Data Normalization

Data processing and normalization tasks evaluate structured output from messy prose and different structures. This capability is essential for catalogue and product pipelines.

Duration

Input Tokens

568

Output Tokens

136

Cost

$0.00

Context

Input

From the extracted description. Extract all colors found in the clothing

Expected output

{
  "colors": [
    "blue",
    "navy",
    "black",
    "beige",
    "brown",
    "gray"
  ]
}

Model output

{
  "colors": [
    "medium blue",
    "navy",
    "medium beige",
    "brown",
    "gray",
    "silver"
  ]
}