feat: enhance training notebook for Colab compatibility and path management

feat: add T4-optimized SFT and RLVR training scripts, evaluation utilities, and updated documentation

docs: update README documentation, add architecture diagram, and include a new training notebook.

feat: add training pipeline with SFT and RLVR support for Qwen 2.5-3B-Instruct

feat: add notebook for Qwen 2.5-7B SFT and RLVR training on CORP-ENV

feat: enhance training environment and documentation for CORP-ENV models

feat: update README and runbook for SFT and GRPO training enhancements

feat: update training scripts and add judge-rerunnable notebook for CORP-ENV