Instructions to use Cloth-splatters/folding-dynamics-gps with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use Cloth-splatters/folding-dynamics-gps with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Cloth-splatters/folding-dynamics-gps", 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
| data_dir: /scratch/project_2015260/data/fold/fold_generated_meshes_seed_1397.h5 | |
| output_dir: /scratch/project_2015260/outputs/dynamics_gps/2026-05-04/16-54-43_6616925 | |
| logging_dir: logs/ | |
| gradient_accumulation_steps: 2 | |
| mixed_precision: bf16 | |
| report_to: wandb | |
| seed: 259 | |
| use_ema: false | |
| gradient_checkpointing: false | |
| scale_lr: false | |
| allow_tf32: true | |
| adam_beta1: 0.9 | |
| adam_beta2: 0.999 | |
| adam_weight_decay: 0.01 | |
| adam_epsilon: 1.0e-08 | |
| max_grad_norm: 1.0 | |
| per_gpu_batch_size: 256 | |
| learning_rate: 0.0003 | |
| num_workers: 8 | |
| compile_model: false | |
| lr_scheduler: cosine | |
| lr_warmup_steps: 2000 | |
| max_train_steps: 900000 | |
| resume_from_checkpoint: null | |
| pretrained_model_name_or_path: null | |
| checkpointing_steps: 2000 | |
| sampling_steps: 1000 | |
| checkpoints_total_limit: 10 | |
| exp_name: GPSDynamics_d256_H8_F5_L8_lr0.0003_bs256_fold_04052026_1656 | |
| num_train_epochs: 999999 | |
| config: configs/train_dynamics_gps.yaml | |
| do_classifier_free_guidance: false | |
| val_inference_steps: 50 | |
| model_cfg: | |
| type: GPSDynamicsModel | |
| num_prev_frames: 3 | |
| num_next_frames: 5 | |
| action_dim: 3 | |
| hidden_dim: 256 | |
| num_layers: 8 | |
| num_heads: 8 | |
| edge_feature_dim: 4 | |
| max_num_nodes: 160 | |
| dropout: 0.1 | |
| num_embeds_ada_norm: 1000 | |
| diffusion_cfg: | |
| type: DDPM_DynamicsGPS | |
| num_train_timesteps: 1000 | |
| beta_schedule: scaled_linear | |
| snr_gamma: 5.0 | |
| history_noise_std: 0.0001 | |
| wandb_cfg: | |
| project_name: cloth_sim_mujoco | |
| entity: null | |
| tags: | |
| - gps | |
| - dynamics | |
| dataset_cfg: | |
| type: ClothDynamicsVariableDataset | |
| data_dir: /scratch/project_2015260/data/fold/fold_generated_meshes_seed_1397.h5 | |
| num_prev_frames: 3 | |
| num_next_frames: 5 | |
| max_num_nodes: 160 | |
| max_num_edges: 832 | |
| val_trajectory_ratio: 0.1 | |