Functions — github.com/arturxe2/ASTRA

CodeHub MCPcopy Create free account

hub / github.com/arturxe2/ASTRA / functions

Functions42 in github.com/arturxe2/ASTRA

⨍ Functions 42 ◇ Types & classes 12

↓ 8 callersFunctionfeats2clip

Auxiliar function to split video features into clips

↓ 6 callersMethod__init__

( self, chunk_size = 8, n_output = 24, baidu = True, audio = False,

↓ 4 callersMethodstep

(self, cur_iteration)

↓ 3 callersMethodload_state_dict

(self, state_dict)

↓ 3 callersFunctiontestSpotting

Function for inference of the action spotting model (and evaluation)

↓ 3 callersFunctiontrainAS

Function to train 1 epoch of the action spotting model

↓ 2 callersFunctionapply_NMS

(predictions, window, thresh=0.0)

↓ 2 callersFunctionpred2vec

(predictions, chunk_size = 32, outputrate = 2, threshold = 0.01, target = False, NMS = False, window = 4)

↓ 1 callersMethod__init__

(self, wC = 1, wD = 1, focal = False, nw = 7, uncertainty = False, uncertainty_mode = 'mse')

↓ 1 callersFunctioncompute_mAP

(preds, labs, framerate = 2, metric = 'tight')

↓ 1 callersFunctionmain

main_challenge.py:18

↓ 1 callersFunctionmain

↓ 1 callersFunctiontrainerAS

Function to train the action spotting model (with validation early stopping)

↓ 1 callersFunctiontrainerAS_test

Function to train the action spotting model (when no validation early stopping - for using all data and evaluate on challenge)

↓ 1 callersMethodwarmup_learning_rate

(self, cur_iteration)

↓ 1 callersFunctionzipResults

(zip_path, target_dir, filename="results_spotting.json")

Method__getitem__

Method__getitem__

(self, alpha = 0.6, beta = 0.4)

(self, optimizer, warmup_iteration, target_lr, after_scheduler=None)

(self, path_labels = "ResNET_TF2", path_store = "SoccerNetSamples",

(self, path_labels = "ResNET_TF2", path_baidu = "Baidu_features", path_audio =

(self, input_dim, output_dim, sigmoid = True, drop = 0.2)

(self, input_dim, output_dim, drop = 0.2)

(self, input_dim, output_dim, drop = 0.2)

(self, alpha = 0.3, beta = 0.3)

(self, p = 0.0, dim = 8576)

(self, p = 0.0)

labelsD: b x 18 predictionsD: b x 2 (with mean in 0 and logvar in 1)

(self, labels, predictions, labelsD = None, predictionsD = None)

(self, featsB = None, featsA = None, labels = None, labelsD = None, inference = False)

(self, x: torch.Tensor)

(self, x: torch.Tensor)

(self, x: torch.Tensor)

(self, featB, featBQ, featA, featAQ, labels, labelsQ, labelsD, labelsDQ)

(self, x: torch.Tensor)

(self, x: torch.Tensor)

Functionget_spot_from_NMS

Non-Maximum Suppression

Functionget_spot_from_SNMS

Soft Non-Maximum Suppression

Methodload_weights

(self, weights=None)