Update model card with v2 benchmark results (MMLU-TR 0.6789, XCOPA-TR 0.690)

Update to v7.1 SFT — improved MMLU_TR (+2.71) and XCOPA_TR (+1.00)

Enhance model card with detailed benchmarks, usage examples, and training details

Add Turkish-LLM-32B-Instruct model

Add Turkish-LLM-32B-Instruct model

initial commit