Skip to content

Buckets

A bucket bundles everything a training run needs — a base model plus one or more datasets — so you can launch reproducible runs from the Training console. A bucket belongs to a pool.

Buckets list — /buckets

Buckets list

All buckets in the selected pool. Create a new bucket (name + base model), or click one to open its detail page. Each row shows its base model and dataset count.

Bucket detail — /buckets/{id}

Bucket detail

  • Header — the bucket's pool, owner, and creation date.
  • Base model — a HuggingFace model id (e.g. Qwen/Qwen2.5-3B) or a previously‑trained model id to fine‑tune from. Save to update it.
  • Datasets — attach training data two ways: Upload a file (JSONL / plain text), or reference a HuggingFace dataset by id with an optional config.
  • Trained runs — models trained from this bucket appear here (and in Inference) once a run exports.
  • Use in training — once a base model and dataset are set, this shows the generated dataset_config snippet. Copy config for the swarm CLI, or Open training console → to launch directly.