Fine-tuning stalls on the boring parts: bad JSONL rows, duplicate examples, and babysitting the job. This pipeline handles the whole path from a Drive folder.
Fine-tuning stalls on the boring parts: bad JSONL rows, duplicate examples, and babysitting the job. This pipeline handles the whole path from a Drive folder.
It pulls your dataset, validates and cleans the format, splits train and validation, then launches the fine-tune and watches it finish. You get back the model ID and its validation loss, so the only thing you do is point it at the folder and choose a base model.
How it runs
Used at step 01 to kick off the pipeline.
Write
Used at step 01 to kick off the pipeline.
WebFetch
Used at step 01 to kick off the pipeline.
WebSearch
Used at step 01 to kick off the pipeline.
Clean malformed rows, deduplicate examples, and split the set into train and validation.
pending
Upload the prepared file and start an OpenAI fine-tuning job with the chosen base model.
pending
Track the run to completion and report the final model ID and validation loss.
pending
Sample output
json
// Sample output
// (generated when the pipeline finishes)
Take a Drive dataset folder, validate and prepare it, launch an OpenAI fine-tune, and report the model ID and loss.
Unlock the rest
The full agent definition, install snippet, and starter task are gated for community members.
Members get the full `.md` agent file, the npm / pnpm install one-liners, a starter prompt that we've tuned against real runs, and the open-source repo when this automation ships there. One email, magic link, done.