I4h workflow dataset annotate vs I4h workflow finetune
I4h workflow dataset annotate is an open-source agent skill for ai models and llm tooling. I4h workflow finetune is an open-source agent skill for ai models and llm tooling.
- I4h workflow dataset annotate is best for label every episode in a recording as success or failure based on a task goal.
- I4h workflow finetune is best for fine-tune a gr00t policy for a surgical robotics task on my lerobot dataset.
- What it is
- I4h workflow dataset annotate is an open-source agent skill for ai models and llm tooling.I4h workflow finetune is an open-source agent skill for ai models and llm tooling.
- Best for
- Label every episode in a recording as success or failure based on a task goalFine-tune a GR00T policy for a surgical robotics task on my LeRobot dataset
- Kind
- SkillSkill
- Pricing
- Open sourceOpen source
- Platforms
- Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, Windsurf, GooseClaude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, Windsurf, Goose
- GitHub stars
- 4K4K
- Last commit
- Nov 2026Nov 2026
- License
- Apache-2.0Apache-2.0
- Open source
- Good for
- Label every episode in a recording as success or failure based on a task goal; Filter a workflow HDF5 file down to episodes that visually meet a criterion; Dry-run camera sampling before paying for a vision model passFine-tune a GR00T policy for a surgical robotics task on my LeRobot dataset; Train an openpi policy for 200 steps with batch size 32; Run a dry-run of the training config before starting a long job
What I4h workflow dataset annotate does
Episode visual grading
Scores each episode in an HDF5 recording against a plain-language success criterion using sampled camera frames.
Dry-run frame sampling
Checks camera coverage and sampled frames without transmitting any images to the model.
Optional episode filtering
Writes a filtered HDF5 file containing only the episodes that pass when filtering is requested.
Endpoint configuration handling
Maps caller-provided vision endpoint and model settings, or starts a local vLLM service, without printing credentials.
Output verification
Inspects annotator summaries and filtered files to reconcile pass and fail counts with the episode total.
What I4h workflow finetune does
Workflow and task resolution
Finds the selected workflow run mode and the remote task manifest that owns the training configuration.
Dataset compatibility checks
Verifies that a LeRobot dataset matches the task's embodiment, cameras, task text and feature widths.
Dry-run configuration review
Resolves the full training configuration before a long run so settings can be inspected.
Foreground training workflow
Runs the project's train entry point with explicit steps, batch size and GPU count.
Checkpoint load verification
Performs a bounded preload through the inference backend to confirm the checkpoint loads.
Disk usage reporting
Reports the size of the task output and selected checkpoint for storage planning.
I4h workflow dataset annotate vs I4h workflow finetune: questions
Is I4h workflow dataset annotate or I4h workflow finetune cheaper?
Both are open source. Compare current plans on each official site.
What is the difference between I4h workflow dataset annotate and I4h workflow finetune?
I4h workflow dataset annotate is an open-source agent skill for ai models and llm tooling. I4h workflow finetune is an open-source agent skill for ai models and llm tooling.
