talk-llama : fix build + sync latest llama.cpp

readme : improve Core ML model conversion guidance (#915)

coreml : support quantized model files

ggml : sync latest ggml

main : fix help for --no-timestamps arg (#908)

extra : update ggml sync script

whisper.objc : enable Core ML in example & fix segmentation fault (#910)

coreml : fix seg fault, double free (#919, #917, #899)

coreml : fix memory leak (#899)

cmake : fix define used for COREML_ALLOW_FALLBACK (#893)

talk-llama : only copy used KV cache in get / set state (#890)

readme : add instructions on converting to GGML + "--no-config" to wget (#874)

cmake : fix options disabling AVX and AVX2 flags (#885)

cmake : add options to disable CPU flags (#860)

ci : add cuBLAS build workflow and fix error causing lines in CMakeLists (#867)

readme : partial OpenCL GPU support via CLBlast (#863)

build : CLBlast support as in llama.cpp (#862)

ggml : fix 32-bit ARM build + quantization

ggml : sync ggml (clBLAST + tensor names)

talk-llama : fix session prompt load (#854)

whisper : add detect-language mode (#853)

talk-llama : add --session support (#845)

bench : improve benchmarks

whisper : add memory sizes for Q8_0 (close #846)

whisper.wasm : fix typo in readme (#832)

release : v1.4.1

whisper : fix quantize bug (#842)

ggml : fix UB (int << 31)

release : v1.4.0

examples : fix + refactor Levenshtein distance

whisper : add integer quantization support (#540)

whisper : add GPU support via cuBLAS (#834)

ggml : fix WASM build

ggml : fix 32-bit ARM NEON (#836)

ggml : use vzip instead of vuzp for consistency

ggml : fix WASM build

ggml : sync with ggml repo (warning fixes + asserts)

whisper : use correct seek_end when offset is used (#833)

tests : add "threads" to run-tests.sh

extra : add sync-ggml.sh script

ggml : sync latest ggml + llama.cpp updates (quantization)

whisper.android : add some tips (#816)

build : add WHISPER_COREML_ALLOW_FALLBACK to make / CMake (#812)

whisper : allow non-CoreML fallback when Core ML cannot be loaded (#812)

whisper : fix bug from previous commit

whisper : avoid designated initializers

minor : improve C++ and Python style (#768)

readme : add logo

main : escape quotes in csv output (#815)

stream : flush upon finishing inference (#811)