JSON Lines
A text data format in which each line contains one JSON object, commonly used to store structured logs and datasets.
Timeline (last 90 days)
The trajectory produced by each Agent run (including user input, model responses, tool-call parameters and returned results) can be stored in a standardized format (such as JSON Lines or Parquet) with metadata attached, so that it can be used to build evaluation sets, fine-tuning datasets or regression test benchmarks.
Multimodal fine-tuning datasets are typically organized in JSON Lines (.jsonl) format, with each sample containing an image path (or Base64 encoding) and the corresponding multi-turn dialogue text.
Many developers adopt structured log formats (such as JSON Lines) to facilitate subsequent data analysis using tools like Python.
All Facts (3)
The trajectory produced by each Agent run (including user input, model responses, tool-call parameters and returned results) can be stored in a standardized format (such as JSON Lines or Parquet) with metadata attached, so that it can be used to build evaluation sets, fine-tuning datasets or regression test benchmarks.
50%UnverifiedMultimodal fine-tuning datasets are typically organized in JSON Lines (.jsonl) format, with each sample containing an image path (or Base64 encoding) and the corresponding multi-turn dialogue text.
50%UnverifiedMany developers adopt structured log formats (such as JSON Lines) to facilitate subsequent data analysis using tools like Python.
50%