Hugging Face's logo

Wawaworker
/
mrnx

Model card Files Community
1

Upload folder using huggingface_hub

#1
by Wawaworker - opened
base: refs/heads/main
←
from: refs/pr/1
Discussion Files changed
+927
-0
Files changed (18) hide show
  1. mrnx2024-06-23_03-36-58-save-245-35-0.safetensors +3 -0
  2. mrnx2024-06-23_03-36-58-save-245-35-0.yaml +100 -0
  3. mrnx2024-06-23_03-44-33-save-490-70-0.safetensors +3 -0
  4. mrnx2024-06-23_03-44-33-save-490-70-0.yaml +100 -0
  5. mrnx2024-06-23_03-52-26-save-735-105-0.safetensors +3 -0
  6. mrnx2024-06-23_03-52-26-save-735-105-0.yaml +100 -0
  7. mrnx2024-06-23_04-00-21-save-980-140-0.safetensors +3 -0
  8. mrnx2024-06-23_04-00-21-save-980-140-0.yaml +100 -0
  9. mrnx2024-06-23_04-08-18-save-1225-175-0.safetensors +3 -0
  10. mrnx2024-06-23_04-08-18-save-1225-175-0.yaml +100 -0
  11. mrnx2024-06-23_04-16-12-save-1470-210-0.safetensors +3 -0
  12. mrnx2024-06-23_04-16-12-save-1470-210-0.yaml +100 -0
  13. mrnx2024-06-23_04-24-07-save-1715-245-0.safetensors +3 -0
  14. mrnx2024-06-23_04-24-07-save-1715-245-0.yaml +100 -0
  15. mrnx2024-06-23_04-32-05-save-1960-280-0.safetensors +3 -0
  16. mrnx2024-06-23_04-32-05-save-1960-280-0.yaml +100 -0
  17. mrnx2024-06-23_04-40-03-save-2205-315-0.safetensors +3 -0
  18. mrnx2024-06-23_04-40-03-save-2205-315-0.yaml +100 -0
mrnx2024-06-23_03-36-58-save-245-35-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:005513975873e15f28feb79c07c77b83811ec4e215c4d4e711b2ab16ec3db9ca
3
+ size 6938084280
mrnx2024-06-23_03-36-58-save-245-35-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine
mrnx2024-06-23_03-44-33-save-490-70-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02c9236748e04330a4668d127381e00a885eab7474bf4f9629137cce71d003bb
3
+ size 6938084280
mrnx2024-06-23_03-44-33-save-490-70-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine
mrnx2024-06-23_03-52-26-save-735-105-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9722df53758d03690b866f7b65e3a35894799d07764891d50ba92affddd4cc48
3
+ size 6938084280
mrnx2024-06-23_03-52-26-save-735-105-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine
mrnx2024-06-23_04-00-21-save-980-140-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e207f485a40ffc4c66f3564bcc40a7799d507bc8aa271aed9a330ca31c0f2ec3
3
+ size 6938084280
mrnx2024-06-23_04-00-21-save-980-140-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine
mrnx2024-06-23_04-08-18-save-1225-175-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5b2cc1e6ceabc06f609d6716613ce2ee8c3c725e2490e69c17d4f74bccf31f0
3
+ size 6938084280
mrnx2024-06-23_04-08-18-save-1225-175-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine
mrnx2024-06-23_04-16-12-save-1470-210-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35dbe6e59be1e77ba58b433b895421f0366b018ad3221c1315bfc75ffeeb6de8
3
+ size 6938084280
mrnx2024-06-23_04-16-12-save-1470-210-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine
mrnx2024-06-23_04-24-07-save-1715-245-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6db26e273e03304186e6c86e11c9f4660686f984de6021c4294830342426f19
3
+ size 6938084280
mrnx2024-06-23_04-24-07-save-1715-245-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine
mrnx2024-06-23_04-32-05-save-1960-280-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62bc7f96625f632c1c4d967e2deb378fdcf1d5c42688bcfe9fd9f85fe65519f0
3
+ size 6938084280
mrnx2024-06-23_04-32-05-save-1960-280-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine
mrnx2024-06-23_04-40-03-save-2205-315-0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd07af660c3e74a1a63afe82ccd492b36297d45f41fa90bddaa1ac8451ec8a32
3
+ size 6938084280
mrnx2024-06-23_04-40-03-save-2205-315-0.yaml ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ model:
2
+ params:
3
+ conditioner_config:
4
+ params:
5
+ emb_models:
6
+ - input_key: txt
7
+ is_trainable: false
8
+ params:
9
+ layer: hidden
10
+ layer_idx: 11
11
+ target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
12
+ - input_key: txt
13
+ is_trainable: false
14
+ params:
15
+ always_return_pooled: true
16
+ arch: ViT-bigG-14
17
+ freeze: true
18
+ layer: penultimate
19
+ legacy: false
20
+ version: laion2b_s39b_b160k
21
+ target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
22
+ - input_key: original_size_as_tuple
23
+ is_trainable: false
24
+ params:
25
+ outdim: 256
26
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
27
+ - input_key: crop_coords_top_left
28
+ is_trainable: false
29
+ params:
30
+ outdim: 256
31
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
32
+ - input_key: target_size_as_tuple
33
+ is_trainable: false
34
+ params:
35
+ outdim: 256
36
+ target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
37
+ target: sgm.modules.GeneralConditioner
38
+ denoiser_config:
39
+ params:
40
+ discretization_config:
41
+ target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
42
+ num_idx: 1000
43
+ scaling_config:
44
+ target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
45
+ weighting_config:
46
+ target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
47
+ target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
48
+ disable_first_stage_autocast: true
49
+ first_stage_config:
50
+ params:
51
+ ddconfig:
52
+ attn_resolutions: []
53
+ attn_type: vanilla-xformers
54
+ ch: 128
55
+ ch_mult:
56
+ - 1
57
+ - 2
58
+ - 4
59
+ - 4
60
+ double_z: true
61
+ dropout: 0.0
62
+ in_channels: 3
63
+ num_res_blocks: 2
64
+ out_ch: 3
65
+ resolution: 256
66
+ z_channels: 4
67
+ embed_dim: 4
68
+ lossconfig:
69
+ target: torch.nn.Identity
70
+ monitor: val/rec_loss
71
+ target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
72
+ network_config:
73
+ params:
74
+ adm_in_channels: 2816
75
+ attention_resolutions:
76
+ - 4
77
+ - 2
78
+ channel_mult:
79
+ - 1
80
+ - 2
81
+ - 4
82
+ context_dim: 2048
83
+ in_channels: 4
84
+ legacy: false
85
+ model_channels: 320
86
+ num_classes: sequential
87
+ num_head_channels: 64
88
+ num_res_blocks: 2
89
+ out_channels: 4
90
+ spatial_transformer_attn_type: softmax-xformers
91
+ transformer_depth:
92
+ - 1
93
+ - 2
94
+ - 10
95
+ use_checkpoint: true
96
+ use_linear_in_transformer: true
97
+ use_spatial_transformer: true
98
+ target: sgm.modules.diffusionmodules.openaimodel.UNetModel
99
+ scale_factor: 0.13025
100
+ target: sgm.models.diffusion.DiffusionEngine