# Changelog

## v0.1.23 — May 2026

### New: Dataset linting API

Validate dataset quality before training with `lr.datasets.linter`. Run all rules or a specific subset on any dataset, with live progress display in notebooks. Use `display_lint_overview` and `display_lint_detailed` to inspect results, and `get_lint_affected_sample_ids` to extract flagged sample IDs for filtering.

See [Datasets — Linting](/python-sdk/dataset-generation/datasets.md#linting).

### New: Reasoning comparison for evals

Compare the reasoning quality of two models side-by-side using an LLM judge during evaluation. Pass `ReasoningComparisonOptions` to `lr.evals.run` or `lr.evals.create`, or use the `reasoning_comparison_sample_size` shorthand on `run_from_training_job`. The judge model, sample count, and instructions are all configurable.

See [Evaluation — Reasoning Comparison](/python-sdk/fine-tuning-beta/evaluation.md#reasoning-comparison).

### New: Eval result download and loading

Download per-model eval rollout results as Parquet files with `lr.evals.download_results`, or load them directly into pandas DataFrames with `lr.evals.load_results`.

See [Evaluation — Downloading Results](/python-sdk/fine-tuning-beta/evaluation.md#downloading-results).

## v0.1.22 — April 2026

### Breaking: training config split (GRPO vs SFT)

The single `TrainingConfig` export is removed. Use **`GRPOTrainingConfig`** for GRPO / forward-looking training (same hyperparameters as before, including `num_rollouts` and `max_response_length`) and **`SFTTrainingConfig`** for supervised fine-tuning (`epochs`, `resume_from`, and shared LoRA fields; no rollouts or max response length).

`lr.training.create`, `estimate_cost`, and `run` accept either config type. `TrainingJob.config` from the API remains a discriminated union of the generated API models.

See [Training](/python-sdk/fine-tuning-beta/training.md) for field tables.

### Breaking: `evals.run` takes dataset and models; training defaults are `run_from_training_job`

`lr.evals.run(dataset, models)` creates an eval job, waits, and shows live progress (same as before minus model inference from the training job). For the previous behavior—base + fine-tuned from a completed **`TrainingJob`**—use **`lr.evals.run_from_training_job(config, job, dataset, *, extra_models=None)`**. **`SFTTrainingConfig`** raises `NotImplementedError` from `run_from_training_job` until SFT eval metrics exist; use `lr.evals.run(dataset, models)` or `lr.evals.create(...)` with an explicit model list.

See [Evaluation](/python-sdk/fine-tuning-beta/evaluation.md).

### New: SFT getting-started notebook

[notebooks/getting\_started/06\_sft\_training.ipynb](https://github.com/lightning-rod-labs/lightningrod-python-sdk/blob/main/notebooks/getting_started/06_sft_training.ipynb) walks through hosted SFT with `SFTTrainingConfig`.

### Docs

The [content-learning agent examples](https://github.com/lightning-rod-labs/lightningrod-python-sdk/blob/main/agent-docs/content-learning-examples.md) SFT section now uses the Lightning Rod training API instead of a raw Tinker-only loop.

## v0.1.21 — April 2026

### New: `KeyDeduplication`

Remove near-duplicate questions from your pipeline with exact or fuzzy field matching. Runs after question generation, before labeling. Pass `KeyDeduplication()` to `QuestionPipeline(deduplication=...)` to enable.

Default behavior matches on `question_text` (90% similarity) and `date_close` (exact). Customize with `KeyMatchConfig` to control which fields are compared and their similarity thresholds.

See [Deduplication](/python-sdk/dataset-generation/deduplication.md).

## v0.1.19 — April 2026

### New: `ContinuousValueOnlyAnswerType`

A new answer type for questions that expect a single scalar point estimate (e.g. `42.5`) rather than a full `{mean, stddev}` distribution. Scored via `CONTINUOUS_VALUE_ONLY_LOG_SCORE`. Use `ContinuousAnswerType` when uncertainty-aware predictions are needed; use `ContinuousValueOnlyAnswerType` when you want a single number.

See [Answer Types](/python-sdk/dataset-generation/answer-types.md#continuousvalueonlyanswertype).

### New: `CsvSeedGenerator`

Generate seeds from a CSV file uploaded via `lr.files.upload()`. Each row becomes a seed. Configure which column maps to seed text, labels, and dates.

See [Seed Generators](/python-sdk/dataset-generation/seed-generators.md#csvseedgenerator).

### New: `TopicTreeSeedGenerator`

Generate diverse seeds by recursively decomposing broad topics into specific subtopics. An LLM breaks each root topic into `tree_degree` subtopics, repeated `tree_depth` levels deep. Produces `tree_degree^tree_depth` seeds per root topic — useful for synthetic data generation without a news or document source.

See [Seed Generators](/python-sdk/dataset-generation/seed-generators.md#topictreeseedgenerator).

### New: `FileSetDocumentContextGenerator`

A new context generator that resolves a **single document** by temporal ordering, downloads its full text, and appends it as context. Supports optional LLM processing before injection and a character limit. Use this instead of `QdrantContextGenerator` when you want the complete text of one specific document rather than RAG chunks from multiple documents.

See [Labeling and Context](/python-sdk/dataset-generation/labeling-and-context.md#filesetdocumentcontextgenerator).

### New: `FileSetDocumentLabeler`

A new labeler that resolves a **single document** by temporal ordering and uses an LLM to extract a structured label from its full text. Use this instead of `QdrantRAGLabeler` when labeling from the complete content of one document (e.g. Federal Reserve Beige Book reports).

See [Labeling and Context](/python-sdk/dataset-generation/labeling-and-context.md#filesetdocumentlabeler).

### Updated: `TemporalConstraint` — new values

`TemporalConstraint` now has five values (previously two). The additions enable single-document resolution:

* `NEXT_DOCUMENT` — first document after the seed timestamp
* `PREVIOUS_DOCUMENT` — most recent document before the seed timestamp
* `EQUAL` — document with an exact matching date

These are primarily used with `FileSetDocumentContextGenerator` and `FileSetDocumentLabeler`.

See [Labeling and Context](/python-sdk/dataset-generation/labeling-and-context.md#temporalconstraint).

### Updated: Multi-model evals and intermediate checkpoint access

The evals API has been updated to accept a `list[EvalModel]` instead of a single `model_id`, enabling multiple models to be evaluated in a single job. The new `EvalModel` class accepts a `model_id` and an optional `label` for display.

Training jobs now expose `model_id_by_step` — a dict mapping training step numbers to intermediate checkpoint model IDs, enabling evaluation of checkpoints before the final model.

**Before:**

```python
lr.evals.run(model_id=job.model_id, dataset=test_dataset, benchmark_model_id="openai/gpt-4.1")
```

**After:**

```python
from lightningrod import EvalModel

lr.evals.run(
    test_dataset,
    [
        EvalModel(model_id=job.model_id, label="fine-tuned"),
        EvalModel(model_id="openai/gpt-4.1", label="baseline"),
    ],
)
```

See [Evaluation](/python-sdk/fine-tuning-beta/evaluation.md).

### New: Example builder utilities

Three helper functions for building formatted question example strings to pass as `examples` / `bad_examples` in question generators:

* `binary_example(question, comment=None)`
* `continuous_example(question, comment=None)`
* `multiple_choice_example(question, options, label=None, comment=None)`

```python
from lightningrod import binary_example, continuous_example, multiple_choice_example
```

See [Answer Types — Example Builder Utilities](/python-sdk/dataset-generation/answer-types.md#example-builder-utilities).


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.lightningrod.ai/python-sdk/getting-started/changelog.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.