package grpc.training.data.v1

Get desktop application:
View/edit binary Protocol Buffers messages

Service to store evaluation metrics

===== GET ENDPOINTS, for external fetch =============

• rpc GetLogs (Query, stream LogLine)

  training_data.proto:32

  Get loglines, based on query

• rpc GetEMetrics (Query, stream EMetrics)

  training_data.proto:35

  Get evaluation metrics records, based on query

• rpc AddEMetrics (EMetrics, AddResponse)

  training_data.proto:41

  Add evaluation metrics record

• rpc AddLogLine (LogLine, AddResponse)

  training_data.proto:45

  Add log line record

• rpc AddEMetricsBatch (EMetricsBatch, AddResponse)

  training_data.proto:49

  Add evaluation metrics record

  message EMetricsBatch

  training_data.proto:128

  • bool force = 1

  • repeated EMetrics emetrics = 2

• rpc AddLogLineBatch (LogLineBatch, AddResponse)

  training_data.proto:53

  Add log line record

  message LogLineBatch

  training_data.proto:96

  • bool force = 1

  • repeated LogLine logLine = 2

• rpc DeleteEMetrics (Query, DeleteResponse)

  training_data.proto:57

  Delete all evaluation metrics belonging to a training job or user id

• rpc DeleteLogLines (Query, DeleteResponse)

  training_data.proto:61

  Delete all log lines belonging to a training job or user id

• rpc DeleteJob (Query, DeleteResponse)

  training_data.proto:65

  Delete all log lines belonging to a training job or user id

• rpc Hello (Empty, HelloResponse)

  training_data.proto:69

  ===== In case you want it to say "Hello" =========

  message Empty

  training_data.proto:176

  (message has no fields)

  message HelloResponse

  training_data.proto:172

  • string msg = 1

***

Used as response type in: TrainingData.AddEMetrics, TrainingData.AddEMetricsBatch, TrainingData.AddLogLine, TrainingData.AddLogLineBatch

• bool success = 1

Used in: EMetrics

• Any.DataType type = 1

• string value = 2

Used in: Any

• STRING = 0

• JSONSTRING = 1

• INT = 2

• FLOAT = 3

• string training_id = 1

  The following two options are exclusive

• string user_id = 2

Used as response type in: TrainingData.DeleteEMetrics, TrainingData.DeleteJob, TrainingData.DeleteLogLines

• bool success = 1

Used as request type in: TrainingData.AddEMetrics

Used as response type in: TrainingData.GetEMetrics

Used as field type in: EMetricsBatch

• optional MetaInfo meta = 1

• map<string, Any> etimes = 2

  Repeated, order-dependent list of temporal keys Example: {"iteration": 209}

• string grouplabel = 3

  Group label, such as test, train, or validate

• map<string, Any> values = 4

  / {"cross_entropy": 0.4430539906024933, "accuracy": 0.8999999761581421}

Used as request type in: TrainingData.AddLogLine

Used as response type in: TrainingData.GetLogs

Used as field type in: LogLineBatch

• optional MetaInfo meta = 1

• string line = 2

Used in: EMetrics, LogLine, Query

• string training_id = 1

  Unique id identifying the training job

• string user_id = 2

  Unique id identifying the user

• int64 time = 3

  Time that the metric occured: representing the number of millisecond since midnight January 1, 1970.

• int64 rindex = 4

  Sequential index

• string subid = 5

  Optional subid

Playing with semi-generalized query request.

Used as request type in: TrainingData.DeleteEMetrics, TrainingData.DeleteJob, TrainingData.DeleteLogLines, TrainingData.GetEMetrics, TrainingData.GetLogs

• Query.SearchType searchType = 1

• optional MetaInfo meta = 3

  The following three options are exclusive

• string since = 4

  representing the number of milliseconds since midnight January 1, 1970.

• int32 pagesize = 5

  Only get this many records

• int64 pos = 6

  The starting position. If positive or zero, count from beginning, if negative, count from end.

Used in: Query

• TERM = 0

• NESTED = 1

• MATCH = 2

• ALL = 3