Explore Optimizing and Running Tests in Parallel for Faster CI in `litdata`

**Title**: Explore Running Tests in Parallel for Faster CI in `litdata`

**Description**:

We should investigate the possibility of running the test suite in parallel to speed up CI and local test execution. As the test base grows, serial execution becomes a bottleneck, especially when contributing frequently or iterating on PRs.
> 🧪 An initial test was conducted in [#608](https://github.com/lit-ai/litdata/pull/608).

### Tasks:

* [ ] Review the current test setup and determine if there are any global state dependencies or side effects.
* [ ] Look into test cases optimizations
* [ ] Explore using [`[pytest-xdist](https://pypi.org/project/pytest-xdist/)`](https://pypi.org/project/pytest-xdist/) or an alternative for parallelization.
* [ ] Identify and resolve any tests that may fail due to parallel execution.
* [ ] Benchmark before/after test runtimes.

### Goal:

Improve test speed while maintaining test accuracy and reproducibility.

Start by reviewing the current test setup and the initial parallelization work in PR #608. Evaluate pytest-xdist or an alternative, check for global-state or side-effect dependencies, and benchmark serial versus parallel runs; done means faster execution with accurate and reproducible tests.

Written by the indexing model from the issue text.