|
accelerate_device: cuda:4 |
|
accelerate_local_process_index: 0 |
|
accelerate_num_processes: 1 |
|
accelerate_optimizer_step_was_skipped: false |
|
accelerate_process_index: 0 |
|
accelerate_sync_gradients: true |
|
accelerate_use_distributed: false |
|
adam_beta1: 0.9 |
|
adam_beta2: 0.999 |
|
adam_epsilon: 1.0e-08 |
|
adam_weight_decay: 0.01 |
|
allow_tf32: false |
|
checkpointing_steps: 500 |
|
checkpoints_total_limit: null |
|
disable_tqdm: false |
|
enable_xformers_memory_efficient_attention: true |
|
eval_dataset_path: null |
|
fp16_backend: auto |
|
fp16_opt_level: O1 |
|
gradient_accumulation_steps: 1 |
|
gradient_checkpointing: true |
|
hub_model_id: lora_tb |
|
hub_token: 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 |
|
input_perturbation: 0 |
|
learning_rate: 8.0e-06 |
|
logging_dir: log |
|
lr_scheduler: cosine |
|
lr_warmup_steps: 0 |
|
max_grad_norm: 1 |
|
max_train_samples: null |
|
max_train_steps: 100 |
|
mixed_precision: fp16 |
|
noise_offset: 1 |
|
num_train_epochs: 2 |
|
num_validation_images: 4 |
|
output_dir: /home/deepanshu/lmflowdiff/Model |
|
push_to_hub: true |
|
report_to: tensorboard |
|
resume_from_checkpoint: null |
|
save_steps: 1000 |
|
save_total_limit: null |
|
scale_lr: false |
|
seed: 123 |
|
snr_gamma: null |
|
tracker_project_name: text2image-fine-tune |
|
train_batch_size: 1 |
|
use_8bit_adam: false |
|
validation_epochs: 10 |
|
|