Add preset, fix make_caption library path

finetune/make_captions.py

@@ -12,7 +12,7 @@ import numpy as np
 import torch
 from torchvision import transforms
 from torchvision.transforms.functional import InterpolationMode
-sys.path.append(os.path.dirname(__file__))
+sys.path.append(os.path.join(os.path.dirname(__file__), '..')) # sys.path.append(os.path.dirname(__file__))
 from blip.blip import blip_decoder
 import library.train_util as train_util
 

presets/lora/SDXL - LoRA dogu_cat v1.0.json

@@ -0,0 +1,97 @@
+{
+    "LoRA_type": "Standard",
+    "adaptive_noise_scale": 0.00357,
+    "additional_parameters": "",
+    "block_alphas": "",
+    "block_dims": "",
+    "block_lr_zero_threshold": "",
+    "bucket_no_upscale": false,
+    "bucket_reso_steps": 32,
+    "cache_latents": true,
+    "cache_latents_to_disk": true,
+    "caption_dropout_every_n_epochs": 0.0,
+    "caption_dropout_rate": 0,
+    "caption_extension": ".txt",
+    "clip_skip": "1",
+    "color_aug": false,
+    "conv_alpha": 4,
+    "conv_block_alphas": "",
+    "conv_block_dims": "",
+    "conv_dim": 4,
+    "decompose_both": false,
+    "dim_from_weights": false,
+    "down_lr_weight": "",
+    "enable_bucket": true,
+    "epoch": 10,
+    "factor": -1,
+    "flip_aug": false,
+    "full_bf16": false,
+    "full_fp16": false,
+    "gradient_accumulation_steps": 1,
+    "gradient_checkpointing": true,
+    "keep_tokens": 1,
+    "learning_rate": 0.0001,
+    "lora_network_weights": "",
+    "lr_scheduler": "cosine",
+    "lr_scheduler_args": "",
+    "lr_scheduler_num_cycles": "1",
+    "lr_scheduler_power": "",
+    "lr_warmup": 0,
+    "max_bucket_reso": 2048,
+    "max_data_loader_n_workers": "0",
+    "max_resolution": "1024,1024",
+    "max_timestep": 1000,
+    "max_token_length": "75",
+    "max_train_epochs": "",
+    "max_train_steps": "",
+    "mem_eff_attn": false,
+    "mid_lr_weight": "",
+    "min_bucket_reso": 64,
+    "min_snr_gamma": 0,
+    "min_timestep": 0,
+    "mixed_precision": "bf16",
+    "module_dropout": 0,
+    "multires_noise_discount": 0,
+    "multires_noise_iterations": 0,
+    "network_alpha": 46,
+    "network_dim": 92,
+    "network_dropout": 0,
+    "no_token_padding": false,
+    "noise_offset": 0.0357,
+    "noise_offset_type": "Original",
+    "num_cpu_threads_per_process": 2,
+    "optimizer": "AdamW8bit",
+    "optimizer_args": "",
+    "persistent_data_loader_workers": false,
+    "prior_loss_weight": 1.0,
+    "random_crop": false,
+    "rank_dropout": 0,
+    "save_every_n_epochs": 2,
+    "save_every_n_steps": 0,
+    "save_last_n_steps": 0,
+    "save_last_n_steps_state": 0,
+    "save_precision": "bf16",
+    "scale_v_pred_loss_like_noise_pred": false,
+    "scale_weight_norms": 1,
+    "sdxl": true,
+    "sdxl_cache_text_encoder_outputs": false,
+    "sdxl_no_half_vae": true,
+    "seed": "12345",
+    "shuffle_caption": false,
+    "stop_text_encoder_training_pct": 0,
+    "text_encoder_lr": 0.0001,
+    "train_batch_size": 1,
+    "train_on_input": false,
+    "training_comment": "20 1024x1024 cropped images, 20 repeats, Blip captions, but 'woman' replaced with 'khls woman'",
+    "unet_lr": 0.0001,
+    "unit": 1,
+    "up_lr_weight": "",
+    "use_cp": false,
+    "use_wandb": false,
+    "v2": false,
+    "v_parameterization": false,
+    "v_pred_like_loss": 0,
+    "vae_batch_size": 0,
+    "weighted_captions": false,
+    "xformers": "xformers"
+}

tools/create_txt_from_images.py

@@ -0,0 +1,32 @@
+import os
+import argparse
+
+def main(folder_path):
+    # Validate if the folder exists
+    if not os.path.exists(folder_path):
+        print("The specified folder does not exist.")
+        return
+    
+    # Loop through all files in the directory
+    for filename in os.listdir(folder_path):
+        # Check if the file is an image file (webp, jpg, png)
+        if filename.lower().endswith(('.webp', '.jpg', '.png')):
+            # Remove the file extension from the filename
+            name_without_extension = os.path.splitext(filename)[0]
+            
+            # Construct the name of the txt file
+            txt_filename = f"{name_without_extension}.txt"
+            
+            # Extract the content before the underscore
+            content = name_without_extension.split("_")[0]
+            
+            # Write the content to the txt file
+            with open(os.path.join(folder_path, txt_filename), "w") as txt_file:
+                txt_file.write(content)
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description='Process a folder.')
+    parser.add_argument('folder_path', type=str, help='Path to the folder to process')
+    
+    args = parser.parse_args()
+    main(args.folder_path)