MCPcopy Create free account

hub / github.com/MIKUZ12/DreamZero-Libero / functions

Functions1,297 in github.com/MIKUZ12/DreamZero-Libero

⨍ Functions 1,297◇ Types & classes 203↳ Endpoints 1
↓ 218 callersMethodappend
(self, data: Dict)
groot/vla/common/utils/io/json_utils.py:240
↓ 146 callersMethodget
(self, name)
groot/vla/common/utils/misc/functional_utils.py:333
↓ 50 callersMethodmean
(self)
groot/vla/common/utils/misc/torch_utils.py:658
↓ 34 callersMethodupdate
(self, value, n=1)
groot/vla/common/utils/misc/torch_utils.py:728
↓ 33 callersFunctionf_join
Join file paths and expand special symbols like `~` for home dir
groot/vla/common/utils/io/file_utils.py:140
↓ 26 callersFunctioncprint
Print colorize text. It accepts arguments of print function.
groot/vla/common/utils/io/termcolor.py:114
↓ 23 callersMethodclose
(self)
groot/vla/common/utils/io/json_utils.py:250
↓ 21 callersMethoddecode
(self, z, scale)
groot/vla/model/dreamzero/modules/wan_video_vae.py:998
↓ 19 callersMethodlog
(self, data)
groot/vla/utils/timer.py:31
↓ 19 callersMethodnorm
(self, x)
groot/vla/model/dreamzero/modules/wan_video_dit.py:124
↓ 17 callersMethod__init__
( self, in_channels: int, out_channels: int, factor_t, factor_s=1,
groot/vla/model/dreamzero/modules/wan_video_vae.py:395
↓ 17 callersMethodget_trajectory_index
Get the index of the trajectory in the dataset by the trajectory ID. This is useful when you need to get the trajectory length or sampling wei
groot/vla/data/dataset/lerobot.py:1349
↓ 15 callersMethodencode
(self, x, scale)
groot/vla/model/dreamzero/modules/wan_video_vae.py:969
↓ 15 callersFunctioninstantiate
Any dict with "cls" or "class" key is considered instantiable. Any key that has the special value "__delete__" will not be passed to the
groot/vla/common/utils/io/config_utils.py:233
↓ 14 callersMethodextend
(self, data_list: List[Dict])
groot/vla/common/utils/io/json_utils.py:246
↓ 13 callersMethoddevice
(self)
groot/vla/model/dreamzero/action_head/wan_flow_matching_action_tf.py:1402
↓ 13 callersMethodeval
(self)
groot/vla/data/transform/base.py:69
↓ 13 callersFunctionhash_state_dict_keys
(state_dict, with_shape=True)
groot/vla/model/dreamzero/modules/utils.py:179
↓ 12 callersFunctionf_expand
(fpath)
groot/vla/common/utils/io/file_utils.py:132
↓ 11 callersMethod__init__
(self, embed_dim=512, image_size=224, patch_size=16,
groot/vla/model/dreamzero/modules/wan_video_image_encoder.py:487
↓ 11 callersFunctionis_numpy
(obj)
groot/vla/common/utils/misc/array_tensor_utils.py:24
↓ 11 callersFunctionis_tensor
(obj)
groot/vla/common/utils/misc/array_tensor_utils.py:28
↓ 11 callersFunctionmprint
(*args, **kwargs)
groot/vla/experiment/utils.py:37
↓ 11 callersMethodreset
(self, reset_info)
eval_utils/policy_server.py:123
↓ 10 callersMethod__init__
(self, in_dim, out_dim, has_pos_emb=False)
groot/vla/model/dreamzero/modules/wan_video_dit.py:351
↓ 10 callersMethodadd
(self, cls)
groot/vla/common/utils/misc/functional_utils.py:330
↓ 10 callersMethodapply
Normalize inputs
groot/vla/model/n1_5/sim_policy.py:568
↓ 10 callersMethodprepare_input
(self, inputs)
groot/vla/model/dreamzero/base_vla.py:243
↓ 10 callersFunctionunsqueeze_dict_values
Unsqueeze the values of a dictionary. This converts the data to be batched of size 1.
groot/vla/model/n1_5/sim_policy.py:1043
↓ 9 callersMethod__init__
(self, dim, out_dim, patch_size, eps=1e-6)
groot/vla/model/dreamzero/modules/wan2_1_submodule.py:537
↓ 9 callersMethod_check_state_is_batched
(self, obs: dict[str, Any])
groot/vla/model/n1_5/sim_policy.py:1021
↓ 9 callersFunctionis_sequence
Returns: True if the sequence is a collections.Sequence and not a string.
groot/vla/common/utils/data_structure/tree_utils.py:20
↓ 9 callersMethodretrieve_data_and_pad
Retrieve the data from the dataset and pad it if necessary. Args: array (np.ndarray): The array to retrieve the data from.
groot/vla/data/dataset/lerobot.py:1370
↓ 9 callersMethodstep
(self, model_output, timestep, sample, to_final=False, **kwargs)
groot/vla/model/dreamzero/modules/flow_match_scheduler.py:43
↓ 8 callersFunctionis_mapping
Returns: True if the sequence is a collections.Mapping
groot/vla/common/utils/data_structure/tree_utils.py:28
↓ 8 callersMethodmodel_dump
(self, *args, **kwargs)
groot/vla/data/transform/concat.py:59
↓ 8 callersFunctionrope_params
(max_seq_len, dim, theta=10000)
groot/vla/model/dreamzero/modules/wan2_1_submodule.py:29
↓ 8 callersFunctionsqueeze_dict_values
Squeeze the values of a dictionary. This removes the batch dimension.
groot/vla/model/n1_5/sim_policy.py:1063
↓ 8 callersFunctiontf_to_torch
(data)
scripts/data/convert_droid.py:72
↓ 8 callersMethodto_numpy
Convert tensor back to numpy array efficiently. Args: frames: tensor of shape [T, C, H, W] in range [0, 1] Returns:
groot/vla/data/transform/video.py:613
↓ 8 callersMethodunapply
Unnormalize actions and convert relative actions to absolute if needed
groot/vla/model/n1_5/sim_policy.py:614
↓ 7 callersMethod_sigma_to_alpha_sigma_t
(sigma)
groot/vla/model/dreamzero/modules/flow_unipc_multistep_scheduler.py:234
↓ 7 callersFunctioncheck_is_instance
(model, module_class)
groot/vla/model/dreamzero/modules/wan_video_vae.py:11
↓ 7 callersMethodvalidate_data
(self, action_head_outputs, backbone_outputs, is_training)
groot/vla/model/dreamzero/base_vla.py:106
↓ 6 callersMethod__init__
(self, dim, out_dim, patch_size, eps=1e-6)
groot/vla/model/dreamzero/modules/wan_video_dit_action_casual_chunk.py:1192
↓ 6 callersFunctionflash_attention
q: [B, Lq, Nq, C1]. k: [B, Lk, Nk, C1]. v: [B, Lk, Nk, C2]. Nq must be divisible by Nk. q_lens
groot/vla/model/dreamzero/modules/attention.py:26
↓ 6 callersFunctionfmt
(d, key)
scripts/compare_loss.py:39
↓ 6 callersFunctionformat_action_dict
(action_dict: dict, threshold_gripper: bool)
eval_utils/diagnose_libero_fit.py:293
↓ 6 callersFunctionhalf
(x)
groot/vla/model/dreamzero/modules/wan2_1_attention.py:82
↓ 6 callersFunctionhalf
(x)
groot/vla/model/dreamzero/modules/attention.py:61
↓ 6 callersMethodlazy_joint_forward_causal
(self, batch, video=None, latent_video=None, state=None, video_only=False, **kwargs)
groot/vla/model/n1_5/sim_policy.py:758
↓ 6 callersFunctionrope_action_apply
(x, freqs, freqs_action, freqs_state, action_register_length, num_action_per_block=32, num_state_per_block=1)
groot/vla/model/dreamzero/modules/wan2_1_submodule.py:92
↓ 6 callersMethodstate_dict
(self)
groot/vla/common/utils/misc/torch_utils.py:332
↓ 6 callersMethodstd
(self)
groot/vla/common/utils/misc/torch_utils.py:666
↓ 5 callersFunction_RMSNorm
(normalized_shape, eps)
groot/vla/model/dreamzero/modules/wan_video_dit.py:132
↓ 5 callersMethod__init__
(self, dim, eps=1e-6)
groot/vla/model/dreamzero/modules/wan_video_text_encoder.py:24
↓ 5 callersMethod_forward
(self, obs: dict)
eval_utils/run_libero_server.py:167
↓ 5 callersFunctionany_to_primitive
(x)
groot/vla/common/utils/io/json_utils.py:67
↓ 5 callersMethodget_task
(self, idx)
scripts/open_loop_yam.py:119
↓ 5 callersMethodget_trajectory_data
Get the trajectory data.
groot/vla/data/dataset/lerobot_sharded.py:1240
↓ 5 callersFunctionload_local_dataset
Load local dataset and return a dict with observations and actions Args: episode_id: Episode ID (used for old format) src_path: S
scripts/data/convert_agibot.py:838
↓ 5 callersMethodrun
(self, rank: int = 0)
socket_test_optimized_AR.py:499
↓ 5 callersMethodtrain
Correctly set self.state from checkpoint so get_train_dataloader can read from it.
groot/vla/experiment/base.py:538
↓ 5 callersMethodwith_label
(self, key)
groot/vla/utils/timer.py:12
↓ 5 callersFunctionwrite_results
(output_dir: Path, results: dict)
eval_utils/run_libero_eval.py:232
↓ 4 callersMethod_as_state
(self, value)
eval_utils/run_libero_server.py:111
↓ 4 callersMethod_locate
(self, idx)
scripts/open_loop_yam.py:105
↓ 4 callersFunction_match_patterns_helper
(element, patterns)
groot/vla/common/utils/misc/misc_utils.py:29
↓ 4 callersFunctionbuild_payload
( agentview_video: np.ndarray, wrist_video: np.ndarray, eef_state: np.ndarray, gripper_positio
eval_utils/diagnose_libero_fit.py:267
↓ 4 callersFunctioncast_to
(weight, dtype, device)
groot/vla/model/dreamzero/modules/vram_management.py:5
↓ 4 callersFunctionconcat_state_or_action
(modality_dict, keys, compute_concat_info=False)
scripts/data/convert_droid.py:95
↓ 4 callersFunctioncount_conv3d
(model)
groot/vla/model/dreamzero/modules/wan_video_vae.py:931
↓ 4 callersFunctionflash_attention
(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor, num_heads: int, compatibility_mode=False)
groot/vla/model/dreamzero/modules/wan_video_dit.py:39
↓ 4 callersMethodfrom_pretrained
( cls, pretrained_model_name_or_path: str, config: VLAConfig = None )
groot/vla/model/dreamzero/base_vla.py:505
↓ 4 callersMethodget_video_path
Get the video file path for a specific trajectory and video key. Args: trajectory_id (int): The ID of the trajectory.
groot/vla/data/dataset/lerobot.py:1422
↓ 4 callersFunctionimplements_method
Returns: True if object implements a method
groot/vla/common/utils/misc/functional_utils.py:50
↓ 4 callersMethodinfer
(self, obs: Dict)
eval_utils/policy_client.py:66
↓ 4 callersFunctionleft_pad_history
(frames: np.ndarray, end_index: int, history_frames: int)
eval_utils/diagnose_libero_fit.py:223
↓ 4 callersFunctionmake_rollout_frame
(obs: dict)
eval_utils/diagnose_libero_fit.py:239
↓ 4 callersFunctionmake_rollout_frame
(obs: dict)
eval_utils/run_libero_eval.py:208
↓ 4 callersFunctionmatch_patterns
Args: include: None to disable `include` filter and delegate to exclude precedence: "include" or "exclude"
groot/vla/common/utils/misc/misc_utils.py:38
↓ 4 callersFunctionsinusoidal_embedding_1d
(dim: int, position: torch.Tensor)
groot/vla/model/dreamzero/modules/wan2_1_submodule.py:16
↓ 3 callersMethod__init__
(self)
groot/vla/model/n1_5/sim_policy.py:72
↓ 3 callersMethod__init__
(self)
groot/vla/model/dreamzero/modules/vram_management.py:12
↓ 3 callersMethod__init__
( self, *args, num_steps_per_shard: int = int(1e4), **kwargs, )
groot/vla/data/dataset/lerobot_sharded.py:23
↓ 3 callersMethod__init__
Args: out_file: file path err_file: file path. If the same as out_file, print both stdout and stderr to one
groot/vla/common/utils/io/print_utils.py:244
↓ 3 callersMethod_ensure_inference_action_fields
Fill training-only action fields required by the action head at inference.
groot/vla/model/n1_5/sim_policy.py:576
↓ 3 callersMethod_forward_inference
r""" Run the diffusion model with kv caching. See Algorithm 2 of CausVid paper https://arxiv.org/abs/2412.07772 for details. T
groot/vla/model/dreamzero/modules/wan_video_dit_action_casual_chunk.py:1863
↓ 3 callersFunction_get_video_info_ffmpeg
Get video metadata using ffprobe.
groot/vla/common/utils/misc/video_utils.py:25
↓ 3 callersFunction_log_action
Pretty-print action shape, range, and timing.
test_client_AR.py:277
↓ 3 callersMethod_run_diffusion_steps
( self, noisy_input: torch.Tensor, timestep: torch.Tensor, action: torch.Tenso
groot/vla/model/dreamzero/action_head/wan_flow_matching_action_tf.py:844
↓ 3 callersFunction_summarize_checkpoint_keys
(keys: set[str], limit: int = 20)
groot/vla/experiment/base.py:83
↓ 3 callersMethodcheck_free_vram
(self)
groot/vla/model/dreamzero/modules/vram_management.py:15
↓ 3 callersMethodclose
(self)
groot/vla/common/utils/misc/image_utils.py:146
↓ 3 callersFunctiondraw_bar
( draw: ImageDraw.ImageDraw, x: int, y: int, w: int, h: int, gt_value: float, pred
eval_utils/visualize_libero_teacher_forced_trace.py:54
↓ 3 callersMethoddtype
(self)
groot/vla/model/dreamzero/action_head/wan_flow_matching_action_tf.py:1406
↓ 3 callersMethodenable_vram_management
(self, num_persistent_param_in_dit=None)
groot/vla/model/dreamzero/action_head/wan_flow_matching_action_tf.py:453
↓ 3 callersFunctionensure_file
Return a valid local path: use `path` if it exists, otherwise download from HuggingFace.
groot/vla/model/dreamzero/action_head/wan_flow_matching_action_tf.py:32
↓ 3 callersMethodflush
Manually flush the replaced stdout/stderr buffers.
groot/vla/common/utils/io/print_utils.py:233
↓ 3 callersMethodforward
(self, x: torch.Tensor)
groot/vla/data/transform/state_action.py:93
next →1–100 of 1,297, ranked by callers