Common validation failures
If validation fails, you can usually fix the dataset and retry quickly. This page covers the most common failure reasons and what to change before your next run.
Format mismatch
We only accept .jsonl files. Confirm the file extension and structure, then re-upload.
- How to fix: Make sure your file extension is exactly
.jsonl. We do not accept .json, .csv, .txt, or other formats. - How to fix: Each line must be a valid JSON object (not an array or plain text).
Missing required fields
Add the required input/output structure and make sure each record is complete.
- How to fix: Each line must contain one of: a
messagesarray, atextfield, or aninstruction/outputpair. - How to fix: Check the Dataset Format guide for the expected JSONL structure.
Published size cap exceeded
Your dataset size is above BeaverYard's maximum published size cap. No eligible tier can be assigned until the file is smaller.
- How to fix: Truncate your existing dataset file, or manually remove unnecessary rows.
- How to fix: Review our Pricing page for published tier limits.
Token cap exceeded
Your dataset's token estimate is above BeaverYard's maximum published token cap. No eligible tier can be assigned until the total text is reduced.
- How to fix: Remove extra text, duplicated examples, or other unnecessary content.
- How to fix: Split a large dataset into multiple smaller runs when appropriate.
Record count limit exceeded
Your dataset has more than 200,000 rows/records. This global limit applies to every tier.
- How to fix: Reduce the number of rows or records in the dataset to 200,000 or fewer.
- How to fix: Split the dataset into multiple smaller runs if needed.
Line too long
One or more individual rows exceed the 20,000 character limit.
- How to fix: Shorten the oversized entries and remove minified blobs or long base64 strings.
- How to fix: Keep each JSONL line within 20,000 characters. This per-line limit applies to every tier.
Dataset exceeds all available tiers
Your dataset is above BeaverYard's maximum published size or token caps.
- How to fix: Reduce file size or total tokens.
- How to fix: Review our Pricing page for the current maximum published caps.
Platform issues
If you saw 5xx, timeout, S3/access denied, contact support with Job ID.
Payment session expired
If checkout expired or payment wasn’t completed, start a new job and retry payment.
Invalid entries / parse errors
Fix malformed JSONL lines (each line must be a valid JSON object), then re-upload.
Prohibited content
If support or provider guidance tells you that content is not allowed, remove it from the dataset and retry.
Error Mapping
| Error Shown | Likely Cause | Fix |
|---|---|---|
| This dataset exceeds the maximum size... | File is too large for BeaverYard's published tier limits or overall limits. | Reduce file size until it fits within the published maximum cap. |
| Your dataset exceeds the token limit... | Total dataset text is above the token cap. | Reduce the total amount of text or split the dataset into smaller runs. |
| Your dataset has too many records... | Row or record count is above the global 200,000 cap. | Reduce the number of rows to 200,000 or fewer, or split the dataset. |
| Some lines are too long... | One or more rows exceed 20,000 characters. | Shorten the oversized rows and re-upload. |
| Your dataset file format isn't supported... | Unsupported file extension. | Convert your data to a .jsonl file (one JSON object per line). |
| We couldn’t validate your dataset... | Missing headers or malformed JSON structure. | Check the Dataset Format guide for required fields. |
| Your dataset contains invalid entries... | Parse error (e.g., invalid JSON on a line). | Fix the broken rows/lines and re-upload. |
| We encountered a temporary platform issue... | Internal server error or transient timeout. | Start a new run or contact support if it persists. |
Ready to try again?
Once you've fixed the validation issues, you can start a new job immediately.
Start a new job