feat: support finetuning and evaluation of VLA models

[learnerljh]

Description

Please include a summary of the changes and which issue is fixed. Include relevant motivation and context.

Fixes # (issue)

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Code refactoring
• Performance improvement
• Test coverage improvement

Checklist

Go over all the following points, and put an x in all the boxes that apply.
If you are unsure about any of these, don't hesitate to ask. We are here to help!

• I have read the CONTRIBUTION guide. (required)
• My change requires a change to the documentation.
• I have updated the tests accordingly. (required for a bug fix or a new feature)
• I have updated the documentation accordingly.
• I have reformatted the code using make format. (required)
• I have checked the code using make lint. (required)
• I have ensured make test pass. (required)

Testing

Please describe the tests that you ran to verify your changes:

• Test A
• Test B

[Copilot]

Pull request overview

This PR adds comprehensive support for finetuning and evaluation of Vision-Language-Action (VLA) models, including OpenPI, OpenVLA, UniVLA, and SmolVLA. The changes introduce a unified CLI interface, model-specific training/evaluation configurations, and extensive infrastructure for model serving and benchmarking.

Key changes:

• Unified CLI interface (vla-arena train/eval) for all VLA models with dynamic model loading
• OpenPI model integration with JAX-based training, Docker deployment, and websocket-based policy serving
• Training and evaluation configurations for OpenPI, OpenVLA (with OFT variant), UniVLA, and SmolVLA
• Removed legacy evaluation utilities and policy implementations in favor of model-specific evaluators

Reviewed changes

Copilot reviewed 124 out of 1401 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
vla_arena/cli/main.py	CLI entry point with train/eval subcommands
vla_arena/cli/train.py	Training orchestrator supporting PyTorch (torchrun) and JAX models
vla_arena/cli/eval.py	Evaluation launcher for model-specific evaluators
vla_arena/models/openpi/*	Complete OpenPI integration including scripts, Docker configs, and client libraries
vla_arena/configs/train/*.yaml	Training configurations for all supported models
vla_arena/configs/evaluation/*.yaml	Evaluation configurations for all supported models
vla_arena/evaluation/*	Removed legacy evaluation utilities and policy base classes
tests/*	Updated test structure and fixtures

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[muchvo]

LGTM.