File size: 3,521 Bytes
e593b16
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
# Sylvia Ritter. AKA: by silvery trait

# variable values
$pretrained_model_name_or_path = "D:\models\v1-5-pruned-mse-vae.ckpt"
$train_dir = "D:\models\blakecreates"
$output_dir = "D:\models\blakecreates\diffusers_fine_tuned_model_bucket"
$training_folder = "dataset\192_cables retake artstyle"

$learning_rate = 1e-6
$dataset_repeats = 192
$train_batch_size = 4
$epoch = 1
$save_every_n_epochs=1
$mixed_precision="fp16"
$num_cpu_threads_per_process=6
$lr_scheduler="constant"

$max_resolution = "576,704"

# You should not have to change values past this point
$data_dir = $train_dir + "\" + $training_folder
# Get smalest image resolution to prevent upscaling to that
$shellApp = New-Object -ComObject 'shell.application'
$folderNamespace = $shellApp.Namespace($data_dir)
$get_smalest_resolution = Get-ChildItem $data_dir -Recurse -File -Include *.png, *.jpg, *.webp |
ForEach-Object {
    $image = $folderNamespace.ParseName($_.Name)

    if ($folderNamespace.GetDetailsOf($image, 31) -match '(?<width>\d+) x (?<height>\d+)') {
        [PsCustomObject]@{
            Image           = $_.FullName
            Width           = $Matches.width
            Height          = $Matches.height
            PixelResolution = [int]($Matches.width) * [int]($Matches.height)
            # IsPortrait = $([int]$Matches.height -gt [int]$Matches.width)
        }
    }
} | Sort-Object -Property PixelResolution | Select-Object -first 1

$width, $height = $max_resolution -split ","

if (([int]$width * [int]$height) -gt $get_smalest_resolution.PixelResolution) {

    $smalest_resolution = $get_smalest_resolution.Width + "," + $get_smalest_resolution.Height
    Write-Output "Smallest resolution of images data found is " $get_smalest_resolution.Image "at $smalest_resolution. Replace image with a larger one, or change resolution to prevent this resolution change."

    $max_resolution = $smalest_resolution
}

# stop script on error
$ErrorActionPreference = "Stop"

# activate venv
cd D:\kohya_ss
.\venv\Scripts\activate

# create caption json file
python D:\kohya_ss\diffusers_fine_tuning\merge_captions_to_metadata.py `
--caption_extention ".caption" $train_dir"\"$training_folder $train_dir"\meta_cap.json"

# create images buckets
python D:\kohya_ss\diffusers_fine_tuning\prepare_buckets_latents.py `
    $train_dir"\"$training_folder `
    $train_dir"\meta_cap.json" `
    $train_dir"\meta_lat.json" `
    $pretrained_model_name_or_path `
    --batch_size 4 --max_resolution $max_resolution --mixed_precision fp16

# Get number of valid images
$image_num = Get-ChildItem "$train_dir\$training_folder" -Recurse -File -Include *.npz | Measure-Object | %{$_.Count}
$repeats = $image_num * $dataset_repeats

# calculate max_train_set
$max_train_set = [Math]::Ceiling($repeats / $train_batch_size * $epoch)


accelerate launch --num_cpu_threads_per_process $num_cpu_threads_per_process D:\kohya_ss\diffusers_fine_tuning\fine_tune.py `
    --pretrained_model_name_or_path=$pretrained_model_name_or_path `
    --in_json $train_dir"\meta_lat.json" `
    --train_data_dir=$train_dir"\"$training_folder `
    --output_dir=$output_dir `
    --train_batch_size=$train_batch_size `
    --dataset_repeats=$dataset_repeats `
    --learning_rate=$learning_rate `
    --max_train_steps=$max_train_set `
    --use_8bit_adam --xformers `
    --mixed_precision=$mixed_precision `
    --save_every_n_epochs=$save_every_n_epochs `
    --lr_scheduler=$lr_scheduler `
    --seed=494481440 `
    --train_text_encoder `
    --save_precision="fp16"