Skip to content

A pure Julia implementation to read/write TFRecord

License

Notifications You must be signed in to change notification settings

pilgrimygy/TFRecord.jl

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

34 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TFRecord

CI

Usage

Install

julia> ] add TFRecord

Write TFRecord

using TFRecord

n = 10
f1 = rand(Bool, n)
f2 = rand(1:5, n)
f3 = rand(("cat", "dog", "chicken", "horse", "goat"), n)
f4 = rand(Float32, n)

TFRecord.write(
    "example.tfrecord",
    (
        Dict(
            "feature1" => f1[i],
            "feature2" => f2[i],
            "feature3" => f3[i],
            "feature4" => f4[i],
        )
        for i in 1:n
    )
)

Here we write 10 observations into the file example.tfrecord. Internally each dictionary is converted into a TFRecord.Example first, which is a known prototype by TensorFlow. Note that the type of key must be AbstractString and the type of value can be one of the following types:

  • Bool, Int64, Float32, AbstractString
  • Vector of the above types

For customized data types, you need to convert it into TFRecord.Example first.

Read TFRecord

for example in TFRecord.read("example.tfrecord")
    println(example)
end

Please refer to test/runtest.jl to get the content encoded in the example.

For more fine-grained control, please read the doc:

julia> ? TFRecord.read

julia> ? TFRecord.write

About

A pure Julia implementation to read/write TFRecord

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Julia 100.0%