Instructions to use noguchis/viewdiff_train_teddybear with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use noguchis/viewdiff_train_teddybear with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("noguchis/viewdiff_train_teddybear", torch_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
| { | |
| "dataset_args": { | |
| "remove_empty_masks": false, | |
| "load_point_clouds": false, | |
| "load_depths": false, | |
| "load_depth_masks": false, | |
| "load_masks": false, | |
| "box_crop": false, | |
| "image_width": null, | |
| "image_height": null, | |
| "pick_sequence": [], | |
| "exclude_sequence": [], | |
| "n_frames_per_sequence": -1 | |
| }, | |
| "batch": { | |
| "n_parallel_images": 3, | |
| "image_width": 256, | |
| "image_height": 256, | |
| "other_selection": "mix", | |
| "other_selection_frame_indices": [], | |
| "sequence_offset": 1, | |
| "crop": "random", | |
| "mask_foreground": false, | |
| "prompt": "Editorial Style Photo, ${category}, 4k --ar 16:9", | |
| "use_blip_prompt": true, | |
| "load_recentered": true, | |
| "replace_pose_with_spherical_start_phi": -400.0, | |
| "replace_pose_with_spherical_end_phi": 360.0, | |
| "replace_pose_with_spherical_phi_endpoint": false, | |
| "replace_pose_with_spherical_radius": 4.0, | |
| "replace_pose_with_spherical_theta": 45.0 | |
| }, | |
| "co3d_root": "/mnt/data/shoji_noguchi/venv/viewdiff/ViewDiff/co3d/dataset", | |
| "category": "teddybear", | |
| "subset": null, | |
| "split": null, | |
| "max_sequences": 50, | |
| "seed": 42, | |
| "training": { | |
| "validation_epochs": 1, | |
| "train_batch_size": 1, | |
| "num_train_epochs": 1000, | |
| "max_train_steps": null, | |
| "dataloader_num_workers": 0, | |
| "local_rank": -1, | |
| "mixed_precision": "no", | |
| "noise_prediction_type": "epsilon", | |
| "remove_cfa_skip_connections_at_iter": -1, | |
| "changed_cfa_last_layer": "no_residual_connection", | |
| "dreambooth_prior_preservation_loss_weight": 0.1, | |
| "dreambooth_prior_preservation_every_nth": 1, | |
| "prob_images_not_noisy": 0.25, | |
| "max_num_images_not_noisy": 2 | |
| }, | |
| "optimizer": { | |
| "learning_rate": 5e-05, | |
| "vol_rend_learning_rate": 0.001, | |
| "vol_rend_adam_weight_decay": 0.0, | |
| "scale_lr": false, | |
| "lr_scheduler": "constant", | |
| "lr_warmup_steps": 500, | |
| "use_8bit_adam": false, | |
| "allow_tf32": false, | |
| "adam_beta1": 0.9, | |
| "adam_beta2": 0.999, | |
| "adam_weight_decay": 0.01, | |
| "adam_epsilon": 1e-08, | |
| "max_grad_norm": 0.005, | |
| "gradient_accumulation_steps": 1, | |
| "only_train_new_layers": false | |
| }, | |
| "model": { | |
| "n_input_images": 3, | |
| "pose_cond_mode": "sa-ca", | |
| "pose_cond_coord_space": "absolute", | |
| "pose_cond_lora_rank": 64, | |
| "pose_cond_dim": 10, | |
| "conditioning_dropout_prob": 0.1, | |
| "use_ema": false, | |
| "enable_xformers_memory_efficient_attention": false, | |
| "gradient_checkpointing": false | |
| }, | |
| "cross_frame_attention": { | |
| "mode": "pretrained", | |
| "n_cfa_down_blocks": 0, | |
| "n_cfa_up_blocks": 0, | |
| "no_cfa_in_mid_block": false, | |
| "to_k_other_frames": 2, | |
| "with_self_attention": true, | |
| "random_others": true, | |
| "last_layer_mode": "no_residual_connection", | |
| "unproj_reproj_mode": "only_unproj_reproj", | |
| "num_3d_layers": 1, | |
| "dim_3d_latent": 16, | |
| "dim_3d_grid": 64, | |
| "vol_rend_proj_in_mode": "multiple", | |
| "vol_rend_proj_out_mode": "multiple", | |
| "vol_rend_aggregator_mode": "ibrnet", | |
| "vol_rend_model_background": true, | |
| "vol_rend_background_grid_percentage": 0.5, | |
| "vol_rend_disparity_at_inf": 0.5, | |
| "n_novel_images": 1, | |
| "use_temb_cond": true | |
| }, | |
| "io": { | |
| "save": { | |
| "image_grids": false, | |
| "pred_files": true, | |
| "pred_video": true, | |
| "pred_gif": false, | |
| "denoise_files": false, | |
| "denoise_video": false, | |
| "cams": true, | |
| "prompts": true, | |
| "rendered_depth": false, | |
| "cond_files": false, | |
| "image_metrics": true | |
| }, | |
| "pretrained_model_name_or_path": "stabilityai/stable-diffusion-2-1-base", | |
| "revision": null, | |
| "output_dir": "outputs/train/teddybear/50_sequences/subset_all/input_3/train/train_teddybear", | |
| "experiment_name": "train_teddybear", | |
| "logging_dir": "logs", | |
| "log_images_every_nth": 500, | |
| "report_to": "custom_tensorboard", | |
| "checkpointing_steps": 500, | |
| "checkpoints_total_limit": 2, | |
| "resume_from_checkpoint": null, | |
| "automatic_checkpoint_resume": false | |
| }, | |
| "selected_sequences": { | |
| "train": [ | |
| "391_46974_93640", | |
| "373_41560_82975", | |
| "46_2596_7562", | |
| "393_47968_95188", | |
| "534_77570_150656", | |
| "444_63772_125902", | |
| "38_1675_5008", | |
| "457_64379_126856", | |
| "42_2104_6430", | |
| "387_46466_92678", | |
| "392_47802_94952", | |
| "353_37579_70740", | |
| "380_44886_89716", | |
| "598_91959_182995", | |
| "392_47443_94381", | |
| "38_1686_5058", | |
| "540_78977_153006", | |
| "46_2595_7548", | |
| "392_47834_94937", | |
| "612_97540_195796", | |
| "387_46699_92990", | |
| "394_48750_96460", | |
| "341_35542_65393", | |
| "391_46884_93389", | |
| "246_26304_51384", | |
| "394_48525_95999", | |
| "482_68882_133595", | |
| "392_47808_94891", | |
| "353_37349_70281", | |
| "387_46523_92748", | |
| "616_99318_199096", | |
| "394_48514_95986", | |
| "387_46530_92758", | |
| "575_84513_166963", | |
| "472_66680_130870", | |
| "537_78247_152118", | |
| "378_44012_87822", | |
| "392_47776_94858", | |
| "377_43705_87007", | |
| "392_47498_94471", | |
| "612_97904_197160", | |
| "604_93996_187948", | |
| "392_47757_94835", | |
| "620_101667_204118", | |
| "504_72676_141358", | |
| "620_101591_203905", | |
| "475_67237_131972", | |
| "444_63708_125709", | |
| "377_43631_86945", | |
| "579_85828_169718" | |
| ], | |
| "val": [ | |
| "581_86292_171292" | |
| ] | |
| } | |
| } |