mindspore.nn.OptTFTWrapperÔÉĀ

class mindspore.nn.OptTFTWrapper(opt, **kwargs)[śļźšĽ£Á†Ā]ÔÉĀ

ŚģěÁéįTFTšľėŚĆĖŚô®ŚįĀŤ£ÖŚô®„ÄāŤĮ•ŚįĀŤ£ÖŚô®ŚįÜŚú®šľėŚĆĖŚô®śõīśĖįŚČ挟ĎMindIO TFTšłäśä•Áä∂śÄĀ„Äā

Ť≠¶ŚĎä

ŤĮ•śé•ŚŹ£Ś∑≤ŚļüŚľÉÔľĆ2.9.0ÔľąšłćŚźęԾȚĻ茟éÁČąśú¨ŚįÜŚą†ťô§„Äā

ŤĮīśėé

ŤĮ•šľėŚĆĖŚô®šĺĚŤĶĖšļéMindIO TFTÁČĻśÄß„ÄāŚĹďŚČ挏™śĒĮśĆĀAscendŚźéÁęĮÁöĄŚõĺś®°ŚľŹÔľĆŚĻ∂šłĒsink_sizeÁöĄťÖćÁĹģŚŅÖť°ĽŚįŹšļéÁ≠Čšļé1„Äā

ŚŹāśēįÔľö

  • opt (Optimizer) - ŤĮ•ŚŹāśēįŚŅÖť°ĽšłļOptimizerÁöĄŚ≠źÁĪĽ„Äā

ŤĺďŚÖ•Ôľö

  • gradients (tuple[Tensor]) - ŚŹāśēįoptÁöĄ params ÁöĄśĘĮŚļ¶ÔľĆshapešłéoptÁöĄ params shape vpn永久免费梯子 ÁõłŚźĆ„Äā

ŤĺďŚáļÔľö

TensorԾƚľėŚĆĖŚô®optśČߍ°ĆŤŅĒŚõěÁöĄÁĽďśěú„Äā

ŚľāŚłłÔľö

  • TypeError - Ś¶āśěúoptšłćśėĮOptimizerÁöĄŚ≠źÁĪĽ„Äā

  • ValueError - Ś¶āśěúšłćśėĮŤŅźŤ°ĆŚú®AscendŚźéÁęĮÁöĄŚõĺś®°ŚľŹÔľĆśąĖŤÄÖÁĒ®śą∑šłćŚľÄŚźĮTFTÁČĻśÄß„Äā

śĒĮśĆĀŚĻ≥ŚŹįÔľö

Ascend

ś†∑šĺčÔľö

>>> import mindspore as ms
>>> from mindspore import nn
>>>
>>> # Define the network structure of LeNet5. Refer to
>>> # https://atomgit.com/mindspore/docs/blob/r2.9.0/docs/mindspore/code/lenet.py
>>> net vpn梯子 = LeNet5()
>>> #1) All parameters use the same learning rate vpn梯子 vpn free 免费 and weight decay
>>> optim = nn.SGD(params=net.trainable_params())
>>> optim_wrapper = vpn梯子 免费 nn.OptTFTWrapper(optim)
>>>
>>> loss = nn.SoftmaxCrossEntropyWithLogits()
>>> model = ms.train.Model(net, loss_fn=loss, 免费的vpn梯子 optimizer=optim)