代码之家 › 专栏 › 技术社区 › Alex Punnen

权值量化后的张量流模型误差

tensorflow-serving tensorflow python

Alex Punnen · 技术社区 · 5 年前

Quantizing 重量以及模型,然后保存为服务和使用它,我从客户端得到以下错误

_集合:<\u RPC的集合,终止于:status=StatusCode.UNIMPLEMENTED状态码未实现details=“在[1,9,4]和尚不支持[221,1,4]。[{{节点筛选的“\u detections/map/while/non\u max\u suppression\u 47/NonMaxSuppressionV3}}]” 调试错误\u字符串= “{”created“:@1552349235.607723578”,“description”:“收到错误” 尚不支持在[1,9,4]和[221,1,4]之间。\n\t[{{node] 锚定\u 3/add \u 2/eightbit}}]]\n\t[{{节点

在TF服务端,我得到以下错误

2019-03-12 00:07:13.149087:一外部/org\u tensorflow/tensorflow/core/kernels/quantized\u add_作品抄送:546] ndims=3 2019-03-12 00:07:13.149233:I bcast.x_reformate()=[1,9,4]2019-03-12 00:07:13.149309:I 外部/org\u tensorflow/tensorflow/core/kernels/quantized\u add_作品抄送:549] bcast.y_reformate()=[221,1,4]2019-03-12 00:07:13.149348:I 外部/org\u tensorflow/tensorflow/core/kernels/quantized\u add_作品抄送:551] bcast.x_bcast()=[221,1,1]2019-03-12 00:07:13.149385:I 外部/org\u tensorflow/tensorflow/core/kernels/quantized\u add_作品抄送:553] bcast.y\u bcast()=[1,9,1]

MetaGraphDef with tag-set: 'serve' contains the following SignatureDefs:

signature_def['serving_default']:
  The given SavedModel SignatureDef contains the following input(s):
    inputs['input_image'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, -1, -1, 3)
        name: input_1_2:0
  The given SavedModel SignatureDef contains the following output(s):
    outputs['filtered_detections/map/TensorArrayStack/TensorArrayGatherV3:0'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 300, 4)
        name: filtered_detections/map/TensorArrayStack/TensorArrayGatherV3:0
    outputs['filtered_detections/map/TensorArrayStack_1/TensorArrayGatherV3:0'] tensor_info:
        dtype: DT_FLOAT
        shape: (-1, 300)
        name: filtered_detections/map/TensorArrayStack_1/TensorArrayGatherV3:0
    outputs['filtered_detections/map/TensorArrayStack_2/TensorArrayGatherV3:0'] tensor_info:
        dtype: DT_INT32
        shape: (-1, 300)
        name: filtered_detections/map/TensorArrayStack_2/TensorArrayGatherV3:0
  Method name is: tensorflow/serving/predict