2
我一直在關注「Tensorflow on Android」的github存儲庫。Inception5h vs Inception V4,什麼是5h
這link,顯示所有的初始模型,但不是先驗5h。 用於在github tensorflow演示應用程序使用inception5h,如圖here
new_http_archive(
name = "inception5h",
build_file = "models.BUILD",
url = "https://storage.googleapis.com/download.tensorflow.org/models/inception5h.zip",
sha256 = "d13569f6a98159de37e92e9c8ec4dae8f674fbf475f69fe6199b514f756d4364"
)
請解釋
1.爲什麼它是inception5h,而不是inceptionV5?
2.爲什麼是以上5h未在上面的型號鏈接中列出?
我[再培訓](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/image_retraining/ retrain.py)和[優化初始](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/tools/optimize_for_inference.py)開始V3。再培訓APK的推理時間爲1200〜2000ms,初始5H模型爲200〜400ms。在我的再培訓APK中,是否有可能獲得類似的推理時間,初始值爲V3? –
Andrew Harp 感謝有關InceptionV5的信息。我已經嘗試重新訓練/微調[inceptionV1](https://github.com/tensorflow/models/blob/master/slim/scripts/finetune_inception_v1_on_flowers.sh)苗條的模型,並面臨[這個問題](http:// stackoverflow .com/questions/43647446/no-operation-named-input-in-the-graph-error-while-fine-tuning-retraining-ince)請幫助我。 –