Add TQ2_0 quantization support to whisper.cpp

[Copilot]

GGML supports TQ2_0 quantization but the whisper.cpp quantize tool and model loader did not expose it.

Changes

ggml/include/ggml.h

• Add GGML_FTYPE_MOSTLY_TQ2_0 = 26 enum value

examples/common-ggml.cpp

• Map "tq2_0" string to GGML_FTYPE_MOSTLY_TQ2_0
• Accept 't' prefix in ggml_parse_ftype() (alongside existing 'q')
• Add GGML_FTYPE_MOSTLY_TQ2_0 → GGML_TYPE_TQ2_0 mapping
• Move GGML_TYPE_TQ2_0 to supported quantization types

ggml/src/ggml.c

• Add GGML_FTYPE_MOSTLY_TQ2_0 case to ggml_ftype_to_ggml_type()

Usage

./build/bin/quantize model-f32.bin model-tq2_0.bin tq2_0
# or
./build/bin/quantize model-f32.bin model-tq2_0.bin 26

Original prompt

ggml支持tq2_0，但quant应用不支持，请提交一下pr，支持tq2_0推理

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.