mario-rc/emotional-rlaif-ppo-gemma-2-2b-it

docs: correct project repository link

Reorder released models and apply requested card layout adjustments

Standardize License section while preserving base-model requirements

Add brief limitations introduction

Specify PyTorch 2.5.1+cu124 from the local training environment

Restore all four framework fields without other card changes

Add short introductions to Model Details and Framework Versions

Check BF16 support in both examples and document evidenced framework versions

Reject unknown response tags, validate before display, and restore license files

Fix interactive response parsing and context limits; preserve both examples

Apply requested bullet and example clarifications; retain both code examples

Apply requested bullet and example clarifications; retain both code examples

Restore both usage examples and apply requested card text edits

Fix Gemma-2-2B PPO card inference and document reproducibility

Restore original model-card text with explicitly requested updates

Restore exact version preceding model-card rewrite

Record Hub-generated LFS attributes in release checksums

Preserve exact frozen reward adapter saved with PPO checkpoint

Publish verified emotional RLAIF adapter and unified 20-model card

Normalize model card tables

Upload updated SML emotional RLAIF adapter

initial commit