{ "
Returns a list of parameters for training
\u8fd4\u56de\u8bad\u7ec3\u7684\u53c2\u6570\u5217\u8868
\n": "\n", "
Backward pass
\n": "\u5411\u540e\u4f20\u7403
\n", "Calculate accuracy
\n": "\u8ba1\u7b97\u7cbe\u5ea6
\n", "Calculate loss
\n": "\u8ba1\u7b97\u635f\u5931
\n", "Filter parameters that require gradients
\n": "\u8fc7\u6ee4\u9700\u8981\u6e10\u53d8\u7684\u53c2\u6570
\n", "Forward pass
\n": "\u5411\u524d\u4f20\u7403
\n", "Get all parameters
\n": "\u83b7\u53d6\u6240\u6709\u53c2\u6570
\n", "Get predictions
\n": "\u83b7\u53d6\u9884\u6d4b
\n", "Iterate through the batches
\n": "\u904d\u5386\u6279\u6b21
\n", "Move targets to the same device as output
\n": "\u5c06\u76ee\u6807\u79fb\u52a8\u5230\u4e0e\u8f93\u51fa\u76f8\u540c\u7684\u8bbe\u5907\u4e0a
\n", "Optimize
\n": "\u4f18\u5316
\n", "Set gradients to zero
\n": "\u5c06\u6e10\u53d8\u8bbe\u7f6e\u4e3a\u96f6
\n", "Set model for train
\n": "\u8bbe\u7f6e\u706b\u8f66\u6a21\u578b
\n", "tracker.add({'loss.scaled': loss})
\n": "tracker.add ({'loss.scaled': loss})
\n", "Returns the loss, output and the target
\u8fd4\u56de\u635f\u5931\u3001\u8f93\u51fa\u548c\u76ee\u6807