MCPcopy Create free account
hub / github.com/Ericonaldo/ILSwiss / train

Method train

rlkit/core/base_algorithm.py:145–154  ·  view source on GitHub ↗
(self, start_epoch=0)

Source from the content-addressed store, hash-verified

143 self.ready_env_ids = np.arange(self.env_num)
144
145 def train(self, start_epoch=0):
146 self.pretrain()
147 if start_epoch == 0:
148 params = self.get_epoch_snapshot(-1)
149 logger.save_itr_params(-1, params)
150 self.training_mode(False)
151 # self._n_env_steps_total = start_epoch * self.num_env_steps_per_epoch
152 gt.reset()
153 gt.set_def_unique(False)
154 self.start_training(start_epoch=start_epoch)
155
156 def pretrain(self):
157 """

Callers 15

experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45
experimentFunction · 0.45

Calls 5

pretrainMethod · 0.95
get_epoch_snapshotMethod · 0.95
training_modeMethod · 0.95
start_trainingMethod · 0.95
resetMethod · 0.45

Tested by

no test coverage detected