anjakuzev commited on
Commit
150e859
1 Parent(s): 86ed055

End of training

Browse files
Files changed (45) hide show
  1. README.md +27 -0
  2. image_0.png +0 -0
  3. image_1.png +0 -0
  4. image_2.png +0 -0
  5. image_3.png +0 -0
  6. logs/text2image-fine-tune/1698226805.5745053/events.out.tfevents.1698226805.146-235-238-191.296963.1 +3 -0
  7. logs/text2image-fine-tune/1698226805.575728/hparams.yml +50 -0
  8. logs/text2image-fine-tune/1698236900.876514/events.out.tfevents.1698236900.146-235-238-191.348206.1 +3 -0
  9. logs/text2image-fine-tune/1698236900.8777218/hparams.yml +52 -0
  10. logs/text2image-fine-tune/1698238676.22221/events.out.tfevents.1698238676.146-235-238-191.363907.1 +3 -0
  11. logs/text2image-fine-tune/1698238676.2234287/hparams.yml +52 -0
  12. logs/text2image-fine-tune/1698248444.7369337/events.out.tfevents.1698248444.146-235-238-191.390678.1 +3 -0
  13. logs/text2image-fine-tune/1698248444.7381709/hparams.yml +52 -0
  14. logs/text2image-fine-tune/1698311467.1536908/events.out.tfevents.1698311467.146-235-238-191.433134.1 +3 -0
  15. logs/text2image-fine-tune/1698311467.1548967/hparams.yml +52 -0
  16. logs/text2image-fine-tune/1698315228.51079/events.out.tfevents.1698315228.146-235-238-191.470849.1 +3 -0
  17. logs/text2image-fine-tune/1698315228.5121017/hparams.yml +52 -0
  18. logs/text2image-fine-tune/1698315448.953754/events.out.tfevents.1698315448.146-235-238-191.471723.1 +3 -0
  19. logs/text2image-fine-tune/1698315448.9550197/hparams.yml +54 -0
  20. logs/text2image-fine-tune/1698322872.9454908/events.out.tfevents.1698322872.146-235-238-191.496107.1 +3 -0
  21. logs/text2image-fine-tune/1698322872.9467914/hparams.yml +54 -0
  22. logs/text2image-fine-tune/1698326536.4699554/events.out.tfevents.1698326536.146-235-238-191.3154.1 +3 -0
  23. logs/text2image-fine-tune/1698326536.4712646/hparams.yml +54 -0
  24. logs/text2image-fine-tune/1698326821.824955/events.out.tfevents.1698326821.146-235-238-191.1795.1 +3 -0
  25. logs/text2image-fine-tune/1698326821.8261724/hparams.yml +54 -0
  26. logs/text2image-fine-tune/1698390861.2813563/events.out.tfevents.1698390861.146-235-238-191.26646.1 +3 -0
  27. logs/text2image-fine-tune/1698390861.2826254/hparams.yml +54 -0
  28. logs/text2image-fine-tune/1698392957.5421095/events.out.tfevents.1698392957.146-235-238-191.30910.1 +3 -0
  29. logs/text2image-fine-tune/1698392957.5432978/hparams.yml +54 -0
  30. logs/text2image-fine-tune/1698394819.4437282/events.out.tfevents.1698394819.146-235-238-191.34747.1 +3 -0
  31. logs/text2image-fine-tune/1698394819.4448159/hparams.yml +54 -0
  32. logs/text2image-fine-tune/events.out.tfevents.1698226805.146-235-238-191.296963.0 +3 -0
  33. logs/text2image-fine-tune/events.out.tfevents.1698236900.146-235-238-191.348206.0 +3 -0
  34. logs/text2image-fine-tune/events.out.tfevents.1698238676.146-235-238-191.363907.0 +3 -0
  35. logs/text2image-fine-tune/events.out.tfevents.1698248444.146-235-238-191.390678.0 +3 -0
  36. logs/text2image-fine-tune/events.out.tfevents.1698311467.146-235-238-191.433134.0 +3 -0
  37. logs/text2image-fine-tune/events.out.tfevents.1698315228.146-235-238-191.470849.0 +3 -0
  38. logs/text2image-fine-tune/events.out.tfevents.1698315448.146-235-238-191.471723.0 +3 -0
  39. logs/text2image-fine-tune/events.out.tfevents.1698322872.146-235-238-191.496107.0 +3 -0
  40. logs/text2image-fine-tune/events.out.tfevents.1698326536.146-235-238-191.3154.0 +3 -0
  41. logs/text2image-fine-tune/events.out.tfevents.1698326821.146-235-238-191.1795.0 +3 -0
  42. logs/text2image-fine-tune/events.out.tfevents.1698390861.146-235-238-191.26646.0 +3 -0
  43. logs/text2image-fine-tune/events.out.tfevents.1698392957.146-235-238-191.30910.0 +3 -0
  44. logs/text2image-fine-tune/events.out.tfevents.1698394819.146-235-238-191.34747.0 +3 -0
  45. pytorch_lora_weights.safetensors +3 -0
README.md ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ license: creativeml-openrail-m
4
+ base_model: SG161222/RealVisXL_V2.0
5
+ dataset: anjakuzev/michael_scott
6
+ tags:
7
+ - stable-diffusion-xl
8
+ - stable-diffusion-xl-diffusers
9
+ - text-to-image
10
+ - diffusers
11
+ - lora
12
+ inference: true
13
+ ---
14
+
15
+ # LoRA text2image fine-tuning - anjakuzev/michael_scott_v19
16
+
17
+ These are LoRA adaption weights for SG161222/RealVisXL_V2.0. The weights were fine-tuned on the anjakuzev/michael_scott dataset. You can find some example images in the following.
18
+
19
+ ![img_0](./image_0.png)
20
+ ![img_1](./image_1.png)
21
+ ![img_2](./image_2.png)
22
+ ![img_3](./image_3.png)
23
+
24
+
25
+ LoRA for the text encoder was enabled: False.
26
+
27
+ Special VAE used for training: None.
image_0.png ADDED
image_1.png ADDED
image_2.png ADDED
image_3.png ADDED
logs/text2image-fine-tune/1698226805.5745053/events.out.tfevents.1698226805.146-235-238-191.296963.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ce8c4f5b31a84cec97885432705c9773dd8f0dc055221ff195fc6e161484298
3
+ size 2447
logs/text2image-fine-tune/1698226805.575728/hparams.yml ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/harry_styles
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/my-text-to-image-model_8
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 500
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 50
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: runwayml/stable-diffusion-v1-5
35
+ push_to_hub: true
36
+ random_flip: true
37
+ rank: 4
38
+ report_to: tensorboard
39
+ resolution: 512
40
+ resume_from_checkpoint: null
41
+ revision: null
42
+ scale_lr: false
43
+ seed: 42
44
+ snr_gamma: null
45
+ train_batch_size: 1
46
+ train_data_dir: null
47
+ use_8bit_adam: false
48
+ validation_epochs: 1
49
+ validation_prompt: Generate an image of Harry Styles with wavy hair and sharp facial
50
+ features. He should be wearing contemporary formal attire with a confident expression.
logs/text2image-fine-tune/1698236900.876514/events.out.tfevents.1698236900.146-235-238-191.348206.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e4d83bb95e9aee0cc7044f129c8a5b09908559fdcdebfd9d8b448e8eb9ab555
3
+ size 2598
logs/text2image-fine-tune/1698236900.8777218/hparams.yml ADDED
@@ -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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 500
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 56
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: runwayml/stable-diffusion-v1-5
35
+ push_to_hub: true
36
+ random_flip: true
37
+ rank: 4
38
+ report_to: tensorboard
39
+ resolution: 512
40
+ resume_from_checkpoint: null
41
+ revision: null
42
+ scale_lr: false
43
+ seed: 42
44
+ snr_gamma: null
45
+ train_batch_size: 1
46
+ train_data_dir: null
47
+ use_8bit_adam: false
48
+ validation_epochs: 1
49
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
50
+ eyes. He should be wearing professional attire, including a dark suit jacket and
51
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
52
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698238676.22221/events.out.tfevents.1698238676.146-235-238-191.363907.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd75d423570be1533fa29584136148e13190b6ec94f4cb03efa0234eec77c858
3
+ size 2601
logs/text2image-fine-tune/1698238676.2234287/hparams.yml ADDED
@@ -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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v2
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 800
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 89
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: runwayml/stable-diffusion-v1-5
35
+ push_to_hub: true
36
+ random_flip: true
37
+ rank: 4
38
+ report_to: tensorboard
39
+ resolution: 512
40
+ resume_from_checkpoint: null
41
+ revision: null
42
+ scale_lr: false
43
+ seed: 1337
44
+ snr_gamma: null
45
+ train_batch_size: 1
46
+ train_data_dir: null
47
+ use_8bit_adam: false
48
+ validation_epochs: 1
49
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
50
+ eyes. He should be wearing professional attire, including a dark suit jacket and
51
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
52
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698248444.7369337/events.out.tfevents.1698248444.146-235-238-191.390678.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a180c88f7c84d0436b97041e60242b57e0e6fb0af593497b2a589212938a9ecf
3
+ size 2608
logs/text2image-fine-tune/1698248444.7381709/hparams.yml ADDED
@@ -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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v3
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 800
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 89
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: stablediffusionapi/cyberrealistic-v32
35
+ push_to_hub: true
36
+ random_flip: true
37
+ rank: 4
38
+ report_to: tensorboard
39
+ resolution: 512
40
+ resume_from_checkpoint: null
41
+ revision: null
42
+ scale_lr: false
43
+ seed: 1337
44
+ snr_gamma: null
45
+ train_batch_size: 1
46
+ train_data_dir: null
47
+ use_8bit_adam: false
48
+ validation_epochs: 1
49
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
50
+ eyes. He should be wearing professional attire, including a dark suit jacket and
51
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
52
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698311467.1536908/events.out.tfevents.1698311467.146-235-238-191.433134.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4227d7261874ff7b64b77ef6c0ff3c51e4813ac251a3f81828f252f8164cac0
3
+ size 2608
logs/text2image-fine-tune/1698311467.1548967/hparams.yml ADDED
@@ -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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v3
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 15000
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 1667
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: stablediffusionapi/cyberrealistic-v32
35
+ push_to_hub: true
36
+ random_flip: true
37
+ rank: 4
38
+ report_to: tensorboard
39
+ resolution: 512
40
+ resume_from_checkpoint: null
41
+ revision: null
42
+ scale_lr: false
43
+ seed: 1337
44
+ snr_gamma: null
45
+ train_batch_size: 1
46
+ train_data_dir: null
47
+ use_8bit_adam: false
48
+ validation_epochs: 1
49
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
50
+ eyes. He should be wearing professional attire, including a dark suit jacket and
51
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
52
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698315228.51079/events.out.tfevents.1698315228.146-235-238-191.470849.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:795d530ff4db1c285670e4f4e6de51309e3742ee67aa0233cfeeb332ab5e72bd
3
+ size 2594
logs/text2image-fine-tune/1698315228.5121017/hparams.yml ADDED
@@ -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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v3
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 800
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 89
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: SG161222/RealVisXL_V2.0
35
+ push_to_hub: true
36
+ random_flip: true
37
+ rank: 4
38
+ report_to: tensorboard
39
+ resolution: 512
40
+ resume_from_checkpoint: null
41
+ revision: null
42
+ scale_lr: false
43
+ seed: 1337
44
+ snr_gamma: null
45
+ train_batch_size: 1
46
+ train_data_dir: null
47
+ use_8bit_adam: false
48
+ validation_epochs: 1
49
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
50
+ eyes. He should be wearing professional attire, including a dark suit jacket and
51
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
52
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698315448.953754/events.out.tfevents.1698315448.146-235-238-191.471723.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b43ad2f4c377571647da0ea3fec3e78fcb6229819874b1a9a1b7f5670e1648b
3
+ size 2651
logs/text2image-fine-tune/1698315448.9550197/hparams.yml ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v3
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 800
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 89
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: SG161222/RealVisXL_V2.0
35
+ pretrained_vae_model_name_or_path: null
36
+ push_to_hub: true
37
+ random_flip: true
38
+ rank: 4
39
+ report_to: tensorboard
40
+ resolution: 1024
41
+ resume_from_checkpoint: null
42
+ revision: null
43
+ scale_lr: false
44
+ seed: 1337
45
+ snr_gamma: null
46
+ train_batch_size: 1
47
+ train_data_dir: null
48
+ train_text_encoder: false
49
+ use_8bit_adam: false
50
+ validation_epochs: 1
51
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
52
+ eyes. He should be wearing professional attire, including a dark suit jacket and
53
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
54
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698322872.9454908/events.out.tfevents.1698322872.146-235-238-191.496107.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:724932b5538d98c8cb5ca087529e15c857ae7ad64e4a7458b234cae9caa627a5
3
+ size 2651
logs/text2image-fine-tune/1698322872.9467914/hparams.yml ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v3
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 500
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 56
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: SG161222/RealVisXL_V2.0
35
+ pretrained_vae_model_name_or_path: null
36
+ push_to_hub: true
37
+ random_flip: true
38
+ rank: 4
39
+ report_to: tensorboard
40
+ resolution: 1024
41
+ resume_from_checkpoint: null
42
+ revision: null
43
+ scale_lr: false
44
+ seed: 1337
45
+ snr_gamma: null
46
+ train_batch_size: 1
47
+ train_data_dir: null
48
+ train_text_encoder: false
49
+ use_8bit_adam: false
50
+ validation_epochs: 1
51
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
52
+ eyes. He should be wearing professional attire, including a dark suit jacket and
53
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
54
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698326536.4699554/events.out.tfevents.1698326536.146-235-238-191.3154.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b74b80a3ba8809b7a0b874057b1ea42fc2ceca32fa14dfabe2c807caf4bf8707
3
+ size 2652
logs/text2image-fine-tune/1698326536.4712646/hparams.yml ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v10
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 300
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 34
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: SG161222/RealVisXL_V2.0
35
+ pretrained_vae_model_name_or_path: null
36
+ push_to_hub: true
37
+ random_flip: true
38
+ rank: 4
39
+ report_to: tensorboard
40
+ resolution: 1024
41
+ resume_from_checkpoint: null
42
+ revision: null
43
+ scale_lr: false
44
+ seed: 1337
45
+ snr_gamma: null
46
+ train_batch_size: 1
47
+ train_data_dir: null
48
+ train_text_encoder: false
49
+ use_8bit_adam: false
50
+ validation_epochs: 1
51
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
52
+ eyes. He should be wearing professional attire, including a dark suit jacket and
53
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
54
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698326821.824955/events.out.tfevents.1698326821.146-235-238-191.1795.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c97098ce80be92c2ffd33b9851442bb12419e24337fc8f9240b5aca1bc20549
3
+ size 2652
logs/text2image-fine-tune/1698326821.8261724/hparams.yml ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v10
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 300
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 34
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: SG161222/RealVisXL_V2.0
35
+ pretrained_vae_model_name_or_path: null
36
+ push_to_hub: true
37
+ random_flip: true
38
+ rank: 4
39
+ report_to: tensorboard
40
+ resolution: 1024
41
+ resume_from_checkpoint: null
42
+ revision: null
43
+ scale_lr: false
44
+ seed: 1337
45
+ snr_gamma: null
46
+ train_batch_size: 1
47
+ train_data_dir: null
48
+ train_text_encoder: false
49
+ use_8bit_adam: false
50
+ validation_epochs: 1
51
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
52
+ eyes. He should be wearing professional attire, including a dark suit jacket and
53
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
54
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698390861.2813563/events.out.tfevents.1698390861.146-235-238-191.26646.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a0672c9fdba5c27f8d24b9cb0422821caaa89ac5eaf3adb9ff26f991c2f43a7d
3
+ size 2652
logs/text2image-fine-tune/1698390861.2826254/hparams.yml ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v19
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 100
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 12
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: SG161222/RealVisXL_V2.0
35
+ pretrained_vae_model_name_or_path: null
36
+ push_to_hub: true
37
+ random_flip: true
38
+ rank: 4
39
+ report_to: tensorboard
40
+ resolution: 1024
41
+ resume_from_checkpoint: null
42
+ revision: null
43
+ scale_lr: false
44
+ seed: 1337
45
+ snr_gamma: null
46
+ train_batch_size: 1
47
+ train_data_dir: null
48
+ train_text_encoder: false
49
+ use_8bit_adam: false
50
+ validation_epochs: 1
51
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
52
+ eyes. He should be wearing professional attire, including a dark suit jacket and
53
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
54
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698392957.5421095/events.out.tfevents.1698392957.146-235-238-191.30910.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fafae653319f483f5536c5082cd06417cc670080758cd99405cc969887476cdf
3
+ size 2652
logs/text2image-fine-tune/1698392957.5432978/hparams.yml ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v19
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 100
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 12
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: SG161222/RealVisXL_V2.0
35
+ pretrained_vae_model_name_or_path: null
36
+ push_to_hub: true
37
+ random_flip: true
38
+ rank: 4
39
+ report_to: tensorboard
40
+ resolution: 1024
41
+ resume_from_checkpoint: null
42
+ revision: null
43
+ scale_lr: false
44
+ seed: 1337
45
+ snr_gamma: null
46
+ train_batch_size: 1
47
+ train_data_dir: null
48
+ train_text_encoder: false
49
+ use_8bit_adam: false
50
+ validation_epochs: 1
51
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
52
+ eyes. He should be wearing professional attire, including a dark suit jacket and
53
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
54
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/1698394819.4437282/events.out.tfevents.1698394819.146-235-238-191.34747.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ce0fff6a850b55af28c2a6268d0aa224d40ee15de8d13a6d7cf104742019aaaf
3
+ size 2652
logs/text2image-fine-tune/1698394819.4448159/hparams.yml ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ cache_dir: null
7
+ caption_column: additional_feature
8
+ center_crop: true
9
+ checkpointing_steps: 3000
10
+ checkpoints_total_limit: null
11
+ dataloader_num_workers: 8
12
+ dataset_config_name: null
13
+ dataset_name: anjakuzev/michael_scott
14
+ enable_xformers_memory_efficient_attention: false
15
+ gradient_accumulation_steps: 4
16
+ gradient_checkpointing: false
17
+ hub_model_id: anjakuzev/michael_scott_v19
18
+ hub_token: null
19
+ image_column: image
20
+ learning_rate: 0.0001
21
+ local_rank: -1
22
+ logging_dir: logs
23
+ lr_scheduler: cosine
24
+ lr_warmup_steps: 0
25
+ max_grad_norm: 1.0
26
+ max_train_samples: null
27
+ max_train_steps: 100
28
+ mixed_precision: null
29
+ noise_offset: 0
30
+ num_train_epochs: 12
31
+ num_validation_images: 4
32
+ output_dir: diffusers/output
33
+ prediction_type: null
34
+ pretrained_model_name_or_path: SG161222/RealVisXL_V2.0
35
+ pretrained_vae_model_name_or_path: null
36
+ push_to_hub: true
37
+ random_flip: true
38
+ rank: 4
39
+ report_to: tensorboard
40
+ resolution: 1024
41
+ resume_from_checkpoint: null
42
+ revision: null
43
+ scale_lr: false
44
+ seed: 1337
45
+ snr_gamma: null
46
+ train_batch_size: 1
47
+ train_data_dir: null
48
+ train_text_encoder: false
49
+ use_8bit_adam: false
50
+ validation_epochs: 1
51
+ validation_prompt: Generate an image of Michael Scott with short dark hair and deep-set
52
+ eyes. He should be wearing professional attire, including a dark suit jacket and
53
+ tie. His expression should range from jovial and smiling to deep in thought or surprised.
54
+ The setting should hint at a workplace environment, preferably an office.
logs/text2image-fine-tune/events.out.tfevents.1698226805.146-235-238-191.296963.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8065922af2ac88de65a6467d27d647cc097a4ff7851d16dee5de51ab37ef873
3
+ size 568
logs/text2image-fine-tune/events.out.tfevents.1698236900.146-235-238-191.348206.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dedc254ab9261a7eb5042d54765f181d021af5eba3f9a14567f24a0ad71e55c3
3
+ size 77564499
logs/text2image-fine-tune/events.out.tfevents.1698238676.146-235-238-191.363907.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:056b1eba649b13b81ed6e3a7257c8734a1ca8b6f852591457996fdbf01aea40f
3
+ size 121312948
logs/text2image-fine-tune/events.out.tfevents.1698248444.146-235-238-191.390678.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a3a88a811a844f1a447535f994ec24b4459c39a6deaaa925354a28d89af59e1
3
+ size 118339618
logs/text2image-fine-tune/events.out.tfevents.1698311467.146-235-238-191.433134.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7e02c25f2b5a1b47afd03f38f41b4a038aad7eeca70e37558ce194930b30b212
3
+ size 169822254
logs/text2image-fine-tune/events.out.tfevents.1698315228.146-235-238-191.470849.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77b18fe1078a04ddf457f33ed4621c0ed8e785a8c432b1a61756dcbcfc1abfe1
3
+ size 88
logs/text2image-fine-tune/events.out.tfevents.1698315448.146-235-238-191.471723.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8693b0f9e73c4b5ad22051f0fef13633609215efcdcbd41aac616adf91a41e6b
3
+ size 101892195
logs/text2image-fine-tune/events.out.tfevents.1698322872.146-235-238-191.496107.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79d15df5045943f01e0a40d8ad11f7d841595f9fbf5866db6556220e5679b935
3
+ size 88
logs/text2image-fine-tune/events.out.tfevents.1698326536.146-235-238-191.3154.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a05833c18add261d3d6b904b3d3ccf93a456f33225151245184e3bd1c853447
3
+ size 88
logs/text2image-fine-tune/events.out.tfevents.1698326821.146-235-238-191.1795.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e4aa171b0e8b45970123c68cea404ad9830154ecd366a717440631f2fd1419f
3
+ size 91322078
logs/text2image-fine-tune/events.out.tfevents.1698390861.146-235-238-191.26646.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab8dde031cb7225a9b054b4159c32a8a9aa50e7f8157b3cf4ad18d3ffe14205b
3
+ size 36540652
logs/text2image-fine-tune/events.out.tfevents.1698392957.146-235-238-191.30910.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e7a94af1c6a9e6d218158740592f57d675cbe9c1c387040728628bf556fd012
3
+ size 88
logs/text2image-fine-tune/events.out.tfevents.1698394819.146-235-238-191.34747.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:745b2cb8f4437cd033359fd78bdc65ebb27afa92154cc15abb897fed829474b1
3
+ size 39249229
pytorch_lora_weights.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b81bbcbc664e314776cede6fc0c99fd6c6fc74248ce9f4b0834b025028dfbba7
3
+ size 23401064