LV.1

langweixian

23积分0赞

3 帖子 8 回复 0 收藏

TA的动态

TA的帖子

TA的回复

MLU270源码编译报错 + 哪里找.pb离线模型文件 我的回复：#1lhscau回复您好，建议使用寒武纪提供的tf容器，保证开发环境统一；目前我司不提供现成的pb文件下载。源码也编译成功了，首先遇到这个问题：ERROR: Skipping './tensorflow/tools/pip_package:build_pip_package': invalid target format './tensorflow/tools/pip_package:build_pip_package': invalid package name './tensorflow/tools/pip_package': package name component contains only '.' characters WARNING: Target pattern parsing failed. ERROR: invalid target format './tensorflow/tools/pip_package:build_pip_package': invalid package name './tensorflow/tools/pip_package': package name component contains only '.' characters INFO: Elapsed time: 2.267s INFO: 0 processes. FAILED: Build did NOT complete successfully (0 packages loaded) why this error happen?然后将 ./ 改成 //解决了，然后遇到after modify ./ to // , I try compling again, but why this happen?ERROR: /home/szu/lang/tensorflow_sdk1.7.0_v1.4.0/src/cambricon_tensorflow/tensorflow/tensorflow/python/keras/BUILD:19:1: no such package '@keras_applications_archive//': java.io.IOException: Error downloading [https://storage.googleapis.com/mirror.tensorflow.org/github.com/keras-team/keras-applications/archive/1.0.8.tar.gz, https://github.com/keras-team/keras-applications/archive/1.0.8.tar.gz] to /home/szu/.cache/bazel/_bazel_szu/3c380e50fc9541f4a8c5b18c881af31d/external/keras_applications_archive/1.0.8.tar.gz: All mirrors are down: [] and referenced by '//tensorflow/python/keras:keras'ERROR: Analysis of target '//tensorflow/tools/pip_package:build_pip_package' failed; build aborted: no such package '@keras_applications_archive//': java.io.IOException: Error downloading [https://storage.googleapis.com/mirror.tensorflow.org/github.com/keras-team/keras-applications/archive/1.0.8.tar.gz, https://github.com/keras-team/keras-applications/archive/1.0.8.tar.gz] to /home/szu/.cache/bazel/_bazel_szu/3c380e50fc9541f4a8c5b18c881af31d/external/keras_applications_archive/1.0.8.tar.gz: All mirrors are down: []INFO: Elapsed time: 89.101sINFO: 0 processes.FAILED: Build did NOT complete successfully (371 packages loaded, 14900 targets configured) currently loading: tensorflow/lite/schema；这是网络问题，通过export http_proxy="http:[your ip]:[your proxy port]", export https_proxy="http:[your ip]:[your proxy port]"解决了 0

MLU270源码编译报错 + 哪里找.pb离线模型文件 我的回复：#3lhscau回复关于该链接，您可以先测试一下，有问题的话，贴出错误日志，供我们分析，并请同时描述一下，您的需求。好的，顺便问一下，寒武纪的板子提供dvfs功能吗 0

MLU270源码编译报错 + 哪里找.pb离线模型文件 我的回复：#1lhscau回复您好，建议使用寒武纪提供的tf容器，保证开发环境统一；目前我司不提供现成的pb文件下载。哦哦谢谢，那请问https://github.com/tensorflow/models/blob/v1.13.0/research/slim/export_inference_graph.py这里的模型可以拿来生成离线模型吗 0

求MLU270的全部官方技术文档 我的回复：好的谢谢 0