elizaos/eliza-1-training-data
elizaos/eliza-1-training-data Training corpus for the eliza-1 model line. All records are in eliza_native_v1 format — the canonical training schema for elizaOS agents. Format Every record is a JSON object with this shape: { "format": "eliza_native_v1", "boundary": "vercel_ai_sdk.generateText", "request": { "system": "...", "messages": [...], "tools": {...}, "settings": {} }, "response": { "text": "...", "finishReason": "stop"… See the full description on the dataset page: https://huggingface.co/datasets/elizaos/eliza-1-training-data.
Upload converted/gpt55-scenarios/test.jsonl with huggingface_hub
Upload converted/gpt55-scenarios/val.jsonl with huggingface_hub
Upload converted/gpt55-scenarios/train.jsonl with huggingface_hub
Update glaive validation manifest with trope-cleaning stats
Add xlam validation manifest
Re-upload glaive with extended trope stripping (38641 records cleaned, 12 legitimate mid-text occurrences remain)
Add Salesforce xlam-function-calling-60k converted to eliza_native_v1 (60000 records)
Remove old 18-record broken glaive file; correct 341k-record version is at converted/glaive/
Add openclaw validation report
Add hermes-v1 validation report
Add openclaw-opencode converted dataset (4250 records)
Add glaive validation report
Add glaive-function-calling-v2 full conversion (341454 records, 99.8% yield)
Upload converted/hermes/hermes-3-validation.json with huggingface_hub
Upload converted/hermes/hermes-3.jsonl with huggingface_hub
Add dataset card with format documentation
Add converted/openclaw/openclaw-opencode-dataset.jsonl (eliza_native_v1 format)
Add converted/hermes/glaive-function-calling-v2.jsonl (eliza_native_v1 format)
Add converted/hermes/hermes-function-calling-v1.jsonl (eliza_native_v1 format)
Add converted/merged/test.jsonl (eliza_native_v1 format)
Add converted/merged/val.jsonl (eliza_native_v1 format)
Add converted/merged/train.jsonl (eliza_native_v1 format)
initial commit