| accelerate launch ^ | |
| --mixed_precision bf16 ^ | |
| --num_cpu_threads_per_process 1 ^ | |
| sd-scripts/flux_train_network.py ^ | |
| --pretrained_model_name_or_path "C:\Users\marko\Downloads\FLUXGYM\fluxgym\models\unet\flux1-dev.sft" ^ | |
| --clip_l "C:\Users\marko\Downloads\FLUXGYM\fluxgym\models\clip\clip_l.safetensors" ^ | |
| --t5xxl "C:\Users\marko\Downloads\FLUXGYM\fluxgym\models\clip\t5xxl_fp16.safetensors" ^ | |
| --ae "C:\Users\marko\Downloads\FLUXGYM\fluxgym\models\vae\ae.sft" ^ | |
| --cache_latents_to_disk ^ | |
| --save_model_as safetensors ^ | |
| --sdpa --persistent_data_loader_workers ^ | |
| --max_data_loader_n_workers 2 ^ | |
| --seed 42 ^ | |
| --gradient_checkpointing ^ | |
| --mixed_precision bf16 ^ | |
| --save_precision bf16 ^ | |
| --network_module networks.lora_flux ^ | |
| --network_dim 4 ^ | |
| --optimizer_type adamw8bit ^--sample_prompts="C:\Users\marko\Downloads\FLUXGYM\fluxgym\outputs\becc\sample_prompts.txt" --sample_every_n_steps="500" ^ | |
| --learning_rate 8e-4 ^ | |
| --cache_text_encoder_outputs ^ | |
| --cache_text_encoder_outputs_to_disk ^ | |
| --fp8_base ^ | |
| --highvram ^ | |
| --max_train_epochs 16 ^ | |
| --save_every_n_epochs 4 ^ | |
| --dataset_config "C:\Users\marko\Downloads\FLUXGYM\fluxgym\outputs\becc\dataset.toml" ^ | |
| --output_dir "C:\Users\marko\Downloads\FLUXGYM\fluxgym\outputs\becc" ^ | |
| --output_name becc ^ | |
| --timestep_sampling shift ^ | |
| --discrete_flow_shift 3.1582 ^ | |
| --model_prediction_type raw ^ | |
| --guidance_scale 1 ^ | |
| --loss_type l2 ^ |