!5734 fix bug of adam optimizer

Merge pull request !5734 from fary86/fix_adam_opt_bug

mindspore/context.py

@@ -219,7 +219,7 @@ class _Context:
         self.set_param(ms_ctx_param.profiling_options, option)
 
     def set_variable_memory_max_size(self, variable_memory_max_size):
-        if not check_input_format(variable_memory_max_size):
+        if not _check_input_format(variable_memory_max_size):
             raise ValueError("Context param variable_memory_max_size should be in correct format! Such as \"5GB\"")
         if int(variable_memory_max_size[:-2]) >= _DEVICE_APP_MEMORY_SIZE:
             raise ValueError("Context param variable_memory_max_size should be less than 31GB.")
@@ -230,7 +230,7 @@ class _Context:
         self.set_param(ms_ctx_param.graph_memory_max_size, graph_memory_max_size_)
 
     def set_max_device_memory(self, max_device_memory):
-        if not check_input_format(max_device_memory):
+        if not _check_input_format(max_device_memory):
             raise ValueError("Context param max_device_memory should be in correct format! Such as \"3.5GB\"")
         max_device_memory_value = float(max_device_memory[:-2])
         if max_device_memory_value == 0:
@@ -289,7 +289,7 @@ class _Context:
         thread_info.debug_runtime = enable
 
 
-def check_input_format(x):
+def _check_input_format(x):
     import re
     pattern = r'[1-9][0-9]*(\.)?[0-9]*GB|0\.[0-9]*GB'
     result = re.match(pattern, x)

mindspore/nn/optim/adam.py

@@ -51,6 +51,7 @@ def _update_run_op(beta1, beta2, eps, lr, weight_decay, param, m, v, gradient, d
     Returns:
         Tensor, the new value of v after updating.
     """
+    success = True
     if optim_filter:
         op_mul = P.Mul()
         op_square = P.Square()
@@ -80,8 +81,8 @@ def _update_run_op(beta1, beta2, eps, lr, weight_decay, param, m, v, gradient, d
         next_param = F.depend(next_param, F.assign(param, op_cast(next_param, F.dtype(param))))
         next_param = F.depend(next_param, F.assign(m, op_cast(next_m, F.dtype(m))))
         next_param = F.depend(next_param, F.assign(v, op_cast(next_v, F.dtype(v))))
-        return next_param
-    return gradient
+        success = F.depend(success, next_param)
+    return success
 
 
 @_adam_opt.register("Function", "Function", "Function", "Function", "Tensor", "Tensor", "Tensor", "Tensor", "Tensor",