feat(studio): add model selector to Safe Synthesizer creation [ASTD-19]

[marcusds]

Summary

Adds a model dropdown under Training Data Configuration in the Safe Synthesizer creation flow (ASTD-19). Bound to spec.config.training.pretrained_model.

Options (in ticket order):

1. HuggingFaceTB/SmolLM3-3B (default)
2. mistralai/Mistral-7B-Instruct-v0.3
3. TinyLlama/TinyLlama-1.1B-Chat-v1.0

Info tooltip: "Model that gets fine-tuned to generate synthetic data."

Changes

• constants.ts — PRETRAINED_MODEL_OPTIONS + DEFAULT_PRETRAINED_MODEL.
• schema.ts — added pretrained_model to the training config + default. Flows into the submitted job payload automatically.
• AdvancedParameters.tsx — ControlledSelect at the top of the Training Data Configuration section.
• AdvancedParameters.test.tsx — tests for render, default (SmolLM3), option order, and selection updating the config.

Acceptance criteria

• Selector shown under Training Data Configuration
• Three options, in the specified order
• Defaults to SmolLM3-3B when untouched
• Saved config reflects the selection

Note

Backend pretrained_model is a free-form string; only SmolLM3-3B is currently pre-provisioned as a model fileset (plugins/nemo-safe-synthesizer/scripts/setup_model_filesets.py). Mistral-7B and TinyLlama resolve via HF at job time and may require provisioning before they run reliably.

Testing

pnpm --filter nemo-studio-ui test src/routes/SafeSynthesizerNewRoute → 71 passed.

Summary by CodeRabbit

• New Features

  • Added a pretrained model selector to the Safe Synthesizer “Training Data Configuration”.
  • The selector offers three supported options and applies a default when none is chosen.
  • The selected pretrained model is included in the saved synthesizer configuration.

• Bug Fixes

  • Ensured the pretrained model field is part of schema validation and form default behavior.

• Tests

  • Expanded coverage for the default selection, available options order, and updating the watched value after user changes.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 9f2c5e79-4d74-44a5-8ba4-42bf7c61c927

📥 Commits

Reviewing files that changed from the base of the PR and between 8e55735 and 945529a.

📒 Files selected for processing (1)

• web/packages/studio/src/routes/SafeSynthesizerNewRoute/schema.ts

🚧 Files skipped from review as they are similar to previous changes (1)

• web/packages/studio/src/routes/SafeSynthesizerNewRoute/schema.ts

---

📝 Walkthrough

Walkthrough

Adds pretrained-model constants and defaults, extends the Safe Synthesizer schema, renders a model selector in advanced parameters, and tests default, option-list, and selection-update behavior.

Changes

Pretrained Model Selection

Layer / File(s)	Summary
Model constants and form contract web/packages/studio/src/routes/SafeSynthesizerNewRoute/constants.ts, web/packages/studio/src/routes/SafeSynthesizerNewRoute/schema.ts	Defines supported models, derives the default model, and adds the optional pretrained-model schema field and form default.
Training configuration selector web/packages/studio/src/routes/SafeSynthesizerNewRoute/components/AdvancedParameters.tsx	Adds a controlled selector bound to the pretrained-model form field and populated from the supported model options.
Selector behavior coverage web/packages/studio/src/routes/SafeSynthesizerNewRoute/components/AdvancedParameters.test.tsx	Tests selector rendering, default state, option ordering, and form updates after selecting another model.

Sequence Diagram(s)

sequenceDiagram
  participant User
  participant AdvancedParameters
  participant ControlledSelect
  participant FormState
  User->>AdvancedParameters: open Training Data Configuration
  AdvancedParameters->>ControlledSelect: render supported model options
  ControlledSelect->>FormState: initialize default pretrained_model
  User->>ControlledSelect: select another model
  ControlledSelect->>FormState: update pretrained_model

Loading

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly matches the main change: adding a model selector to Safe Synthesizer creation.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch astd-19-add-model-selector-training-data-config/mschwab

---

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@web/packages/studio/src/routes/SafeSynthesizerNewRoute/schema.ts`:
- Line 106: Update the pretrained_model field in the schema to validate against
the same supported-model tuple used by the selector, rather than accepting any
string. Reuse the existing PRETRAINED_MODEL_OPTIONS symbol from the
SafeSynthesizerNewRoute implementation and preserve optionality.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: d5be11d0-bd74-46ea-a012-4a9257b80158

📥 Commits

Reviewing files that changed from the base of the PR and between 26319c9 and 8e55735.

📒 Files selected for processing (4)

• web/packages/studio/src/routes/SafeSynthesizerNewRoute/components/AdvancedParameters.test.tsx
• web/packages/studio/src/routes/SafeSynthesizerNewRoute/components/AdvancedParameters.tsx
• web/packages/studio/src/routes/SafeSynthesizerNewRoute/constants.ts
• web/packages/studio/src/routes/SafeSynthesizerNewRoute/schema.ts

[github-actions]

Suite	Lines Covered	Line Rate	Branch Rate
Unit Tests	27146/34858	77.9%	62.1%
Integration Tests	15964/33570	47.5%	20.0%