Please help to find best "Dreambooth Lora" parameters for train SD model with 15 images of face 🙏

[hosein-moayedi]

Hey everybody! 😇 I am working on find best parameters for training my stable diffusion model with my face in different ways with 16 VRAM. After about 2.5 months, finally I found best parameters for train model using by Dreambooth this version (https://github.com/ShivamShrirao/diffusers/blob/main/examples/dreambooth/train_dreambooth.py) but the main problem is size of final result that is same as a base model (about 2.5 Gig) and that is very huge. 😕 Now I am working on Kohya Dreambooth Lora training (https://github.com/bmaltais/kohya_ss) and finding best parameters for that. I watch these tutorials and try that parameters but the version of kohya_ss that used in these videos is for about 3 or 4 months ago and when I using from that parameters that mentioned in videos, that has not good results. 😔

is there any guys for help me for find best parameters for train my SD model with just 13 images of my face?

tutorials: https://www.youtube.com/watch?v=k5imq01uvUY&t=1860s https://www.youtube.com/watch?v=70H03cv57-o https://www.youtube.com/watch?v=TpuDOsuKIBo https://www.youtube.com/watch?v=3uzCNrQao3o

This is my last Dreambooth Lora params that I am getting better result but that is not enought!

Note: (Max Training Step is 1500)

{ "pretrained_model_name_or_path": "/home/hoseinmoayedi98/perfect-ai/base_model/deliberate_v2.ckpt", "v2": false, "v_parameterization": false, "logging_dir": "/home/hoseinmoayedi98/perfect-ai/training_data/log", "train_data_dir": "/home/hoseinmoayedi98/perfect-ai/training_data/img", "reg_data_dir": "", "output_dir": "/home/hoseinmoayedi98/perfect-ai/stable-diffusion-webui/models/Lora/", "max_resolution": "512,512", "learning_rate": 1e-05, "lr_scheduler": "constant", "lr_warmup": 0, "train_batch_size": 1, "epoch": 1, "save_every_n_epochs": 1, "mixed_precision": "fp16", "save_precision": "fp16", "seed": "1234", "num_cpu_threads_per_process": 4, "cache_latents": true, "cache_latents_to_disk": false, "caption_extension": "", "enable_bucket": false, "gradient_checkpointing": false, "full_fp16": false, "no_token_padding": false, "stop_text_encoder_training": 0, "xformers": false, "save_model_as": "safetensors", "shuffle_caption": false, "save_state": false, "resume": "", "prior_loss_weight": 0.1, "text_encoder_lr": 1e-05, "unet_lr": 1e-05, "network_dim": 128, "lora_network_weights": "", "dim_from_weights": false, "color_aug": false, "flip_aug": false, "clip_skip": 2, "gradient_accumulation_steps": 1.0, "mem_eff_attn": false, "output_name": "trained_lora_v2.0.7", "model_list": "custom", "max_token_length": "75", "max_train_epochs": "", "max_data_loader_n_workers": "1", "network_alpha": 128, "training_comment": "", "keep_tokens": "0", "lr_scheduler_num_cycles": "", "lr_scheduler_power": "", "persistent_data_loader_workers": false, "bucket_no_upscale": true, "random_crop": false, "bucket_reso_steps": 64.0, "caption_dropout_every_n_epochs": 0.0, "caption_dropout_rate": 0, "optimizer": "AdamW8bit", "optimizer_args": "", "noise_offset_type": "Original", "noise_offset": 0, "adaptive_noise_scale": 0, "multires_noise_iterations": 0, "multires_noise_discount": 0, "LoRA_type": "Standard", "conv_dim": 1, "conv_alpha": 1, "sample_every_n_steps": 0, "sample_every_n_epochs": 0, "sample_sampler": "euler_a", "sample_prompts": "", "additional_parameters": "", "vae_batch_size": 0, "min_snr_gamma": 0, "down_lr_weight": "", "mid_lr_weight": "", "up_lr_weight": "", "block_lr_zero_threshold": "", "block_dims": "", "block_alphas": "", "conv_dims": "", "conv_alphas": "", "weighted_captions": false, "unit": 1, "save_every_n_steps": 0, "save_last_n_steps": 0, "save_last_n_steps_state": 0, "use_wandb": false, "wandb_api_key": "", "scale_v_pred_loss_like_noise_pred": false, "scale_weight_norms": 0, "network_dropout": 0, "rank_dropout": 0, "module_dropout": 0 }

[hosein-moayedi]

Please help me @FurkanGozukara @bmaltais @sukhysall @Elldreth

[FurkanGozukara]

Please help me @FurkanGozukara @bmaltais @sukhysall @Elldreth

you already saw my tutorial

i am yet to test more parameters but i have a new workflow in dreambooth extension

hopefully will publish a new video for that

that can possibly help you

[hosein-moayedi]

Thanks @FurkanGozukara ! I am waiting for your video man 😇 I am launching my project, so I very need that

Thanks

[hosein-moayedi]

@FurkanGozukara I have another question that I think you know the answer. I am not very familiar with the structure of Stable Diffusion models and DreamBooth of training but I have a idea that I don't know is it worked or no. You have a base SD model called stable-diffusion-1.5.safetensors and when you trained this model by Dreambooth with your images you will have a new stable-diffusion-1.5 model that trained by your face and we saved that with this name: sd_trained.safetensors file and that has also same size like 2.5 gig.

I want to know, is there any ways for compare and save the main differences between base model and trained model in all important elements like weights or etc.. ?

and when you want access to trained model, you can make it by apply your differences that you saved on base model.

is it possible? (Base Model) & (Your Images) ----- Dreambooth -----> (Trained Model)

compare (Base Model) with (Trained Model) -----> (Save Differences)

for access to trained model: Base Model --- (Apply Differences) ---> Trained Model