kelvinzhaozg's picture
Upload policy weights, train config and readme
93b8f79 verified
{
"type": "flow_matching_transformer",
"n_obs_steps": 1,
"normalization_mapping": {
"VISUAL": "MEAN_STD",
"STATE": "MIN_MAX",
"ACTION": "MIN_MAX"
},
"input_features": {
"observation.images.rgb": {
"type": "VISUAL",
"shape": [
3,
240,
424
]
},
"observation.state": {
"type": "STATE",
"shape": [
24
]
}
},
"output_features": {
"action": {
"type": "ACTION",
"shape": [
21
]
}
},
"device": "cuda",
"use_amp": false,
"push_to_hub": true,
"repo_id": "kelvinzhaozg/flow_matching_transformer_digit_third_arm_mujoco_walking",
"private": null,
"tags": null,
"license": null,
"horizon": 64,
"n_action_steps": 32,
"drop_n_last_frames": 7,
"vision_backbone": "resnet18",
"crop_shape": [
224,
224
],
"crop_is_random": true,
"pretrained_backbone_weights": null,
"use_group_norm": true,
"spatial_softmax_num_keypoints": 32,
"use_separate_rgb_encoder_per_camera": true,
"n_conditioning_layers": 2,
"n_decoder_layers": 8,
"n_attention_heads": 4,
"attention_embed_dim": 512,
"attention_dropout": 0.1,
"embedding_dropout": 0.1,
"use_causal_attention": true,
"diffusion_step_embed_dim": 128,
"noise_scheduler_type": "DDPM",
"num_train_timesteps": 100,
"beta_schedule": "squaredcos_cap_v2",
"beta_start": 0.0001,
"beta_end": 0.02,
"prediction_type": "epsilon",
"clip_sample": true,
"clip_sample_range": 1.0,
"num_inference_steps": null,
"do_mask_loss_for_padding": false,
"optimizer_lr": 0.0001,
"optimizer_betas": [
0.95,
0.999
],
"optimizer_eps": 1e-08,
"optimizer_weight_decay": 1e-06,
"scheduler_name": "cosine",
"scheduler_warmup_steps": 500,
"flow_matching_type": "CondOT",
"num_integration_steps": 50,
"fm_time_embed_dim": 128,
"fm_min_period": 0.004,
"fm_max_period": 4.0
}