如何使用 SavedModel 在 Tensorflowjs 中读取 predict() 结果

Question

代码使用 tfjs-node:

const model = await tf.node.loadSavedModel(modelPath);
const data = fs.readFileSync(imgPath);
const tfimage = tf.node.decodeImage(data, 3);
const expanded = tfimage.expandDims(0);
const result = model.predict(expanded);
console.log(result);
for (r of result) {
   console.log(r.dataSync());
}

输出：

(8) [Tensor, Tensor, Tensor, Tensor, Tensor, Tensor, Tensor, Tensor]
Float32Array(100) [48700, 48563, 48779, 48779, 49041, 48779, ...]
Float32Array(400) [0.10901492834091187, 0.18931034207344055, 0.9181075692176819, 0.8344497084617615, ...]
Float32Array(100) [61, 88, 65, 84, 67, 51, 62, 20, 59, 9, 18, ...]
Float32Array(9000) [0.009332209825515747, 0.003941178321838379, 0.0005068182945251465, 0.001926332712173462, 0.0020033419132232666, 0.000742495059967041, 0.022082984447479248, 0.0032682716846466064, 0.05071520805358887, 0.000018596649169921875, ...]
Float32Array(100) [0.6730095148086548, 0.1356855034828186, 0.12674063444137573, 0.12360832095146179, 0.10837388038635254, 0.10075071454048157, ...]
Float32Array(1) [100]
Float32Array(196416) [0.738592267036438, 0.4373246729373932, 0.738592267036438, 0.546840488910675, -0.010780575685203075, 0.00041256844997406006, 0.03478313609957695, 0.11279871314764023, -0.0504981130361557, -0.11237315833568573, 0.02907072752714157, 0.06638012826442719, 0.001794634386897087, 0.0009463857859373093, ...]
Float32Array(4419360) [0.0564018189907074, 0.016801774501800537, 0.025803595781326294, 0.011671125888824463, 0.014013528823852539, 0.008442580699920654, ...]

如何读取对象检测的 predict() 响应？我期待一本包含 num_detections、detection_boxes、detection_classes 等的字典，如 here.

所述

我也试过使用 tf.execute()，但它抛出以下错误：UnhandledPromiseRejectionWarning: Error: execute() of TFSavedModel is not supported yet.

我正在使用从 here 下载的 efficientdet/d0。

Answer 1

当您使用 dataSync() 下载张量时，它只会保留值。如果你想要一个没有张量的描述每个结果的对象，你只需要 console.log(result)。然后你在浏览器控制台中扩展你的日志结果，它应该 return 像这样：

Tensor {
  "dataId": Object {},
  "dtype": "float32",
  "id": 160213,
  "isDisposedInternal": false,
  "kept": false,
  "rankType": "2",
  "scopeId": 365032,
  "shape": Array [
    1,
    3,
  ],
  "size": 3,
  "strides": Array [
    3,
  ],
}

您的 console.log(result) 的输出中包含 8 tensors，这表明它是正确的。您正在遍历每个结果，每个输出都应遵循以下格式：

['num_detections', 'detection_boxes', 'detection_classes', 'detection_scores', 'raw_detection_boxes', 'raw_detection_scores, 'detection_anchor_indices', 'detection_multiclass_scores']

如何使用 SavedModel 在 Tensorflowjs 中读取 predict() 结果

How to read predict() result in Tensorflowjs using a SavedModel

node.js

tensorflow

object-detection

tensorflow.js

efficientnet