ggml-quants : provide ggml_vqtbl1q_u8 for 64bit compatibility (llama/5711)

IQ3_S: a much better alternative to Q3_K (llama/5676)

ggml : always define ggml_fp16_t as uint16_t (llama/5666)

ggml : 32-bit arm compat (#1891)

sync : llama.cpp (ggml/0)

ggml : restore vec dot stride arg names (llama/5453)

ci : fix wikitext url + compile warnings (llama/5569)

1.5 bit quantization (llama/5453)

ggml-quants : fix compiler warnings (shadow variable) (llama/5472)

ggml : fix compile warnings (unused vars) (llama/4966)

ggml : add mmla kernels for quantized GEMM (llama/4966)

ggml : fix `error C2078: too many initializers` for MSVC ARM64 (llama/5404)

Slight quantization improvement for Q4_K and Q5_K (llama/5361)

iq2_xxs: tune quantization (llama/5320)

Faster AVX2 dot product for IQ2_XS (llama/5187)

SOTA 3-bit quants (llama/5196)

ggml : add IQ2 to test-backend-ops + refactoring (llama/4990)

ggml : importance matrix support for legacy quants (llama/4969)

Add ability to use importance matrix for all k-quants (llama/4930)

2-bit quantizations (llama/4897)

ggml : fix 32-bit ARM compat for IQ2_XS (#1758)

ggml : SOTA 2-bit quants (add IQ2_XS) (llama/4856)

ggml : fix vld1q_s8_x4 32-bit compat (llama/4828)

SOTA 2-bit quants (llama/4773)

ggml : add ggml_vdotq_s32 alias (llama/4715)

sync : ggml (VMM, sync-ggml-am, dotprod ARM fixes, CUDA fixes) (#1691)

sync : ggml (ggml_scale, ggml_row_size, etc.) (#1677)

sync : ggml (Metal fixes, new ops, tests) (#1633)

sync : ggml (new ops, new backend, etc) (#1602)

sync : ggml (ggml-alloc + linker + gguf fixes) (#1501)

sync : ggml (backend v2, k-quants, CUDA opts, Metal opts, etc.) (#1422)