def _parse_model(config, model_name, batch_size):
if config.max_batch_size == 0:
if batch_size != 1:
raise Exception("batching not supported for model '" + model_name + "'")
else: # max_batch_size > 0
if batch_size > config.max_batch_size:
raise Exception(
"expecting batch size <= {} for model '{}'".format(config.max_batch_size, model_name))
input_ = config.input[0]
output = config.output[0]
if input_.format == model_config_pb2.ModelInput.FORMAT_NHWC:
h = input_.dims[0]
w = input_.dims[1]
c = input_.dims[2]
else:
c = input_.dims[0]
h = input_.dims[1]
w = input_.dims[2]
return input_.name, output.name, c, h, w, input_.format, tensor_dtype_to_np_dtype(input_.data_type)
def parse_model(config, model_name, batch_size):
"""
Check the configuration of a model to make sure it meets the
requirements for an image classification network (as expected by
this client)
"""
if config.name != model_name:
raise Exception("Get model config exception, config name: {}".format(config.name))
if len(config.input) != 1:
raise Exception("expecting 1 input, got {}".format(len(config.input)))
if len(config.output) != 1:
raise Exception("expecting 1 output, got {}".format(len(config.output)))
input_ = config.input[0]
output = config.output[0]
if output.data_type != types_pb2.DT_FLOAT:
raise Exception("expecting output data type to be DT_FLOAT, model '" +
model_name + "' output type is " +
model_config.DataType.Name(output.data_type))
# Output is expected to be a vector. But allow any number of
# dimensions as long as all but 1 is size 1 (e.g. { 10 }, { 1, 10
# }, { 10, 1, 1 } are all ok).
non_one_cnt = 0
for dim in output.dims:
if dim > 1:
non_one_cnt += 1
if non_one_cnt > 1:
raise Exception("expecting model output to be a vector")
# Model specifying maximum batch size of 0 indicates that batching
# is not supported and so the input tensors do not expect an "N"
# dimension (and 'batch_size' should be 1 so that only a single
# image instance is inferred at a time).
max_batch_size = config.max_batch_size
if max_batch_size == 0:
if batch_size != 1:
raise Exception("batching not supported for model '" + model_name + "'")
else: # max_batch_size > 0
if batch_size > max_batch_size:
raise Exception(
"expecting batch size <= {} for model '{}'".format(max_batch_size, model_name))
# Model input must have 3 dims, either CHW or HWC
if len(input_.dims) != 3:
raise Exception(
"expecting input to have 3 dimensions, model '{}' input has {}".format(
model_name, len(input_.dims)))
if ((input_.format != model_config.ModelInput.FORMAT_NCHW) and
(input_.format != model_config.ModelInput.FORMAT_NHWC)):
raise Exception("unexpected input format " + model_config.ModelInput.Format.Name(input_.format) +
", expecting " +
model_config.ModelInput.Format.Name(model_config.ModelInput.FORMAT_NCHW) +
" or " +
model_config.ModelInput.Format.Name(model_config.ModelInput.FORMAT_NHWC))
if input_.format == model_config.ModelInput.FORMAT_NHWC:
h = input_.dims[0]
w = input_.dims[1]
c = input_.dims[2]
else:
c = input_.dims[0]
h = input_.dims[1]
w = input_.dims[2]
output_size = 1
for dim in output.dims:
output_size = output_size * dim
output_size = output_size * np.dtype(tensor_dtype_to_np_dtype(output.data_type)).itemsize
return input_.name, output.name, c, h, w, input_.format, tensor_dtype_to_np_dtype(input_.data_type), output_size