size 557
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8560fc69de46fcd9b437220ab2e8590aeb0fad454c7831b9c2c403da6461c941
|
3 |
+
size 563
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "stabilityai/stable-diffusion-2-1",
|
3 |
+
"architectures": [
|
4 |
+
"CLIPTextModel"
|
5 |
+
],
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"dropout": 0.0,
|
9 |
+
"eos_token_id": 2,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_size": 1024,
|
12 |
+
"initializer_factor": 1.0,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 4096,
|
15 |
+
"layer_norm_eps": 1e-05,
|
16 |
+
"max_position_embeddings": 77,
|
17 |
+
"model_type": "clip_text_model",
|
18 |
+
"num_attention_heads": 16,
|
19 |
+
"num_hidden_layers": 23,
|
20 |
+
"pad_token_id": 1,
|
21 |
+
"projection_dim": 512,
|
22 |
+
"torch_dtype": "float32",
|
23 |
+
"transformers_version": "4.29.1",
|
24 |
+
"vocab_size": 49408
|
25 |
+
}
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1a88a8cd21c5ba684e7e957e0a47d9c3b6b417ce825c02c8f2a76cad99ea043c
|
3 |
+
size 1361679905
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "UNet2DConditionModel",
|
3 |
+
"_diffusers_version": "0.17.0.dev0",
|
4 |
+
"_name_or_path": "stabilityai/stable-diffusion-2-1",
|
5 |
+
"act_fn": "silu",
|
6 |
+
"addition_embed_type": null,
|
7 |
+
"addition_embed_type_num_heads": 64,
|
8 |
+
"attention_head_dim": [
|
9 |
+
5,
|
10 |
+
10,
|
11 |
+
20,
|
12 |
+
20
|
13 |
+
],
|
14 |
+
"block_out_channels": [
|
15 |
+
320,
|
16 |
+
640,
|
17 |
+
1280,
|
18 |
+
1280
|
19 |
+
],
|
20 |
+
"center_input_sample": false,
|
21 |
+
"class_embed_type": null,
|
22 |
+
"class_embeddings_concat": false,
|
23 |
+
"conv_in_kernel": 3,
|
24 |
+
"conv_out_kernel": 3,
|
25 |
+
"cross_attention_dim": 1024,
|
26 |
+
"cross_attention_norm": null,
|
27 |
+
"down_block_types": [
|
28 |
+
"CrossAttnDownBlock2D",
|
29 |
+
"CrossAttnDownBlock2D",
|
30 |
+
"CrossAttnDownBlock2D",
|
31 |
+
"DownBlock2D"
|
32 |
+
],
|
33 |
+
"downsample_padding": 1,
|
34 |
+
"dual_cross_attention": false,
|
35 |
+
"encoder_hid_dim": null,
|
36 |
+
"flip_sin_to_cos": true,
|
37 |
+
"freq_shift": 0,
|
38 |
+
"in_channels": 4,
|
39 |
+
"layers_per_block": 2,
|
40 |
+
"mid_block_only_cross_attention": null,
|
41 |
+
"mid_block_scale_factor": 1,
|
42 |
+
"mid_block_type": "UNetMidBlock2DCrossAttn",
|
43 |
+
"norm_eps": 1e-05,
|
44 |
+
"norm_num_groups": 32,
|
45 |
+
"num_class_embeds": null,
|
46 |
+
"only_cross_attention": false,
|
47 |
+
"out_channels": 4,
|
48 |
+
"projection_class_embeddings_input_dim": null,
|
49 |
+
"resnet_out_scale_factor": 1.0,
|
50 |
+
"resnet_skip_time_act": false,
|
51 |
+
"resnet_time_scale_shift": "default",
|
52 |
+
"sample_size": 96,
|
53 |
+
"time_cond_proj_dim": null,
|
54 |
+
"time_embedding_act_fn": null,
|
55 |
+
"time_embedding_dim": null,
|
56 |
+
"time_embedding_type": "positional",
|
57 |
+
"timestep_post_act": null,
|
58 |
+
"up_block_types": [
|
59 |
+
"UpBlock2D",
|
60 |
+
"CrossAttnUpBlock2D",
|
61 |
+
"CrossAttnUpBlock2D",
|
62 |
+
"CrossAttnUpBlock2D"
|
63 |
+
],
|
64 |
+
"upcast_attention": true,
|
65 |
+
"use_linear_projection": true
|
66 |
+
}
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3d2e761597b019bf757aa3bdd39e7e0e13e359debbf5bb25c14ce3f8ac25f92a
|
3 |
+
size 3463934693
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:07f4112fe8dd7c5e35a0f4566e398be46a3972fb6578b9fc3209044a7c12f4dc
|
3 |
+
size 486015104
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "stabilityai/stable-diffusion-2-1",
|
3 |
+
"architectures": [
|
4 |
+
"CLIPTextModel"
|
5 |
+
],
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"dropout": 0.0,
|
9 |
+
"eos_token_id": 2,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_size": 1024,
|
12 |
+
"initializer_factor": 1.0,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 4096,
|
15 |
+
"layer_norm_eps": 1e-05,
|
16 |
+
"max_position_embeddings": 77,
|
17 |
+
"model_type": "clip_text_model",
|
18 |
+
"num_attention_heads": 16,
|
19 |
+
"num_hidden_layers": 23,
|
20 |
+
"pad_token_id": 1,
|
21 |
+
"projection_dim": 512,
|
22 |
+
"torch_dtype": "float32",
|
23 |
+
"transformers_version": "4.29.1",
|
24 |
+
"vocab_size": 49408
|
25 |
+
}
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9472547fc09233c48c0a616a3be5983f00987b9769491a02c8ec73ae88c51d69
|
3 |
+
size 1361679905
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "UNet2DConditionModel",
|
3 |
+
"_diffusers_version": "0.17.0.dev0",
|
4 |
+
"_name_or_path": "stabilityai/stable-diffusion-2-1",
|
5 |
+
"act_fn": "silu",
|
6 |
+
"addition_embed_type": null,
|
7 |
+
"addition_embed_type_num_heads": 64,
|
8 |
+
"attention_head_dim": [
|
9 |
+
5,
|
10 |
+
10,
|
11 |
+
20,
|
12 |
+
20
|
13 |
+
],
|
14 |
+
"block_out_channels": [
|
15 |
+
320,
|
16 |
+
640,
|
17 |
+
1280,
|
18 |
+
1280
|
19 |
+
],
|
20 |
+
"center_input_sample": false,
|
21 |
+
"class_embed_type": null,
|
22 |
+
"class_embeddings_concat": false,
|
23 |
+
"conv_in_kernel": 3,
|
24 |
+
"conv_out_kernel": 3,
|
25 |
+
"cross_attention_dim": 1024,
|
26 |
+
"cross_attention_norm": null,
|
27 |
+
"down_block_types": [
|
28 |
+
"CrossAttnDownBlock2D",
|
29 |
+
"CrossAttnDownBlock2D",
|
30 |
+
"CrossAttnDownBlock2D",
|
31 |
+
"DownBlock2D"
|
32 |
+
],
|
33 |
+
"downsample_padding": 1,
|
34 |
+
"dual_cross_attention": false,
|
35 |
+
"encoder_hid_dim": null,
|
36 |
+
"flip_sin_to_cos": true,
|
37 |
+
"freq_shift": 0,
|
38 |
+
"in_channels": 4,
|
39 |
+
"layers_per_block": 2,
|
40 |
+
"mid_block_only_cross_attention": null,
|
41 |
+
"mid_block_scale_factor": 1,
|
42 |
+
"mid_block_type": "UNetMidBlock2DCrossAttn",
|
43 |
+
"norm_eps": 1e-05,
|
44 |
+
"norm_num_groups": 32,
|
45 |
+
"num_class_embeds": null,
|
46 |
+
"only_cross_attention": false,
|
47 |
+
"out_channels": 4,
|
48 |
+
"projection_class_embeddings_input_dim": null,
|
49 |
+
"resnet_out_scale_factor": 1.0,
|
50 |
+
"resnet_skip_time_act": false,
|
51 |
+
"resnet_time_scale_shift": "default",
|
52 |
+
"sample_size": 96,
|
53 |
+
"time_cond_proj_dim": null,
|
54 |
+
"time_embedding_act_fn": null,
|
55 |
+
"time_embedding_dim": null,
|
56 |
+
"time_embedding_type": "positional",
|
57 |
+
"timestep_post_act": null,
|
58 |
+
"up_block_types": [
|
59 |
+
"UpBlock2D",
|
60 |
+
"CrossAttnUpBlock2D",
|
61 |
+
"CrossAttnUpBlock2D",
|
62 |
+
"CrossAttnUpBlock2D"
|
63 |
+
],
|
64 |
+
"upcast_attention": true,
|
65 |
+
"use_linear_projection": true
|
66 |
+
}
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:787845ed0b6308aae1f2d2b33eec0d10b21d07ad8c8cadab66b4ca7c7030dc16
|
3 |
+
size 3463934693
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8c03dbee968df4399980c4c6394b0a5acd2cb27db1986671e1f5e7adb33c5c84
|
3 |
+
size 2724
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
adam_beta1: 0.9
|
2 |
+
adam_beta2: 0.999
|
3 |
+
adam_epsilon: 1.0e-08
|
4 |
+
adam_weight_decay: 0.01
|
5 |
+
allow_tf32: false
|
6 |
+
center_crop: false
|
7 |
+
checkpointing_steps: 200
|
8 |
+
checkpoints_total_limit: 3
|
9 |
+
class_data_dir: =./dataset/office_building_sampled
|
10 |
+
class_prompt: rendering of office building.
|
11 |
+
dataloader_num_workers: 0
|
12 |
+
enable_xformers_memory_efficient_attention: false
|
13 |
+
gradient_accumulation_steps: 1
|
14 |
+
gradient_checkpointing: true
|
15 |
+
hub_model_id: null
|
16 |
+
hub_token: null
|
17 |
+
instance_data_dir: ./dataset/office_building_pngProcessed
|
18 |
+
instance_prompt: rendering of ttarch office building.
|
19 |
+
learning_rate: 1.0e-06
|
20 |
+
local_rank: 0
|
21 |
+
logging_dir: logs
|
22 |
+
lr_num_cycles: 1
|
23 |
+
lr_power: 1.0
|
24 |
+
lr_scheduler: constant
|
25 |
+
lr_warmup_steps: 0
|
26 |
+
max_grad_norm: 1.0
|
27 |
+
max_train_steps: 3201
|
28 |
+
mixed_precision: null
|
29 |
+
num_class_images: 100
|
30 |
+
num_train_epochs: 214
|
31 |
+
num_validation_images: 4
|
32 |
+
offset_noise: false
|
33 |
+
output_dir: ./models/office-dbModel-v0-4
|
34 |
+
pretrained_model_name_or_path: stabilityai/stable-diffusion-2-1
|
35 |
+
prior_generation_precision: null
|
36 |
+
prior_loss_weight: 1.0
|
37 |
+
push_to_hub: false
|
38 |
+
report_to: tensorboard
|
39 |
+
resolution: 768
|
40 |
+
resume_from_checkpoint: null
|
41 |
+
revision: null
|
42 |
+
sample_batch_size: 4
|
43 |
+
scale_lr: false
|
44 |
+
seed: null
|
45 |
+
set_grads_to_none: false
|
46 |
+
tokenizer_name: null
|
47 |
+
train_batch_size: 1
|
48 |
+
train_text_encoder: true
|
49 |
+
use_8bit_adam: true
|
50 |
+
validation_prompt: null
|
51 |
+
validation_steps: 100
|
52 |
+
with_prior_preservation: false
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
adam_beta1: 0.9
|
2 |
+
adam_beta2: 0.999
|
3 |
+
adam_epsilon: 1.0e-08
|
4 |
+
adam_weight_decay: 0.01
|
5 |
+
allow_tf32: false
|
6 |
+
center_crop: false
|
7 |
+
checkpointing_steps: 200
|
8 |
+
checkpoints_total_limit: 3
|
9 |
+
class_data_dir: =./dataset/office_building_sampled
|
10 |
+
class_prompt: rendering of office building.
|
11 |
+
dataloader_num_workers: 0
|
12 |
+
enable_xformers_memory_efficient_attention: false
|
13 |
+
gradient_accumulation_steps: 1
|
14 |
+
gradient_checkpointing: true
|
15 |
+
hub_model_id: null
|
16 |
+
hub_token: null
|
17 |
+
instance_data_dir: ./dataset/office_building_pngProcessed
|
18 |
+
instance_prompt: rendering of ttarch office building.
|
19 |
+
learning_rate: 1.0e-06
|
20 |
+
local_rank: 0
|
21 |
+
logging_dir: logs
|
22 |
+
lr_num_cycles: 1
|
23 |
+
lr_power: 1.0
|
24 |
+
lr_scheduler: constant
|
25 |
+
lr_warmup_steps: 0
|
26 |
+
max_grad_norm: 1.0
|
27 |
+
max_train_steps: 3201
|
28 |
+
mixed_precision: null
|
29 |
+
num_class_images: 100
|
30 |
+
num_train_epochs: 214
|
31 |
+
num_validation_images: 4
|
32 |
+
offset_noise: false
|
33 |
+
output_dir: ./models/office-dbModel-v0-4
|
34 |
+
pretrained_model_name_or_path: stabilityai/stable-diffusion-2-1
|
35 |
+
prior_generation_precision: null
|
36 |
+
prior_loss_weight: 1.0
|
37 |
+
push_to_hub: false
|
38 |
+
report_to: tensorboard
|
39 |
+
resolution: 768
|
40 |
+
resume_from_checkpoint: null
|
41 |
+
revision: null
|
42 |
+
sample_batch_size: 4
|
43 |
+
scale_lr: false
|
44 |
+
seed: null
|
45 |
+
set_grads_to_none: false
|
46 |
+
tokenizer_name: null
|
47 |
+
train_batch_size: 1
|
48 |
+
train_text_encoder: true
|
49 |
+
use_8bit_adam: true
|
50 |
+
validation_prompt: null
|
51 |
+
validation_steps: 100
|
52 |
+
with_prior_preservation: false
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6d520fda267a79aa9193b2363ee2d9005d74cba6a1c2419f1ce4e8e956778b1a
|
3 |
+
size 268550
|