02 2024 档案
kubernetes loadbalancer: metallb
摘要：https://zhuanlan.zhihu.com/p/617807098
阅读全文
run kubeflow/serving on a kubernetes cluster
摘要：1. build kubeflow/serving container image which contains serving_model [1] # run container tensorflow/serving, the image of this container # is the ba
阅读全文
tensorflow/serving: REST request
摘要：1. save the trained model # in module file of tfx component trainer def _apply_preprocessing(raw_features, tft_layer): transformed_features = tft_laye
阅读全文
tensorflow distributed training in tfx pipeline run by kubeflow
摘要：1. deploy worker, parameter server on kubernetes cluster 1.1 build container image of worker, parameter server $ git clone https://github.com/tensorfl
阅读全文
Debug: tf distribute strategy parameter server: tfx component trainer: model.save(): failed to connect to all addresses
摘要：[ERROR: tf distribute strategy parameter server: tfx component trainer: model.save(): failed to connect to all addresses] log of pod tfx-component-tra
阅读全文
Debug: tf distribute strategy parameter server: tfx component trainer: OutOfRangeError(), End of sequence
摘要：[ERROR: tf distribute strategy parameter server: tfx component trainer: OutOfRangeError(), Node: 'cond/IteratorGetNext' End of sequence] log of pod tf
阅读全文
Debug: tf distribute strategy parameter server: NOT_FOUND: No such file or directory
摘要：[ERROR: NOT_FOUND: /tfx/tfx_pv/pipelines/detect_anomolies_on_wafer_tfdv_schema/ImportExampleGen/examples/67/Split-train/data_tfrecord-00000-of-00001.g
阅读全文
Debug : kfp.Client().upload_pipeline(): Failed to start a transaction to create a new pipeline and a new pipeline version: dial tcp: lookup mysql on 10.96.0.10:53: no such host","
摘要：[ERROR: Failed to start a transaction to create a new pipeline and a new pipeline version: dial tcp: lookup mysql on 10.96.0.10:53: no such host","] >
阅读全文
Debug: tf distribute strategy parameter server: stuck at "INFO:tensorflow:ParameterServerStrategyV2 is now connecting to cluster
摘要：[ERROR: stuck at "INFO:tensorflow:ParameterServerStrategyV2 is now connecting to cluster with cluster_spec: ClusterSpec({'ps': ['dist-strat-example-ps
阅读全文
Debug: tf_ditribute_strategy_worker.yaml: unknown field "spec.template.spec.nodeAffinity"
摘要：[ERROR: unknown field "spec.template.spec.nodeAffinity"] (base) maye@maye-Inspiron-5547:~/github_repository/tensorflow_ecosystem/distribution_strategy
阅读全文
Debug: tf_ditribute_strategy_worker.yaml: resource mapping not found for name: "dist-strat-example-worker-0" namespace: "" from "maye_template.yaml": no matches for kind "Deployment" in version "v1"
摘要：[ERROR: resource mapping not found for name: "dist-strat-example-worker-0" namespace: "" from "maye_template.yaml": no matches for kind "Deployment" i
阅读全文
Debug: tf_distribute_strategy_worker.yaml: Exit Code: 132, and log of pod is empty.
摘要：[ERROR: Exit Code: 132, and log of pod is empty.] (base) maye@maye-Inspiron-5547:~/github_repository/tensorflow_ecosystem/distribution_strategy$ kubec
阅读全文
Debug: tf-distribute-strategy-worker: json.decoder.JSONDecodeError: Expecting property name enclosed in double quotes: line 1 column 182
摘要：[ERROR: json.decoder.JSONDecodeError: Expecting property name enclosed in double quotes: line 1 column 182] # in file pipeline.yaml - name: TF_CONFIG
阅读全文
Start a kubernetes cluster
摘要：1. check if service of container runtime -- containerd -- is running on all computers who want to join the kubernetes cluster. $ systemctl status cont
阅读全文
Mechanism of Machine Learning
摘要：what is Machine Learning? Terms: L : loss, expectation of differences between
Debug: javascript: while loop never exit due to state variable
摘要：1. Illustration In a javascript function, if a variable is not declared before assigning value, it is considered as a state variable of the function,
阅读全文
Run a tfx pipeline using kubeflow pipeline
摘要：1. what is kubeflow pipeline for tfx pipeline ？ kubeflow pipeline is an ochetrator of tfx pipeline, which runs on a kubernetes cluster. LocalDagRuner
阅读全文
Deploy standalone kubeflow pipeline on a kubernetes cluster
摘要：1. Computer environment OS: ubuntu 20.04 kubenetes: 1.26.12 Attention: kubeflow pipeline 2.0 is compatible up to kubernetes v1.26. 2. Prepare file kub
阅读全文
ml-pipeline-ui of kubeflow pipeline
摘要：1. Creating a pipeline on ml-pipeline-ui webpage is saving the pipeline to database mlpipeline, delete a pipeline on ml-ppeline-ui webpage is deleting
阅读全文
nerdctl build -- command to build container image from docker file
摘要：1. Prerequisite of using nerdctl build buildctl needs to be installed and buildkitd needs to be running. 2. check if buildctl installed $ nerdctl vers
阅读全文
Solution of downloading from github.com too slow
摘要：githubusercontent和github加速镜像 加速地址一览 fastgit.org：https://doc.fastgit.org/ http://cnpmjs.org：https://github.com.cnpmjs.org/ http://gitclone.com：https://
阅读全文
Install nerdctl -- cli of containerd, compatible with docker
摘要：1. Download binary package of nerdctl-full Attention: Downloading from https://github.com/containerd/nerdctl/releases may be very slow in china, and e
阅读全文
