GPT-QModel v7.2.0

What's Changed

• [MODEL] support gemma4_unified by @HaozheZhang6 in #2921
• feat(models): register gemma4_unified_text text-only GPTQ definition by @Anai-Guo in #2925
• Fix Mllama (Llama3.2-11B-Vision) quantization by @JeevanBhoot in #2926
• Bump actions/checkout from 6 to 7 in the github-actions group by @dependabot[bot] in #2927
• [MODEL] support glm4v_moe_text, llama4_text and mllama_text_model by @ZX-ModelCloud in #2912
• [MODEL] support hy_v3 and ministral3 by @ZX-ModelCloud in #2911
• [MODEL] support cohere2_moe by @ZX-ModelCloud in #2929
• support minimax_m3_vl by @ZX-ModelCloud in #2930
• [MODEL] support lfm2 by @ZX-ModelCloud in #2932
• [MODEL] support lfm2_vl by @ZX-ModelCloud in #2933
• Fix Mllama GPTQ loading for skipped cross-attention layers by @JeevanBhoot in #2934
• Bump version from 7.1.0 to 7.2.0 by @Qubitium in #2935

New Contributors

• @HaozheZhang6 made their first contribution in #2921
• @JeevanBhoot made their first contribution in #2926

Full Changelog: v7.1.0...v7.2.0