MCPcopy Create free account
hub / github.com/deepspeedai/DeepSpeed / train_batch_size

Method train_batch_size

deepspeed/runtime/engine.py:1264–1265  ·  view source on GitHub ↗
(self)

Source from the content-addressed store, hash-verified

1262 return self._config.sparse_gradients_enabled
1263
1264 def train_batch_size(self):
1265 return self._config.train_batch_size
1266
1267 def train_micro_batch_size_per_gpu(self):
1268 return self._config.train_micro_batch_size_per_gpu

Callers 10

__init__Method · 0.95
_take_model_stepMethod · 0.95
_autotuning_exitMethod · 0.95
_load_checkpointMethod · 0.95
test_unmanaged_varying_backward_countMethod · 0.80
__init__Method · 0.80
train_batchMethod · 0.80
get_dataloader_and_lr_scheduler_for_variable_batch_size_deepspeedFunction · 0.80
_take_lr_scheduler_stepFunction · 0.80
mainFunction · 0.80

Calls

no outgoing calls

Tested by 1

test_unmanaged_varying_backward_countMethod · 0.64