Skip to main content

Logs

Upload, tag, and delete training logs in AI Studio.

This article explains how to manage training logs in Hyperstack AI Studio. It covers file upload via signed URLs, saving logs in real time, deleting logs, required formats like JSONL, and tagging strategies to organize your data effectively.

Prefer the API?

You can also upload and manage training logs programmatically. See the Training Data API reference for all log operations, including request formats, parameters, and example responses.

Manage Logs​

Upload Logs​

You can upload logs directly through the AI Studio UI using the following steps:

  1. Open the Logs Page

    Navigate to the Logs & Datasets page.

  2. Upload Your .jsonl File

    Ensure your logs meet the required file format as outlined in the JSONL File Format guidelines.

    Click the Upload Logs button in the top-right corner, then either select your .jsonl file from your device or drag and drop it into the upload area.

  3. Add Tags

    Enter at least one tag to help categorize your logs (e.g., testing).

  4. Validate and Upload

    Click Validate & Upload.

    The system will automatically check your file format and structure, then upload the logs if validation succeeds.

Once uploaded, you can:

  • View and manage logs individually.
  • Associate logs with datasets for training.
  • Add metadata tags to improve filtering and traceability.

Delete Logs​

You can delete logs directly through the AI Studio UI using the following steps:

  1. Open the Logs & Datasets Page

    Navigate to the Logs & Datasets page and switch to the Logs tab.

  2. Filter Logs for Deletion

    Use the filtering options to select logs based on:

    • Model
    • Tags

    To delete individual logs, first apply a unique tag to only those logs, then filter by that tag.

  3. Delete Selected Logs

    Once the desired logs have been selected, click the Delete button to remove them permanently.

Permanent Data Removal

Deleting logs is irreversible—once deleted, the data cannot be recovered.

JSONL File Format​

The JSONL file must contain one JSON object per line, where each object represents a conversation or interaction. Here's an example of the expected format:

{"messages": [{"role": "user", "content": "What's the capital of Australia?"}, {"role": "assistant", "content": "The capital of Australia is Canberra."}]}
{"messages": [{"role": "system", "content": "You are a travel advisor."}, {"role": "user", "content": "Where should I go in Europe for a summer vacation?"}, {"role": "assistant", "content": "Consider Italy, Spain, or Greece—they offer great weather, food, and culture in the summer!"}]}
{"messages": [{"role": "user", "content": "What's the capital of Australia?"}, {"role": "assistant", "content": "The capital of Australia is Canberra."}]}

Each line in the JSONL file must be a valid JSON object containing:

  • messages: An array of message objects
  • Each message object must have:
    • role: Either "system", "user", or "assistant"
    • content: The text content of the message

Make sure your JSONL file:

  • Has one complete JSON object per line
  • Uses proper JSON formatting
  • Contains the required fields for each message
  • Has no trailing commas
  • Uses UTF-8 encoding
Context Length Limit

The total number of tokens in your input (system prompt, prior messages, and current prompt) plus the model’s generated output must not exceed the maximum context length of 8192 tokens.


Back to top

On this page