ai-toolkit_config.yaml

---
job: extension
config:
  # this name will be the folder and filename name
  name: "my_first_flux_lora_v1"
  process:
    - type: 'sd_trainer'
      # root folder to save training sessions/samples/weights
      training_folder: "output"
      # uncomment to see performance stats in the terminal every N steps

      #      performance_log_every: 1000
      device: cuda:0
      # if a trigger word is specified, it will be added to captions of training data if it does not already exist
      # alternatively, in your captions you can add [trigger] and it will be replaced with the trigger word

      #      trigger_word: "p3r5on"
      network:
        type: "lora"
        linear: 16
        linear_alpha: 16
      save:
        dtype: float16 # precision to save
        save_every: 1000 # save every this many steps
        max_step_saves_to_keep: 4 # how many intermittent saves to keep
        push_to_hub: false #change this to True to push your trained model to Hugging Face.
        # You can either set up a HF_TOKEN env variable or you'll be prompted to log-in         
      #       hf_repo_id: your-username/your-model-slug
      #       hf_private: true #whether the repo is private or public

      datasets:
        # datasets are a folder of images. captions need to be txt files with the same name as the image
        # for instance image2.jpg and image2.txt. Only jpg, jpeg, and png are supported currently
        # images will automatically be resized and bucketed into the resolution specified
        # on windows, escape back slashes with another backslash so
        # "C:\\path\\to\\images\\folder"
        - folder_path: "/workspace/ai-toolkit/images"
          caption_ext: "txt"
          caption_dropout_rate: 0.05 # will drop out the caption 5% of time
          shuffle_tokens: false # shuffle caption order, split by commas
          cache_latents_to_disk: true # leave this true unless you know what you're doing
          resolution: [512, 768, 1024] # flux enjoys multiple resolutions
      train:
        batch_size: 1
        steps: 3000 # total number of steps to train 500 - 4000 is a good range
        gradient_accumulation_steps: 1
        train_unet: true
        train_text_encoder: false # probably won't work with flux
        gradient_checkpointing: true # need the on unless you have a ton of vram
        noise_scheduler: "flowmatch" # for training only
        optimizer: "adamw8bit"
        lr: 0.0001
        #        skip_first_sample: true
        # uncomment to completely disable sampling
        #        disable_sampling: true
        # uncomment to use new vell curved weighting. Experimental but may produce better results
        #        linear_timesteps: true

        # ema will smooth out learning, but could slow it down. Recommended to leave on.
        ema_config:
          use_ema: true
          ema_decay: 0.99
        # will probably need this if gpu supports it for flux, other dtypes may not work correctly
        dtype: bf16
      model:
        # huggingface model name or path
        name_or_path: "black-forest-labs/FLUX.1-dev"
        is_flux: true
        quantize: true # run 8bit mixed precision
        #        low_vram: true  # uncomment this if the GPU is connected to your monitors. It will use less vram to quantize, but is slower.
      sample:
        sampler: "flowmatch" # must match train.noise_scheduler
        sample_every: 500 # sample every this many steps
        width: 1024
        height: 1024
        prompts:
          - portrait photograph, a man standing, wearing a TOK black hoodie, background is slightly out of focus, strong color contrast
          - A portrait of a man standing on a bustling city street at night, wearing a TOK black hoodie. Neon lights illuminate the background, creating a strong color contrast, with a slightly out-of-focus backdrop of blurred car lights and urban scenery.
          - A portrait of a man standing on a serene forest path, wearing a TOK black hoodie. The green foliage creates a lush and vibrant backdrop, slightly out of focus, with the dappled sunlight creating a strong color contrast against the dark hoodie.
          - A portrait of a man standing on a modern urban rooftop at sunset, wearing a TOK black hoodie. The background showcases a city skyline with glowing buildings, slightly out of focus, emphasizing the contrast between the warm sunset hues and the dark clothing.
          - A portrait of a man standing inside an industrial warehouse, wearing a TOK black hoodie. The background features metal beams and machinery, with a gritty, slightly out-of-focus look that contrasts sharply with the dark hoodie.
          - A portrait of a man standing at a dimly lit subway station, wearing a TOK black hoodie. The background shows blurred trains and station signs, slightly out of focus, creating a strong color contrast between the artificial lighting and the dark hoodie.
        neg: "" # not used on flux
        seed: 42
        walk_seed: true
        guidance_scale: 4
        sample_steps: 20
      trigger_word: TOK
# you can add any additional meta info here. [name] is replaced with config name at top
meta:
  name: "[name]"
  version: '1.0'