Instructions to use kmpartner/bk-test with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use kmpartner/bk-test with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("kmpartner/bk-test", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
| adam_beta1: 0.9 | |
| adam_beta2: 0.999 | |
| adam_epsilon: 1.0e-08 | |
| adam_weight_decay: 0.01 | |
| allow_tf32: false | |
| cache_dir: null | |
| center_crop: true | |
| checkpointing_steps: 5000 | |
| checkpoints_total_limit: null | |
| dataloader_num_workers: 0 | |
| dataset_config_name: null | |
| enable_xformers_memory_efficient_attention: false | |
| gradient_accumulation_steps: 4 | |
| gradient_checkpointing: true | |
| lambda_kd_feat: 1.0 | |
| lambda_kd_output: 1.0 | |
| lambda_sd: 1.0 | |
| learning_rate: 5.0e-05 | |
| local_rank: -1 | |
| logging_dir: logs | |
| lr_scheduler: constant | |
| lr_warmup_steps: 0 | |
| max_grad_norm: 1.0 | |
| max_train_samples: null | |
| max_train_steps: 400000 | |
| mixed_precision: fp16 | |
| non_ema_revision: null | |
| num_train_epochs: 4939 | |
| num_valid_images: 2 | |
| output_dir: ./results/v2-base_kd_bk_tiny | |
| pretrained_model_name_or_path: stabilityai/stable-diffusion-2-1-base | |
| random_flip: true | |
| report_to: all | |
| resolution: 512 | |
| resume_from_checkpoint: null | |
| revision: null | |
| scale_lr: false | |
| seed: 1234 | |
| train_batch_size: 32 | |
| train_data_dir: ./data/laion_aes/preprocessed_11k | |
| unet_config_name: bk_tiny | |
| unet_config_path: ./src/unet_config_v2-base | |
| use_8bit_adam: false | |
| use_copy_weight_from_teacher: true | |
| use_ema: true | |
| valid_prompt: a golden vase with different flowers | |
| valid_steps: 500 | |