File size: 1,269 Bytes
cda9c65
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
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-schnell.safetensors" \
  --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 \--sample_prompts="/app/fluxgym/outputs/another-lora/sample_prompts.txt" --sample_every_n_steps="2" \
  --learning_rate 8e-4 \
  --cache_text_encoder_outputs \
  --cache_text_encoder_outputs_to_disk \
  --fp8_base \
  --highvram \
  --max_train_epochs 1 \
  --save_every_n_epochs 4 \
  --dataset_config "/app/fluxgym/outputs/another-lora/dataset.toml" \
  --output_dir "/app/fluxgym/outputs/another-lora" \
  --output_name another-lora \
  --timestep_sampling shift \
  --discrete_flow_shift 3.1582 \
  --model_prediction_type raw \
  --guidance_scale 1 \
  --loss_type l2 \