File size: 1,301 Bytes
cd65b87 |
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 |
accelerate launch \
--mixed_precision bf16 \
--num_cpu_threads_per_process 1 \
sd-scripts/flux_train_network.py \
--pretrained_model_name_or_path "/app/fluxgym/models/unet/flux1-dev.sft" \
--clip_l "/app/fluxgym/models/clip/clip_l.safetensors" \
--t5xxl "/app/fluxgym/models/clip/t5xxl_fp16.safetensors" \
--ae "/app/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 \
--learning_rate 8e-4 \
--cache_text_encoder_outputs \
--cache_text_encoder_outputs_to_disk \
--fp8_base \
--highvram \
--max_train_epochs 20 \
--save_every_n_epochs 4 \
--dataset_config "/app/fluxgym/outputs/dayan/dataset.toml" \
--output_dir "/app/fluxgym/outputs/dayan" \
--output_name dayan \
--timestep_sampling shift \
--discrete_flow_shift 3.1582 \
--model_prediction_type raw \
--guidance_scale 1 \
--loss_type l2 \
--metadata_author wei \
--metadata_tags dayan \
--multires_noise_iterations 8 \
--skip_until_initial_step \
--use_lion_optimizer \
--v_parameterization |