Functions — github.com/SpatialVLA/SpatialVLA

hub / github.com/SpatialVLA/SpatialVLA / functions

Functions244 in github.com/SpatialVLA/SpatialVLA

⨍ Functions 244◇ Types & classes 35 ↳ Endpoints 1

Method__init__

( self, data_root_dir, output_dir, data_mix, image_size=224, m

data/dataset.py:19

Method__init__

( self, batch_size: int, world_size: int, dataset: Optional[Dataset] = None,

train/monkey_patch.py:117

Method__init__

(self, processor)

train/monkey_patch.py:333

Method__init__

(self, *args, **kwargs)

train/monkey_patch.py:345

train/monkey_patch.py:153

train/monkey_patch.py:150

Method_check_and_enable_sdpa

Overloads `PreTrainedModel._check_and_enable_sdpa` so as to DISABLE torch SDPA by default on Gemma2 models. SDPA reduces the model pe

model/modeling_gemma2.py:554

Function_get_train_sampler

(self)

train/monkey_patch.py:158

Method_init_weights

(self, module)

model/modeling_gemma2.py:542

Method_init_weights

(self, module)

model/modeling_spatialvla.py:143

Methodaction_token_begin_idx

(self)

model/action_tokenizer.py:340

Functionadd_pad_mask_dict

Adds a dictionary indicating which elements of the observation/task should be treated as padding. =>> traj["observation"|"task"]["pad_mas

data/traj_transforms.py:87

Functionasu_table_top_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:731

Augments images, skipping padding images.

data/obs_transforms.py:17

Functionaustin_buds_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:442

Functionaustin_sailor_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:546

Functionaustin_sirius_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:564

Functionbc_z_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:582

Functionberkeley_autolab_ur5_dataset_transform

( trajectory: Dict[str, Any] )

data/oxe/transforms.py:305

Functionberkeley_cable_routing_dataset_transform

( trajectory: Dict[str, Any] )

data/oxe/transforms.py:217

Functionberkeley_fanuc_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:802

Functionberkeley_mvp_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:656

Functionberkeley_rpt_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:660

Functionbridge_orig_dataset_transform

Applies to original version of Bridge V2 from the official project website. Note =>> In original Bridge V2 dataset, the first timestep has a

data/oxe/transforms.py:65

Functionbridge_oxe_dataset_transform

Applies to version of Bridge V2 in Open X-Embodiment mixture. Note =>> In original Bridge V2 dataset, the first timestep has an all-zero act

data/oxe/transforms.py:31

Functioncartesian_to_spherical

(x, y, z)

data/utils/data_utils.py:210

Functionchunk_act_obs

Chunks actions and observations into the given window_size. "observation" keys are given a new axis (at index 1) of size `window_size` conta

data/traj_transforms.py:12

Functioncmu_franka_exploration_dataset_transform

( trajectory: Dict[str, Any] )

data/oxe/transforms.py:517

Functioncmu_playing_with_food_dataset_transform

( trajectory: Dict[str, Any] )

data/oxe/transforms.py:819

Functioncmu_stretch_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:846

Functioncompute_loss

How the loss is computed by Trainer. By default, all models return the loss in the first element. Subclass and override for custom behavior.

train/monkey_patch.py:222

Functionconcat_pad_data_collator

(features, pad_id=0)

train/monkey_patch.py:21

Functiondecode_and_resize

Decodes images and depth images, and then optionally resizes them.

data/obs_transforms.py:47

Methoddecode_token_ids_to_actions

decode token ids to continuous actions. action_token_id: np.ndarray, (n,), token ids. return: np.ndarray, (n, 3), continuous actions

model/action_tokenizer.py:121

Methoddecode_token_ids_to_actions

decode token ids to continuous actions. action_token_id: np.ndarray, (n,), token ids. return: np.ndarray, (n, 3), continuous actions

model/action_tokenizer.py:190

Methoddecode_token_ids_to_actions

decode token ids to continuous actions. action_token_id: np.ndarray, (n,), token ids. return: np.ndarray, (n, 1), continuous actions

model/action_tokenizer.py:235

Methoddecode_token_ids_to_actions

decode token ids to continuous actions. action_token_ids: np.ndarray, (n, 3), token ids.

model/action_tokenizer.py:321

Functiondelete_task_conditioning

Randomly drops out either the goal images or the language instruction. Only does something if both of these are present. Args: t

data/utils/task_augmentation.py:14

Functiondlr_edan_shared_control_dataset_transform

( trajectory: Dict[str, Any] )

data/oxe/transforms.py:717

Functiondlr_sara_grid_clamp_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:712

Functiondlr_sara_pour_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:708

Functiondobbe_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:894

Functiondroid_baseact_transform

DROID dataset transformation for actions expressed in *base* frame of the robot.

data/utils/droid_utils.py:68

Functiondroid_finetuning_transform

DROID dataset transformation for actions expressed in *base* frame of the robot.

data/utils/droid_utils.py:132

Functiondroid_wristact_transform

DROID dataset transformation for actions expressed in *wrist* frame of the robot.

data/utils/droid_utils.py:100

Functioneager_attention_forward

( config: Gemma2Config, query: torch.Tensor, key: torch.Tensor, value: torch.Tensor, mask:

model/modeling_gemma2.py:169

Methodextra_repr

(self)

model/modeling_gemma2.py:76

Functionfastumi_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:990

Functionflash_attention_forward

( config: Gemma2Config, query: torch.Tensor, key: torch.Tensor, value: torch.Tensor, mask:

model/modeling_gemma2.py:198

Functionflex_attention_forward

( config: Gemma2Config, query: torch.Tensor, key: torch.Tensor, value: torch.Tensor, mask:

model/modeling_gemma2.py:244

Functionfmb_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:882

Methodforward

(self, x)

model/modeling_gemma2.py:69

Methodforward

(self, x)

model/modeling_gemma2.py:91

Methodforward

(self, x, position_ids, seq_len=None)

model/modeling_gemma2.py:106

Methodforward

( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None,

model/modeling_gemma2.py:364

Methodforward

( self, hidden_states: torch.Tensor, attention_mask: Optional[torch.Tensor] = None,

model/modeling_gemma2.py:451

Methodforward

( self, input_ids: torch.LongTensor = None, attention_mask: Optional[torch.Tensor] = N

model/modeling_gemma2.py:680

Methodforward

r""" Args: labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing

model/modeling_gemma2.py:920

Methodforward

r""" labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): Labels for computing the sequence classification/regression

model/modeling_gemma2.py:1126

Methodforward

r""" labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): Labels for computing the sequence classification/regression

model/modeling_gemma2.py:1234

Methodforward

Forward pass, xyz is (B, N, 3or6), output (B, N, F).

model/modeling_spatialvla.py:93

Methodforward

(self, image_features)

model/modeling_spatialvla.py:126

Methodforward

( self, input_ids: torch.LongTensor = None, pixel_values: torch.FloatTensor = None,

model/modeling_spatialvla.py:335

Functionfurniture_bench_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:492

Methodget_decoder

(self)

model/modeling_gemma2.py:915

Methodget_decoder

(self)

model/modeling_spatialvla.py:240

Methodget_input_embeddings

(self)

model/modeling_gemma2.py:673

Methodget_input_embeddings

(self)

model/modeling_gemma2.py:900

Methodget_input_embeddings

(self)

model/modeling_gemma2.py:1119

Methodget_input_embeddings

(self)

model/modeling_gemma2.py:1222

Methodget_output_embeddings

(self)

model/modeling_spatialvla.py:231

Functionget_train_dataloader

Returns the training [`~torch.utils.data.DataLoader`]. Will use no sampler if `train_dataset` does not implement `__len__`, a random sampler

train/monkey_patch.py:182

Functiongnm_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:861

Functioniamlab_pick_insert_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:752

Methodignore_index

(self)

model/configuration_spatialvla.py:105

Functionimperial_wristcam_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:747

Functionjaco_play_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:191

Functionkaist_nonprehensible_dataset_transform

( trajectory: Dict[str, Any] )

data/oxe/transforms.py:664

Functionkuka_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:131

Functionlanguage_table_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:351

Functionlibero_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:953

Functionmaniskill_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:485

Methodmodel_input_names

(self)

model/processing_spatialvla.py:211

Functionnormalize_action_and_proprio

Normalizes the action and proprio fields of a trajectory using the given metadata.

data/utils/data_utils.py:58

Functionnyu_door_opening_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:260

Functionnyu_franka_play_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:461

Functionnyu_rot_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:407

Methodon_save

(self, args, state, control, **kwargs)

train/monkey_patch.py:336

Functionplayfusion_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:835

Functionppgm_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:95

Methodprepare_inputs_for_generation

( self, input_ids, past_key_values=None, attention_mask=None, inputs_e

model/modeling_gemma2.py:1015

Methodprepare_inputs_for_generation

( self, input_ids, past_key_values=None, inputs_embeds=None, cache_pos

model/modeling_spatialvla.py:445

Functionpusht_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:375

Functionrestructure

(traj)

data/rlds.py:130

Functionrh20t_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:918

Functionrmat_to_euler

(rot_mat)

data/utils/droid_utils.py:9

Functionrobo_net_dataset_transform

(trajectory: Dict[str, Any])

data/oxe/transforms.py:634

← previous next →101–200 of 244, ranked by callers