PseudoTerminal X
commited on
Commit
•
bd0317f
1
Parent(s):
920db01
Trained for 0 epochs and 400 steps.
Browse filesTrained with datasets ['text-embeds-sdxl', 'training-test']
Learning rate 4e-07, batch size 4, and 4 gradient accumulation steps.
Used DDPM noise scheduler for training with v_prediction prediction type and rescaled_betas_zero_snr=True
Using 'trailing' timestep spacing.
Base model: ptx0/terminus-xl-velocity-v2
VAE: madebyollin/sdxl-vae-fp16-fix
- README.md +6 -110
- model_index.json +15 -7
- scheduler/scheduler_config.json +10 -3
- tokenizer/tokenizer_config.json +20 -23
- tokenizer_2/tokenizer_config.json +28 -23
- unet/config.json +1 -1
- vae/config.json +3 -1
README.md
CHANGED
@@ -58,116 +58,12 @@ You may reuse the base model text encoder for inference.
|
|
58 |
|
59 |
## Datasets
|
60 |
|
61 |
-
###
|
62 |
-
- Repeats:
|
63 |
-
- Total number of images:
|
64 |
-
- Total number of aspect buckets:
|
65 |
- Resolution: 1.0 megapixels
|
66 |
- Cropped: True
|
67 |
-
- Crop style:
|
68 |
-
- Crop aspect:
|
69 |
-
### movieposters
|
70 |
-
- Repeats: 0
|
71 |
-
- Total number of images: 208
|
72 |
-
- Total number of aspect buckets: 3
|
73 |
-
- Resolution: 1.0 megapixels
|
74 |
-
- Cropped: True
|
75 |
-
- Crop style: random
|
76 |
-
- Crop aspect: random
|
77 |
-
### pixel-art
|
78 |
-
- Repeats: 0
|
79 |
-
- Total number of images: 112
|
80 |
-
- Total number of aspect buckets: 3
|
81 |
-
- Resolution: 1.0 megapixels
|
82 |
-
- Cropped: True
|
83 |
-
- Crop style: random
|
84 |
-
- Crop aspect: random
|
85 |
-
### moviecollection
|
86 |
-
- Repeats: 0
|
87 |
-
- Total number of images: 208
|
88 |
-
- Total number of aspect buckets: 3
|
89 |
-
- Resolution: 1.0 megapixels
|
90 |
-
- Cropped: True
|
91 |
-
- Crop style: random
|
92 |
-
- Crop aspect: random
|
93 |
-
### ethnic
|
94 |
-
- Repeats: 0
|
95 |
-
- Total number of images: 368
|
96 |
-
- Total number of aspect buckets: 2
|
97 |
-
- Resolution: 1.0 megapixels
|
98 |
-
- Cropped: True
|
99 |
-
- Crop style: random
|
100 |
-
- Crop aspect: random
|
101 |
-
### shutterstock
|
102 |
-
- Repeats: 0
|
103 |
-
- Total number of images: 2608
|
104 |
-
- Total number of aspect buckets: 3
|
105 |
-
- Resolution: 1.0 megapixels
|
106 |
-
- Cropped: True
|
107 |
-
- Crop style: random
|
108 |
-
- Crop aspect: random
|
109 |
-
### cinemamix-1mp
|
110 |
-
- Repeats: 0
|
111 |
-
- Total number of images: 1104
|
112 |
-
- Total number of aspect buckets: 1
|
113 |
-
- Resolution: 1.0 megapixels
|
114 |
-
- Cropped: True
|
115 |
-
- Crop style: random
|
116 |
-
- Crop aspect: random
|
117 |
-
### nsfw-1024
|
118 |
-
- Repeats: 0
|
119 |
-
- Total number of images: 1328
|
120 |
-
- Total number of aspect buckets: 3
|
121 |
-
- Resolution: 1.0 megapixels
|
122 |
-
- Cropped: True
|
123 |
-
- Crop style: random
|
124 |
-
- Crop aspect: random
|
125 |
-
### anatomy
|
126 |
-
- Repeats: 5
|
127 |
-
- Total number of images: 2032
|
128 |
-
- Total number of aspect buckets: 3
|
129 |
-
- Resolution: 1.0 megapixels
|
130 |
-
- Cropped: True
|
131 |
-
- Crop style: random
|
132 |
-
- Crop aspect: random
|
133 |
-
### bg20k-1024
|
134 |
-
- Repeats: 0
|
135 |
-
- Total number of images: 11136
|
136 |
-
- Total number of aspect buckets: 3
|
137 |
-
- Resolution: 1.0 megapixels
|
138 |
-
- Cropped: True
|
139 |
-
- Crop style: random
|
140 |
-
- Crop aspect: random
|
141 |
-
### yoga
|
142 |
-
- Repeats: 0
|
143 |
-
- Total number of images: 448
|
144 |
-
- Total number of aspect buckets: 2
|
145 |
-
- Resolution: 1.0 megapixels
|
146 |
-
- Cropped: True
|
147 |
-
- Crop style: random
|
148 |
-
- Crop aspect: random
|
149 |
-
### photo-aesthetics
|
150 |
-
- Repeats: 0
|
151 |
-
- Total number of images: 4128
|
152 |
-
- Total number of aspect buckets: 3
|
153 |
-
- Resolution: 1.0 megapixels
|
154 |
-
- Cropped: True
|
155 |
-
- Crop style: random
|
156 |
-
- Crop aspect: random
|
157 |
-
### text-1mp
|
158 |
-
- Repeats: 5
|
159 |
-
- Total number of images: 1616
|
160 |
-
- Total number of aspect buckets: 3
|
161 |
-
- Resolution: 1.0 megapixels
|
162 |
-
- Cropped: True
|
163 |
-
- Crop style: random
|
164 |
-
- Crop aspect: random
|
165 |
-
### photo-concept-bucket
|
166 |
-
- Repeats: 0
|
167 |
-
- Total number of images: 70912
|
168 |
-
- Total number of aspect buckets: 3
|
169 |
-
- Resolution: 1.0 megapixels
|
170 |
-
- Cropped: True
|
171 |
-
- Crop style: random
|
172 |
-
- Crop aspect: random
|
173 |
|
|
|
58 |
|
59 |
## Datasets
|
60 |
|
61 |
+
### training-test
|
62 |
+
- Repeats: 1
|
63 |
+
- Total number of images: 192
|
64 |
+
- Total number of aspect buckets: 17
|
65 |
- Resolution: 1.0 megapixels
|
66 |
- Cropped: True
|
67 |
+
- Crop style: corner
|
68 |
+
- Crop aspect: preserve
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
69 |
|
model_index.json
CHANGED
@@ -1,19 +1,27 @@
|
|
1 |
{
|
2 |
"_class_name": "StableDiffusionXLPipeline",
|
3 |
-
"_diffusers_version": "0.
|
4 |
-
"_name_or_path": "ptx0/
|
|
|
|
|
|
|
|
|
5 |
"force_zeros_for_empty_prompt": true,
|
|
|
|
|
|
|
|
|
6 |
"scheduler": [
|
7 |
"diffusers",
|
8 |
-
"
|
9 |
],
|
10 |
"text_encoder": [
|
11 |
-
|
12 |
-
|
13 |
],
|
14 |
"text_encoder_2": [
|
15 |
-
|
16 |
-
|
17 |
],
|
18 |
"tokenizer": [
|
19 |
"transformers",
|
|
|
1 |
{
|
2 |
"_class_name": "StableDiffusionXLPipeline",
|
3 |
+
"_diffusers_version": "0.29.0.dev0",
|
4 |
+
"_name_or_path": "ptx0/terminus-xl-velocity-v2",
|
5 |
+
"feature_extractor": [
|
6 |
+
null,
|
7 |
+
null
|
8 |
+
],
|
9 |
"force_zeros_for_empty_prompt": true,
|
10 |
+
"image_encoder": [
|
11 |
+
null,
|
12 |
+
null
|
13 |
+
],
|
14 |
"scheduler": [
|
15 |
"diffusers",
|
16 |
+
"EulerDiscreteScheduler"
|
17 |
],
|
18 |
"text_encoder": [
|
19 |
+
null,
|
20 |
+
null
|
21 |
],
|
22 |
"text_encoder_2": [
|
23 |
+
null,
|
24 |
+
null
|
25 |
],
|
26 |
"tokenizer": [
|
27 |
"transformers",
|
scheduler/scheduler_config.json
CHANGED
@@ -1,17 +1,24 @@
|
|
1 |
{
|
2 |
-
"_class_name": "
|
3 |
-
"_diffusers_version": "0.
|
4 |
"beta_end": 0.012,
|
5 |
"beta_schedule": "scaled_linear",
|
6 |
"beta_start": 0.00085,
|
7 |
"clip_sample": false,
|
8 |
"clip_sample_range": 1.0,
|
9 |
"dynamic_thresholding_ratio": 0.995,
|
|
|
10 |
"interpolation_type": "linear",
|
11 |
"num_train_timesteps": 1000,
|
12 |
"prediction_type": "v_prediction",
|
13 |
"rescale_betas_zero_snr": true,
|
14 |
"sample_max_value": 1.0,
|
|
|
|
|
15 |
"skip_prk_steps": true,
|
16 |
-
"
|
|
|
|
|
|
|
|
|
17 |
}
|
|
|
1 |
{
|
2 |
+
"_class_name": "EulerDiscreteScheduler",
|
3 |
+
"_diffusers_version": "0.29.0.dev0",
|
4 |
"beta_end": 0.012,
|
5 |
"beta_schedule": "scaled_linear",
|
6 |
"beta_start": 0.00085,
|
7 |
"clip_sample": false,
|
8 |
"clip_sample_range": 1.0,
|
9 |
"dynamic_thresholding_ratio": 0.995,
|
10 |
+
"final_sigmas_type": "zero",
|
11 |
"interpolation_type": "linear",
|
12 |
"num_train_timesteps": 1000,
|
13 |
"prediction_type": "v_prediction",
|
14 |
"rescale_betas_zero_snr": true,
|
15 |
"sample_max_value": 1.0,
|
16 |
+
"sigma_max": null,
|
17 |
+
"sigma_min": null,
|
18 |
"skip_prk_steps": true,
|
19 |
+
"steps_offset": 0,
|
20 |
+
"timestep_spacing": "trailing",
|
21 |
+
"timestep_type": "discrete",
|
22 |
+
"trained_betas": null,
|
23 |
+
"use_karras_sigmas": false
|
24 |
}
|
tokenizer/tokenizer_config.json
CHANGED
@@ -1,33 +1,30 @@
|
|
1 |
{
|
2 |
"add_prefix_space": false,
|
3 |
-
"
|
4 |
-
"
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
},
|
|
|
11 |
"clean_up_tokenization_spaces": true,
|
12 |
"do_lower_case": true,
|
13 |
-
"eos_token":
|
14 |
-
"__type": "AddedToken",
|
15 |
-
"content": "<|endoftext|>",
|
16 |
-
"lstrip": false,
|
17 |
-
"normalized": true,
|
18 |
-
"rstrip": false,
|
19 |
-
"single_word": false
|
20 |
-
},
|
21 |
"errors": "replace",
|
22 |
"model_max_length": 77,
|
23 |
"pad_token": "<|endoftext|>",
|
24 |
"tokenizer_class": "CLIPTokenizer",
|
25 |
-
"unk_token":
|
26 |
-
"__type": "AddedToken",
|
27 |
-
"content": "<|endoftext|>",
|
28 |
-
"lstrip": false,
|
29 |
-
"normalized": true,
|
30 |
-
"rstrip": false,
|
31 |
-
"single_word": false
|
32 |
-
}
|
33 |
}
|
|
|
1 |
{
|
2 |
"add_prefix_space": false,
|
3 |
+
"added_tokens_decoder": {
|
4 |
+
"49406": {
|
5 |
+
"content": "<|startoftext|>",
|
6 |
+
"lstrip": false,
|
7 |
+
"normalized": true,
|
8 |
+
"rstrip": false,
|
9 |
+
"single_word": false,
|
10 |
+
"special": true
|
11 |
+
},
|
12 |
+
"49407": {
|
13 |
+
"content": "<|endoftext|>",
|
14 |
+
"lstrip": false,
|
15 |
+
"normalized": true,
|
16 |
+
"rstrip": false,
|
17 |
+
"single_word": false,
|
18 |
+
"special": true
|
19 |
+
}
|
20 |
},
|
21 |
+
"bos_token": "<|startoftext|>",
|
22 |
"clean_up_tokenization_spaces": true,
|
23 |
"do_lower_case": true,
|
24 |
+
"eos_token": "<|endoftext|>",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
25 |
"errors": "replace",
|
26 |
"model_max_length": 77,
|
27 |
"pad_token": "<|endoftext|>",
|
28 |
"tokenizer_class": "CLIPTokenizer",
|
29 |
+
"unk_token": "<|endoftext|>"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
}
|
tokenizer_2/tokenizer_config.json
CHANGED
@@ -1,33 +1,38 @@
|
|
1 |
{
|
2 |
"add_prefix_space": false,
|
3 |
-
"
|
4 |
-
"
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
},
|
|
|
11 |
"clean_up_tokenization_spaces": true,
|
12 |
"do_lower_case": true,
|
13 |
-
"eos_token":
|
14 |
-
"__type": "AddedToken",
|
15 |
-
"content": "<|endoftext|>",
|
16 |
-
"lstrip": false,
|
17 |
-
"normalized": true,
|
18 |
-
"rstrip": false,
|
19 |
-
"single_word": false
|
20 |
-
},
|
21 |
"errors": "replace",
|
22 |
"model_max_length": 77,
|
23 |
"pad_token": "!",
|
24 |
"tokenizer_class": "CLIPTokenizer",
|
25 |
-
"unk_token":
|
26 |
-
"__type": "AddedToken",
|
27 |
-
"content": "<|endoftext|>",
|
28 |
-
"lstrip": false,
|
29 |
-
"normalized": true,
|
30 |
-
"rstrip": false,
|
31 |
-
"single_word": false
|
32 |
-
}
|
33 |
}
|
|
|
1 |
{
|
2 |
"add_prefix_space": false,
|
3 |
+
"added_tokens_decoder": {
|
4 |
+
"0": {
|
5 |
+
"content": "!",
|
6 |
+
"lstrip": false,
|
7 |
+
"normalized": false,
|
8 |
+
"rstrip": false,
|
9 |
+
"single_word": false,
|
10 |
+
"special": true
|
11 |
+
},
|
12 |
+
"49406": {
|
13 |
+
"content": "<|startoftext|>",
|
14 |
+
"lstrip": false,
|
15 |
+
"normalized": true,
|
16 |
+
"rstrip": false,
|
17 |
+
"single_word": false,
|
18 |
+
"special": true
|
19 |
+
},
|
20 |
+
"49407": {
|
21 |
+
"content": "<|endoftext|>",
|
22 |
+
"lstrip": false,
|
23 |
+
"normalized": true,
|
24 |
+
"rstrip": false,
|
25 |
+
"single_word": false,
|
26 |
+
"special": true
|
27 |
+
}
|
28 |
},
|
29 |
+
"bos_token": "<|startoftext|>",
|
30 |
"clean_up_tokenization_spaces": true,
|
31 |
"do_lower_case": true,
|
32 |
+
"eos_token": "<|endoftext|>",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
33 |
"errors": "replace",
|
34 |
"model_max_length": 77,
|
35 |
"pad_token": "!",
|
36 |
"tokenizer_class": "CLIPTokenizer",
|
37 |
+
"unk_token": "<|endoftext|>"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
}
|
unet/config.json
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
{
|
2 |
"_class_name": "UNet2DConditionModel",
|
3 |
"_diffusers_version": "0.29.0.dev0",
|
4 |
-
"_name_or_path": "
|
5 |
"act_fn": "silu",
|
6 |
"addition_embed_type": "text_time",
|
7 |
"addition_embed_type_num_heads": 64,
|
|
|
1 |
{
|
2 |
"_class_name": "UNet2DConditionModel",
|
3 |
"_diffusers_version": "0.29.0.dev0",
|
4 |
+
"_name_or_path": "/home/user/training/lite-models/checkpoint-400",
|
5 |
"act_fn": "silu",
|
6 |
"addition_embed_type": "text_time",
|
7 |
"addition_embed_type_num_heads": 64,
|
vae/config.json
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
{
|
2 |
"_class_name": "AutoencoderKL",
|
3 |
-
"_diffusers_version": "0.
|
4 |
"_name_or_path": "madebyollin/sdxl-vae-fp16-fix",
|
5 |
"act_fn": "silu",
|
6 |
"block_out_channels": [
|
@@ -18,6 +18,8 @@
|
|
18 |
"force_upcast": false,
|
19 |
"in_channels": 3,
|
20 |
"latent_channels": 4,
|
|
|
|
|
21 |
"layers_per_block": 2,
|
22 |
"norm_num_groups": 32,
|
23 |
"out_channels": 3,
|
|
|
1 |
{
|
2 |
"_class_name": "AutoencoderKL",
|
3 |
+
"_diffusers_version": "0.29.0.dev0",
|
4 |
"_name_or_path": "madebyollin/sdxl-vae-fp16-fix",
|
5 |
"act_fn": "silu",
|
6 |
"block_out_channels": [
|
|
|
18 |
"force_upcast": false,
|
19 |
"in_channels": 3,
|
20 |
"latent_channels": 4,
|
21 |
+
"latents_mean": null,
|
22 |
+
"latents_std": null,
|
23 |
"layers_per_block": 2,
|
24 |
"norm_num_groups": 32,
|
25 |
"out_channels": 3,
|