Searched refs:dist_inputs (Results 1 – 3 of 3) sorted by relevance
159 def distributed_train_step(dist_inputs): argument161 train_step, args=(dist_inputs,))
176 def train_step(dist_inputs): argument189 per_replica_losses = strategy.run(step_fn, args=(dist_inputs,))194 return step_fn(dist_inputs)
2391 def train_step(dist_inputs): argument2418 per_replica_losses = distribution.run(step_fn, args=(dist_inputs,))