├── .gitmodules
├── models
├── .gitignore
├── requirements-openvino.txt
├── requirements-coreml.txt
├── for-tests-ggml-base.bin
├── for-tests-ggml-tiny.bin
├── for-tests-ggml-base.en.bin
├── for-tests-ggml-large.bin
├── for-tests-ggml-medium.bin
├── for-tests-ggml-small.bin
├── for-tests-ggml-tiny.en.bin
├── for-tests-ggml-medium.en.bin
└── for-tests-ggml-small.en.bin
├── bindings
├── go
│ ├── .gitignore
│ ├── samples
│ │ └── jfk.wav
│ ├── pkg
│ │ └── whisper
│ │ │ ├── doc.go
│ │ │ ├── util_test.go
│ │ │ └── consts.go
│ ├── doc.go
│ ├── go.mod
│ └── examples
│ │ ├── go-whisper
│ │ └── color.go
│ │ └── go-model-download
│ │ └── context.go
├── javascript
│ ├── .gitignore
│ ├── package.json
│ └── package-tmpl.json
├── ruby
│ ├── .gitignore
│ ├── tests
│ │ ├── jfk_reader
│ │ │ ├── extconf.rb
│ │ │ └── .gitignore
│ │ ├── test_error.rb
│ │ └── helper.rb
│ ├── ext
│ │ ├── .gitignore
│ │ ├── metal.mk
│ │ ├── cpu.mk
│ │ └── ruby_whisper.h
│ └── extsources.rb
├── java
│ ├── settings.gradle
│ ├── gradle
│ │ └── wrapper
│ │ │ ├── gradle-wrapper.jar
│ │ │ └── gradle-wrapper.properties
│ ├── src
│ │ ├── main
│ │ │ └── java
│ │ │ │ └── io
│ │ │ │ └── github
│ │ │ │ └── ggerganov
│ │ │ │ └── whispercpp
│ │ │ │ ├── ggml
│ │ │ │ ├── GgmlTensor.java
│ │ │ │ └── GgmlType.java
│ │ │ │ ├── model
│ │ │ │ ├── WhisperState.java
│ │ │ │ └── EModel.java
│ │ │ │ ├── params
│ │ │ │ ├── WhisperFilters.java
│ │ │ │ ├── WhisperSamplingStrategy.java
│ │ │ │ ├── WhisperHParams.java
│ │ │ │ ├── GreedyParams.java
│ │ │ │ ├── BeamSearchParams.java
│ │ │ │ └── CBool.java
│ │ │ │ └── callbacks
│ │ │ │ └── WhisperProgressCallback.java
│ │ └── test
│ │ │ └── java
│ │ │ └── io
│ │ │ └── github
│ │ │ └── ggerganov
│ │ │ └── whispercpp
│ │ │ └── WhisperJnaLibraryTest.java
│ └── gradle.properties
└── CMakeLists.txt
├── examples
├── whisper.android
│ ├── app
│ │ ├── .gitignore
│ │ ├── src
│ │ │ ├── main
│ │ │ │ ├── res
│ │ │ │ │ ├── values
│ │ │ │ │ │ ├── strings.xml
│ │ │ │ │ │ └── themes.xml
│ │ │ │ │ ├── mipmap-anydpi
│ │ │ │ │ │ └── ic_launcher.xml
│ │ │ │ │ └── xml
│ │ │ │ │ │ ├── backup_rules.xml
│ │ │ │ │ │ └── data_extraction_rules.xml
│ │ │ │ └── java
│ │ │ │ │ └── com
│ │ │ │ │ └── whispercppdemo
│ │ │ │ │ ├── ui
│ │ │ │ │ └── theme
│ │ │ │ │ │ └── Color.kt
│ │ │ │ │ └── MainActivity.kt
│ │ │ ├── test
│ │ │ │ └── java
│ │ │ │ │ └── com
│ │ │ │ │ └── whispercppdemo
│ │ │ │ │ └── ExampleUnitTest.kt
│ │ │ └── androidTest
│ │ │ │ └── java
│ │ │ │ └── com
│ │ │ │ └── whispercppdemo
│ │ │ │ └── ExampleInstrumentedTest.kt
│ │ └── proguard-rules.pro
│ ├── lib
│ │ ├── .gitignore
│ │ └── src
│ │ │ └── main
│ │ │ └── AndroidManifest.xml
│ ├── .idea
│ │ ├── .name
│ │ ├── .gitignore
│ │ ├── compiler.xml
│ │ ├── vcs.xml
│ │ ├── misc.xml
│ │ └── gradle.xml
│ ├── gradle
│ │ └── wrapper
│ │ │ ├── gradle-wrapper.jar
│ │ │ └── gradle-wrapper.properties
│ ├── .gitignore
│ ├── build.gradle
│ └── settings.gradle
├── talk-llama
│ ├── llama-quant.h
│ ├── .gitignore
│ ├── llama-cparams.cpp
│ ├── llama-memory.cpp
│ ├── speak.bat
│ ├── llama-io.cpp
│ ├── speak.ps1
│ ├── unicode-data.h
│ ├── llama-sampling.h
│ ├── llama-memory.h
│ └── llama-io.h
├── whisper.android.java
│ ├── app
│ │ ├── .gitignore
│ │ ├── src
│ │ │ ├── main
│ │ │ │ ├── res
│ │ │ │ │ ├── values
│ │ │ │ │ │ ├── strings.xml
│ │ │ │ │ │ ├── colors.xml
│ │ │ │ │ │ └── themes.xml
│ │ │ │ │ ├── mipmap-hdpi
│ │ │ │ │ │ ├── ic_launcher.png
│ │ │ │ │ │ └── ic_launcher_round.png
│ │ │ │ │ ├── mipmap-mdpi
│ │ │ │ │ │ ├── ic_launcher.png
│ │ │ │ │ │ └── ic_launcher_round.png
│ │ │ │ │ ├── mipmap-xhdpi
│ │ │ │ │ │ ├── ic_launcher.png
│ │ │ │ │ │ └── ic_launcher_round.png
│ │ │ │ │ ├── mipmap-xxhdpi
│ │ │ │ │ │ ├── ic_launcher.png
│ │ │ │ │ │ └── ic_launcher_round.png
│ │ │ │ │ ├── mipmap-xxxhdpi
│ │ │ │ │ │ ├── ic_launcher.png
│ │ │ │ │ │ └── ic_launcher_round.png
│ │ │ │ │ ├── mipmap-anydpi-v26
│ │ │ │ │ │ ├── ic_launcher.xml
│ │ │ │ │ │ └── ic_launcher_round.xml
│ │ │ │ │ └── values-night
│ │ │ │ │ │ └── themes.xml
│ │ │ │ ├── java
│ │ │ │ │ └── com
│ │ │ │ │ │ ├── litongjava
│ │ │ │ │ │ └── whisper
│ │ │ │ │ │ │ └── android
│ │ │ │ │ │ │ └── java
│ │ │ │ │ │ │ └── app
│ │ │ │ │ │ │ └── App.java
│ │ │ │ │ │ └── whispercpp
│ │ │ │ │ │ └── java
│ │ │ │ │ │ └── whisper
│ │ │ │ │ │ └── WhisperCpuConfig.java
│ │ │ │ └── AndroidManifest.xml
│ │ │ └── test
│ │ │ │ └── java
│ │ │ │ └── com
│ │ │ │ └── litongjava
│ │ │ │ └── whisper
│ │ │ │ └── android
│ │ │ │ └── java
│ │ │ │ └── ExampleUnitTest.java
│ │ └── proguard-rules.pro
│ ├── settings.gradle
│ ├── README_files
│ │ └── 1.jpg
│ ├── gradle
│ │ └── wrapper
│ │ │ ├── gradle-wrapper.jar
│ │ │ └── gradle-wrapper.properties
│ ├── .gitignore
│ └── build.gradle
├── addon.node
│ ├── .gitignore
│ ├── package.json
│ └── __test__
│ │ └── whisper.spec.js
├── whisper.swiftui
│ ├── .gitignore
│ ├── whisper.swiftui.demo
│ │ ├── Resources
│ │ │ ├── models
│ │ │ │ └── .gitignore
│ │ │ └── samples
│ │ │ │ └── .gitignore
│ │ ├── Supporting files
│ │ │ ├── Assets.xcassets
│ │ │ │ ├── Contents.json
│ │ │ │ └── AccentColor.colorset
│ │ │ │ │ └── Contents.json
│ │ │ ├── Preview Content
│ │ │ │ └── Preview Assets.xcassets
│ │ │ │ │ └── Contents.json
│ │ │ └── WhisperCppDemo.entitlements
│ │ ├── WhisperCppDemoApp.swift
│ │ ├── Utils
│ │ │ └── RiffWaveUtils.swift
│ │ └── Models
│ │ │ └── Model.swift
│ └── whisper.swiftui.xcodeproj
│ │ ├── .gitignore
│ │ └── project.xcworkspace
│ │ ├── .gitignore
│ │ └── xcshareddata
│ │ └── IDEWorkspaceChecks.plist
├── quantize
│ ├── README.md
│ └── CMakeLists.txt
├── command
│ ├── commands.txt
│ └── CMakeLists.txt
├── whisper.objc
│ ├── whisper.objc
│ │ ├── Assets.xcassets
│ │ │ ├── Contents.json
│ │ │ ├── AccentColor.colorset
│ │ │ │ └── Contents.json
│ │ │ └── AppIcon.appiconset
│ │ │ │ └── Contents.json
│ │ ├── AppDelegate.h
│ │ ├── SceneDelegate.h
│ │ ├── main.m
│ │ └── Info.plist
│ └── whisper.objc.xcodeproj
│ │ └── project.xcworkspace
│ │ ├── contents.xcworkspacedata
│ │ └── xcshareddata
│ │ └── IDEWorkspaceChecks.plist
├── wchess
│ ├── wchess.wasm
│ │ └── chessboardjs-1.0.0
│ │ │ ├── img
│ │ │ └── chesspieces
│ │ │ │ └── wikipedia
│ │ │ │ ├── bB.png
│ │ │ │ ├── bK.png
│ │ │ │ ├── bN.png
│ │ │ │ ├── bP.png
│ │ │ │ ├── bQ.png
│ │ │ │ ├── bR.png
│ │ │ │ ├── wB.png
│ │ │ │ ├── wK.png
│ │ │ │ ├── wN.png
│ │ │ │ ├── wP.png
│ │ │ │ ├── wQ.png
│ │ │ │ └── wR.png
│ │ │ ├── css
│ │ │ └── chessboard-1.0.0.min.css
│ │ │ └── js
│ │ │ └── chessboard-1.0.0
│ │ │ └── package.json
│ ├── wchess.cmd
│ │ └── CMakeLists.txt
│ ├── CMakeLists.txt
│ └── libwchess
│ │ └── CMakeLists.txt
├── deprecation-warning
│ ├── CMakeLists.txt
│ └── README.md
├── python
│ └── test_whisper_processor.py
├── bench
│ └── CMakeLists.txt
├── cli
│ └── CMakeLists.txt
├── sycl
│ ├── ls-sycl-device.cpp
│ ├── CMakeLists.txt
│ ├── run-whisper.sh
│ └── build.sh
├── lsp
│ └── CMakeLists.txt
├── stream
│ └── CMakeLists.txt
├── server
│ └── CMakeLists.txt
├── common-ggml.h
├── stream.wasm
│ └── README.md
├── console.h
├── bench.wasm
│ └── README.md
└── command.wasm
│ └── README.md
├── tests
├── .gitignore
├── test-c.c
└── en-2-ref.txt
├── samples
├── .gitignore
├── jfk.mp3
├── jfk.wav
└── README.md
├── scripts
├── sync-ggml.last
├── sha-all.sh
├── apple
│ └── validate-apps.sh
├── convert-all.sh
├── gen-authors.sh
├── sync-llama.sh
└── build-info.sh
├── ggml
├── .gitignore
├── src
│ ├── ggml-cuda
│ │ ├── argmax.cuh
│ │ ├── out-prod.cuh
│ │ ├── argsort.cuh
│ │ ├── fattn.cuh
│ │ ├── gla.cuh
│ │ ├── fattn-tile-f16.cuh
│ │ ├── fattn-tile-f32.cuh
│ │ ├── fattn-wmma-f16.cuh
│ │ ├── acc.cuh
│ │ ├── pad.cuh
│ │ ├── clamp.cuh
│ │ ├── scale.cuh
│ │ ├── arange.cuh
│ │ ├── concat.cuh
│ │ ├── im2col.cuh
│ │ ├── pool2d.cuh
│ │ ├── wkv6.cuhold
│ │ ├── upscale.cuh
│ │ ├── count-equal.cuh
│ │ ├── diagmask.cuh
│ │ ├── opt-step-adamw.cuh
│ │ ├── tsembd.cuh
│ │ ├── conv-transpose-1d.cuh
│ │ ├── template-instances
│ │ │ ├── mmq-instance-iq1_s.cu
│ │ │ ├── mmq-instance-iq2_s.cu
│ │ │ ├── mmq-instance-iq3_s.cu
│ │ │ ├── mmq-instance-q2_k.cu
│ │ │ ├── mmq-instance-q3_k.cu
│ │ │ ├── mmq-instance-q4_0.cu
│ │ │ ├── mmq-instance-q4_1.cu
│ │ │ ├── mmq-instance-q4_k.cu
│ │ │ ├── mmq-instance-q5_0.cu
│ │ │ ├── mmq-instance-q5_1.cu
│ │ │ ├── mmq-instance-q5_k.cu
│ │ │ ├── mmq-instance-q6_k.cu
│ │ │ ├── mmq-instance-q8_0.cu
│ │ │ ├── mmq-instance-iq2_xs.cu
│ │ │ ├── mmq-instance-iq2_xxs.cu
│ │ │ ├── mmq-instance-iq3_xxs.cu
│ │ │ ├── mmq-instance-iq4_nl.cu
│ │ │ ├── mmq-instance-iq4_xs.cu
│ │ │ ├── fattn-vec-f16-instance-hs64-f16-f16.cu
│ │ │ ├── fattn-vec-f32-instance-hs64-f16-f16.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-f16-f16.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-f16-q4_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-f16-q4_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-f16-q5_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-f16-q5_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-f16-q8_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_0-f16.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_0-q4_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_0-q4_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_0-q5_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_0-q5_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_0-q8_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_1-f16.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_1-q4_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_1-q4_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_1-q5_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_1-q5_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q4_1-q8_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_0-f16.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_0-q4_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_0-q4_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_0-q5_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_0-q5_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_0-q8_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_1-f16.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_1-q4_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_1-q4_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_1-q5_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_1-q5_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q5_1-q8_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q8_0-f16.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q8_0-q4_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q8_0-q4_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q8_0-q5_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q8_0-q5_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs128-q8_0-q8_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs256-f16-f16.cu
│ │ │ ├── fattn-vec-f16-instance-hs64-f16-q4_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs64-f16-q4_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs64-f16-q5_0.cu
│ │ │ ├── fattn-vec-f16-instance-hs64-f16-q5_1.cu
│ │ │ ├── fattn-vec-f16-instance-hs64-f16-q8_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-f16-f16.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-f16-q4_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-f16-q4_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-f16-q5_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-f16-q5_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-f16-q8_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_0-f16.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_0-q4_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_0-q4_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_0-q5_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_0-q5_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_0-q8_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_1-f16.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_1-q4_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_1-q4_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_1-q5_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_1-q5_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q4_1-q8_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_0-f16.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_0-q4_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_0-q4_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_0-q5_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_0-q5_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_0-q8_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_1-f16.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_1-q4_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_1-q4_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_1-q5_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_1-q5_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q5_1-q8_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q8_0-f16.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q8_0-q4_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q8_0-q4_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q8_0-q5_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q8_0-q5_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs128-q8_0-q8_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs256-f16-f16.cu
│ │ │ ├── fattn-vec-f32-instance-hs64-f16-q4_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs64-f16-q4_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs64-f16-q5_0.cu
│ │ │ ├── fattn-vec-f32-instance-hs64-f16-q5_1.cu
│ │ │ ├── fattn-vec-f32-instance-hs64-f16-q8_0.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_1-ncols2_8.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_2-ncols2_4.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_2-ncols2_8.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_4-ncols2_2.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_4-ncols2_4.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_4-ncols2_8.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_8-ncols2_1.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_8-ncols2_2.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_8-ncols2_4.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_8-ncols2_8.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_16-ncols2_1.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_16-ncols2_2.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_16-ncols2_4.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_32-ncols2_1.cu
│ │ │ ├── fattn-mma-f16-instance-ncols1_32-ncols2_2.cu
│ │ │ └── fattn-mma-f16-instance-ncols1_64-ncols2_1.cu
│ │ ├── sum.cuh
│ │ ├── sumrows.cuh
│ │ ├── rope.cuh
│ │ ├── wkv.cuh
│ │ ├── softmax.cuh
│ │ ├── cross-entropy-loss.cuh
│ │ ├── getrows.cuh
│ │ ├── cpy.cuh
│ │ ├── convert.cuh
│ │ ├── norm.cuh
│ │ ├── mmvq.cuh
│ │ ├── binbcast.cuh
│ │ ├── vendors
│ │ │ └── cuda.h
│ │ └── mmv.cuh
│ ├── ggml-vulkan
│ │ ├── vulkan-shaders
│ │ │ ├── test_coopmat2_support.comp
│ │ │ ├── test_coopmat_support.comp
│ │ │ ├── generic_head.comp
│ │ │ ├── CMakeLists.txt
│ │ │ ├── dequant_head.comp
│ │ │ ├── cos.comp
│ │ │ ├── sin.comp
│ │ │ ├── square.comp
│ │ │ ├── clamp.comp
│ │ │ ├── dequant_f32.comp
│ │ │ ├── copy.comp
│ │ │ ├── relu.comp
│ │ │ ├── sigmoid.comp
│ │ │ ├── tanh.comp
│ │ │ ├── silu.comp
│ │ │ ├── leaky_relu.comp
│ │ │ ├── scale.comp
│ │ │ ├── gelu_quick.comp
│ │ │ ├── gelu.comp
│ │ │ ├── silu_back.comp
│ │ │ ├── div.comp
│ │ │ ├── mul.comp
│ │ │ ├── repeat.comp
│ │ │ └── count_equal.comp
│ │ └── cmake
│ │ │ └── host-toolchain.cmake.in
│ ├── ggml-cpu
│ │ ├── ggml-cpu-hbm.h
│ │ ├── ggml-cpu-aarch64.h
│ │ ├── amx
│ │ │ ├── amx.h
│ │ │ └── mmq.h
│ │ ├── llamafile
│ │ │ └── sgemm.h
│ │ └── kleidiai
│ │ │ └── kleidiai.h
│ ├── ggml-sycl
│ │ ├── gla.hpp
│ │ ├── wkv6.hpp
│ │ ├── outprod.hpp
│ │ ├── wkv.hpp
│ │ ├── cpy.hpp
│ │ ├── sycl_hw.cpp
│ │ ├── sycl_hw.hpp
│ │ ├── concat.hpp
│ │ ├── conv.hpp
│ │ ├── softmax.hpp
│ │ ├── tsembd.hpp
│ │ ├── rope.hpp
│ │ ├── getrows.hpp
│ │ ├── im2col.hpp
│ │ ├── mmvq.hpp
│ │ ├── dmmv.hpp
│ │ ├── convert.hpp
│ │ └── backend.hpp
│ ├── ggml-rpc
│ │ └── CMakeLists.txt
│ ├── ggml-threading.h
│ ├── ggml-threading.cpp
│ ├── ggml-amx
│ │ └── mmq.h
│ ├── ggml-kompute
│ │ └── kompute-shaders
│ │ │ ├── op_scale.comp
│ │ │ ├── op_relu.comp
│ │ │ ├── op_scale_8.comp
│ │ │ ├── op_silu.comp
│ │ │ ├── op_getrows.comp
│ │ │ ├── op_gelu.comp
│ │ │ ├── op_mul_mv_q_n_pre.comp
│ │ │ ├── op_addrow.comp
│ │ │ ├── op_diagmask.comp
│ │ │ ├── op_getrows_f32.comp
│ │ │ └── op_getrows_f16.comp
│ ├── ggml-opencl
│ │ └── kernels
│ │ │ └── embed_kernel.py
│ └── ggml-cann
│ │ └── kernels
│ │ └── ascendc_kernels.h
├── include
│ ├── ggml-blas.h
│ ├── ggml-opencl.h
│ └── ggml-rpc.h
└── cmake
│ └── GitVars.cmake
├── README_sycl.md
├── SDL2
├── lib
│ ├── x64
│ │ ├── SDL2.dll
│ │ ├── SDL2.lib
│ │ ├── SDL2main.lib
│ │ └── SDL2test.lib
│ └── x86
│ │ ├── SDL2.dll
│ │ ├── SDL2.lib
│ │ ├── SDL2main.lib
│ │ └── SDL2test.lib
├── docs
│ ├── README-hg.md
│ ├── README-platforms.md
│ ├── README-wince.md
│ ├── README-git.md
│ └── README-pandora.md
├── include
│ ├── SDL_revision.h
│ └── SDL_opengles2_gl2platform.h
├── README-SDL.txt
├── BUGS.txt
└── README.txt
├── cmake
├── whisper.pc.in
├── DefaultTargetOptions.cmake
└── git-vars.cmake
├── grammars
└── colors.gbnf
├── .github
└── workflows
│ └── bindings-go.yml
├── .devops
├── main.Dockerfile
└── cublas.Dockerfile
└── src
└── coreml
└── whisper-encoder.h
/.gitmodules:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/models/.gitignore:
--------------------------------------------------------------------------------
1 | *.bin
2 |
--------------------------------------------------------------------------------
/bindings/go/.gitignore:
--------------------------------------------------------------------------------
1 | build
2 | models
3 |
--------------------------------------------------------------------------------
/bindings/javascript/.gitignore:
--------------------------------------------------------------------------------
1 | publish.log
2 |
--------------------------------------------------------------------------------
/examples/whisper.android/app/.gitignore:
--------------------------------------------------------------------------------
1 | /build
--------------------------------------------------------------------------------
/examples/whisper.android/lib/.gitignore:
--------------------------------------------------------------------------------
1 | /build
--------------------------------------------------------------------------------
/examples/talk-llama/llama-quant.h:
--------------------------------------------------------------------------------
1 | #pragma once
2 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/.gitignore:
--------------------------------------------------------------------------------
1 | /build
--------------------------------------------------------------------------------
/examples/whisper.android/.idea/.name:
--------------------------------------------------------------------------------
1 | WhisperCppDemo
--------------------------------------------------------------------------------
/tests/.gitignore:
--------------------------------------------------------------------------------
1 | *.wav
2 | *.ogg
3 | *.wav.txt
4 |
--------------------------------------------------------------------------------
/samples/.gitignore:
--------------------------------------------------------------------------------
1 | *
2 | !jfk.wave
3 | !jfk.mp3
4 |
5 |
--------------------------------------------------------------------------------
/bindings/ruby/.gitignore:
--------------------------------------------------------------------------------
1 | LICENSE
2 | pkg/
3 | lib/whisper.*
4 |
--------------------------------------------------------------------------------
/examples/talk-llama/.gitignore:
--------------------------------------------------------------------------------
1 | audio.mp3
2 | to_speak.txt
3 |
--------------------------------------------------------------------------------
/bindings/java/settings.gradle:
--------------------------------------------------------------------------------
1 | rootProject.name = "whispercpp"
2 |
--------------------------------------------------------------------------------
/examples/addon.node/.gitignore:
--------------------------------------------------------------------------------
1 | .idea
2 | node_modules
3 | build
4 |
--------------------------------------------------------------------------------
/examples/talk-llama/llama-cparams.cpp:
--------------------------------------------------------------------------------
1 | #include "llama-cparams.h"
2 |
--------------------------------------------------------------------------------
/examples/talk-llama/llama-memory.cpp:
--------------------------------------------------------------------------------
1 | #include "llama-memory.h"
2 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/.gitignore:
--------------------------------------------------------------------------------
1 | xcuserdata
2 | xcshareddata
3 |
--------------------------------------------------------------------------------
/scripts/sync-ggml.last:
--------------------------------------------------------------------------------
1 | 7b08f4cd9e32781e769a52cf0dcaeb2c556632c3
2 |
--------------------------------------------------------------------------------
/tests/test-c.c:
--------------------------------------------------------------------------------
1 | #include "whisper.h"
2 |
3 | int main(void) {}
4 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/Resources/models/.gitignore:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/Resources/samples/.gitignore:
--------------------------------------------------------------------------------
1 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.xcodeproj/.gitignore:
--------------------------------------------------------------------------------
1 | xcuserdata/
2 |
--------------------------------------------------------------------------------
/models/requirements-openvino.txt:
--------------------------------------------------------------------------------
1 | openvino-dev[pytorch,onnx]
2 | openai-whisper
--------------------------------------------------------------------------------
/ggml/.gitignore:
--------------------------------------------------------------------------------
1 | src/ggml-vulkan-shaders.hpp
2 | src/ggml-vulkan-shaders.cpp
3 |
--------------------------------------------------------------------------------
/README_sycl.md:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/README_sycl.md
--------------------------------------------------------------------------------
/samples/jfk.mp3:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/samples/jfk.mp3
--------------------------------------------------------------------------------
/samples/jfk.wav:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/samples/jfk.wav
--------------------------------------------------------------------------------
/models/requirements-coreml.txt:
--------------------------------------------------------------------------------
1 | torch
2 | coremltools
3 | openai-whisper
4 | ane_transformers
5 |
--------------------------------------------------------------------------------
/SDL2/lib/x64/SDL2.dll:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/SDL2/lib/x64/SDL2.dll
--------------------------------------------------------------------------------
/SDL2/lib/x64/SDL2.lib:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/SDL2/lib/x64/SDL2.lib
--------------------------------------------------------------------------------
/SDL2/lib/x86/SDL2.dll:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/SDL2/lib/x86/SDL2.dll
--------------------------------------------------------------------------------
/SDL2/lib/x86/SDL2.lib:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/SDL2/lib/x86/SDL2.lib
--------------------------------------------------------------------------------
/bindings/ruby/tests/jfk_reader/extconf.rb:
--------------------------------------------------------------------------------
1 | require "mkmf"
2 |
3 | create_makefile("jfk_reader")
4 |
--------------------------------------------------------------------------------
/examples/whisper.android/.idea/.gitignore:
--------------------------------------------------------------------------------
1 | # Default ignored files
2 | /shelf/
3 | /workspace.xml
4 |
--------------------------------------------------------------------------------
/SDL2/lib/x64/SDL2main.lib:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/SDL2/lib/x64/SDL2main.lib
--------------------------------------------------------------------------------
/SDL2/lib/x64/SDL2test.lib:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/SDL2/lib/x64/SDL2test.lib
--------------------------------------------------------------------------------
/SDL2/lib/x86/SDL2main.lib:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/SDL2/lib/x86/SDL2main.lib
--------------------------------------------------------------------------------
/SDL2/lib/x86/SDL2test.lib:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/SDL2/lib/x86/SDL2test.lib
--------------------------------------------------------------------------------
/examples/whisper.android.java/settings.gradle:
--------------------------------------------------------------------------------
1 | include ':app'
2 | rootProject.name = "whisper.android.java"
--------------------------------------------------------------------------------
/bindings/go/samples/jfk.wav:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/bindings/go/samples/jfk.wav
--------------------------------------------------------------------------------
/examples/talk-llama/speak.bat:
--------------------------------------------------------------------------------
1 | @powershell -ExecutionPolicy Bypass -F examples\talk-llama\speak.ps1 %1 %2
2 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.xcodeproj/project.xcworkspace/.gitignore:
--------------------------------------------------------------------------------
1 | contents.xcworkspacedata
2 |
--------------------------------------------------------------------------------
/examples/quantize/README.md:
--------------------------------------------------------------------------------
1 | # quantize
2 |
3 | Tool for integer quantization of Whisper `ggml` model files
4 |
--------------------------------------------------------------------------------
/models/for-tests-ggml-base.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-base.bin
--------------------------------------------------------------------------------
/models/for-tests-ggml-tiny.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-tiny.bin
--------------------------------------------------------------------------------
/models/for-tests-ggml-base.en.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-base.en.bin
--------------------------------------------------------------------------------
/models/for-tests-ggml-large.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-large.bin
--------------------------------------------------------------------------------
/models/for-tests-ggml-medium.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-medium.bin
--------------------------------------------------------------------------------
/models/for-tests-ggml-small.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-small.bin
--------------------------------------------------------------------------------
/models/for-tests-ggml-tiny.en.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-tiny.en.bin
--------------------------------------------------------------------------------
/SDL2/docs/README-hg.md:
--------------------------------------------------------------------------------
1 | We are no longer hosted in Mercurial. Please see README-git.md for details.
2 |
3 | Thanks!
4 |
5 |
--------------------------------------------------------------------------------
/models/for-tests-ggml-medium.en.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-medium.en.bin
--------------------------------------------------------------------------------
/models/for-tests-ggml-small.en.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/models/for-tests-ggml-small.en.bin
--------------------------------------------------------------------------------
/bindings/go/pkg/whisper/doc.go:
--------------------------------------------------------------------------------
1 | /*
2 | This is the higher-level speech-to-text whisper.cpp API for go
3 | */
4 | package whisper
5 |
--------------------------------------------------------------------------------
/bindings/ruby/tests/jfk_reader/.gitignore:
--------------------------------------------------------------------------------
1 | Makefile
2 | jfk_reader.o
3 | jfk_reader.so
4 | jfk_reader.bundle
5 | jfk_reader.dll
6 |
--------------------------------------------------------------------------------
/examples/command/commands.txt:
--------------------------------------------------------------------------------
1 | enable
2 | disable
3 | cat
4 | dog
5 | apple
6 | red
7 | blue
8 | green
9 | lightblue
10 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/argmax.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_argmax(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/out-prod.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_out_prod(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
--------------------------------------------------------------------------------
/examples/whisper.android/app/src/main/res/values/strings.xml:
--------------------------------------------------------------------------------
1 |
2 | WhisperCppDemo
3 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/argsort.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_op_argsort(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/fattn.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_flash_attn_ext(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
--------------------------------------------------------------------------------
/bindings/java/gradle/wrapper/gradle-wrapper.jar:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/bindings/java/gradle/wrapper/gradle-wrapper.jar
--------------------------------------------------------------------------------
/examples/whisper.android.java/README_files/1.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/README_files/1.jpg
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/values/strings.xml:
--------------------------------------------------------------------------------
1 |
2 | whisper.android.java
3 |
--------------------------------------------------------------------------------
/examples/whisper.objc/whisper.objc/Assets.xcassets/Contents.json:
--------------------------------------------------------------------------------
1 | {
2 | "info" : {
3 | "author" : "xcode",
4 | "version" : 1
5 | }
6 | }
7 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/gla.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_op_gated_linear_attn(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/fattn-tile-f16.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_flash_attn_ext_tile_f16(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/fattn-tile-f32.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_flash_attn_ext_tile_f32(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/fattn-wmma-f16.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_flash_attn_ext_wmma_f16(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
--------------------------------------------------------------------------------
/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat2_support.comp:
--------------------------------------------------------------------------------
1 | #version 460
2 |
3 | #extension GL_NV_cooperative_matrix2 : require
4 |
5 | void main()
6 | {
7 | }
8 |
--------------------------------------------------------------------------------
/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat_support.comp:
--------------------------------------------------------------------------------
1 | #version 460
2 |
3 | #extension GL_KHR_cooperative_matrix : require
4 |
5 | void main()
6 | {
7 | }
8 |
--------------------------------------------------------------------------------
/bindings/java/src/main/java/io/github/ggerganov/whispercpp/ggml/GgmlTensor.java:
--------------------------------------------------------------------------------
1 | package io.github.ggerganov.whispercpp.ggml;
2 |
3 | public class GgmlTensor {
4 | }
5 |
--------------------------------------------------------------------------------
/examples/whisper.android/gradle/wrapper/gradle-wrapper.jar:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android/gradle/wrapper/gradle-wrapper.jar
--------------------------------------------------------------------------------
/bindings/go/pkg/whisper/util_test.go:
--------------------------------------------------------------------------------
1 | package whisper_test
2 |
3 | const (
4 | ModelPath = "../../models/ggml-small.en.bin"
5 | SamplePath = "../../samples/jfk.wav"
6 | )
7 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/acc.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_ACC_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_acc(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/pad.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_PAD_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_pad(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/bindings/go/doc.go:
--------------------------------------------------------------------------------
1 | /*
2 | github.com/ggerganov/whisper.cpp/bindings/go
3 | provides a speech-to-text service bindings for the Go programming language.
4 | */
5 | package whisper
6 |
--------------------------------------------------------------------------------
/bindings/java/src/main/java/io/github/ggerganov/whispercpp/model/WhisperState.java:
--------------------------------------------------------------------------------
1 | package io.github.ggerganov.whispercpp.model;
2 |
3 | public class WhisperState {
4 | }
5 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/gradle/wrapper/gradle-wrapper.jar:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/gradle/wrapper/gradle-wrapper.jar
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/Assets.xcassets/Contents.json:
--------------------------------------------------------------------------------
1 | {
2 | "info" : {
3 | "author" : "xcode",
4 | "version" : 1
5 | }
6 | }
7 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/clamp.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_CLAMP_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_clamp(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/scale.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_SCALE_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_scale(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/scripts/sha-all.sh:
--------------------------------------------------------------------------------
1 | #!/bin/bash
2 |
3 | # Compute the SHA1 of all model files in ./models/ggml-*.bin
4 |
5 | for f in ./models/ggml-*.bin; do
6 | shasum "$f" -a 1
7 | done
8 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/arange.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_ARANGE_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_arange(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/concat.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_CONCAT_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_concat(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/im2col.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_IM2COL_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_im2col(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/pool2d.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_POOL2D_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_pool2d(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/wkv6.cuhold:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_WKV_BLOCK_SIZE 64
4 |
5 | void ggml_cuda_op_rwkv_wkv6(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/examples/whisper.android/lib/src/main/AndroidManifest.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/upscale.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_UPSCALE_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_upscale(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/count-equal.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_COUNT_EQUAL_CHUNK_SIZE 128
4 |
5 | void ggml_cuda_count_equal(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/scripts/apple/validate-apps.sh:
--------------------------------------------------------------------------------
1 | #!/bin/bash
2 | ./scripts/apple/validate-ios.sh
3 | ./scripts/apple/validate-macos.sh
4 | ./scripts/apple/validate-visionos.sh
5 | ./scripts/apple/validate-tvos.sh
6 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/Preview Content/Preview Assets.xcassets/Contents.json:
--------------------------------------------------------------------------------
1 | {
2 | "info" : {
3 | "author" : "xcode",
4 | "version" : 1
5 | }
6 | }
7 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/diagmask.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_DIAG_MASK_INF_BLOCK_SIZE 32
4 |
5 | void ggml_cuda_op_diag_mask_inf(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-hdpi/ic_launcher.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-hdpi/ic_launcher.png
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-mdpi/ic_launcher.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-mdpi/ic_launcher.png
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-xhdpi/ic_launcher.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-xhdpi/ic_launcher.png
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/opt-step-adamw.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_OPT_STEP_ADAMW_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_opt_step_adamw(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/bindings/ruby/ext/.gitignore:
--------------------------------------------------------------------------------
1 | Makefile
2 | whisper.so
3 | whisper.bundle
4 | whisper.dll
5 | scripts/get-flags.mk
6 | *.o
7 | /*/**/*.c
8 | /*/**/*.cpp
9 | /*/**/*.h
10 | /*/**/*.m
11 | /*/**/*.metal
12 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-xxhdpi/ic_launcher.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-xxhdpi/ic_launcher.png
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-xxxhdpi/ic_launcher.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-xxxhdpi/ic_launcher.png
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/tsembd.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_TIMESTEP_EMBEDDING_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_timestep_embedding(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bB.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bB.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bK.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bK.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bN.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bN.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bP.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bP.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bQ.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bQ.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bR.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/bR.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wB.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wB.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wK.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wK.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wN.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wN.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wP.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wP.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wQ.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wQ.png
--------------------------------------------------------------------------------
/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wR.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/wchess/wchess.wasm/chessboardjs-1.0.0/img/chesspieces/wikipedia/wR.png
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-hdpi/ic_launcher_round.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-hdpi/ic_launcher_round.png
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-mdpi/ic_launcher_round.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-mdpi/ic_launcher_round.png
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/conv-transpose-1d.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_CONV_TRANPOSE_1D_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_conv_transpose_1d(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-iq1_s.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_IQ1_S);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_s.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_IQ2_S);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-iq3_s.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_IQ3_S);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q2_k.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q2_K);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q3_k.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q3_K);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_k.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q4_K);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_k.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q5_K);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q6_k.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q6_K);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-xhdpi/ic_launcher_round.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-xhdpi/ic_launcher_round.png
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-xxhdpi/ic_launcher_round.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-xxhdpi/ic_launcher_round.png
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-xxxhdpi/ic_launcher_round.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/Mozer/talk-llama-fast/HEAD/examples/whisper.android.java/app/src/main/res/mipmap-xxxhdpi/ic_launcher_round.png
--------------------------------------------------------------------------------
/ggml/src/ggml-cpu/ggml-cpu-hbm.h:
--------------------------------------------------------------------------------
1 | #pragma once
2 |
3 | #include "ggml-backend.h"
4 | #include "ggml.h"
5 |
6 | // GGML CPU internal header
7 |
8 | ggml_backend_buffer_type_t ggml_backend_cpu_hbm_buffer_type(void);
9 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_xs.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_IQ2_XS);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_xxs.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_IQ2_XXS);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-iq3_xxs.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_IQ3_XXS);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-iq4_nl.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_IQ4_NL);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/mmq-instance-iq4_xs.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../mmq.cuh"
4 |
5 | DECL_MMQ_CASE(GGML_TYPE_IQ4_XS);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cpu/ggml-cpu-aarch64.h:
--------------------------------------------------------------------------------
1 | #pragma once
2 |
3 | #include "ggml-cpu-traits.h"
4 | #include "ggml.h"
5 |
6 | // GGML internal header
7 |
8 | ggml_backend_buffer_type_t ggml_backend_cpu_aarch64_buffer_type(void);
9 |
--------------------------------------------------------------------------------
/examples/quantize/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | set(TARGET quantize)
2 | add_executable(${TARGET} quantize.cpp)
3 |
4 | include(DefaultTargetOptions)
5 |
6 | target_link_libraries(${TARGET} PRIVATE common whisper ${CMAKE_THREAD_LIBS_INIT})
7 |
--------------------------------------------------------------------------------
/examples/whisper.android/.idea/compiler.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
--------------------------------------------------------------------------------
/examples/whisper.android/app/src/main/res/values/themes.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
--------------------------------------------------------------------------------
/SDL2/docs/README-platforms.md:
--------------------------------------------------------------------------------
1 | Platforms
2 | =========
3 |
4 | We maintain the list of supported platforms on our wiki now, and how to
5 | build and install SDL for those platforms:
6 |
7 | https://wiki.libsdl.org/Installation
8 |
9 |
--------------------------------------------------------------------------------
/bindings/ruby/ext/metal.mk:
--------------------------------------------------------------------------------
1 | ggml/src/ggml-metal/ggml-metal.o: \
2 | ggml/src/ggml-metal/ggml-metal.m \
3 | ggml/src/ggml-metal/ggml-metal-impl.h \
4 | ggml/include/ggml-metal.h \
5 | ggml/include/ggml.h
6 | $(CC) $(CFLAGS) -c $< -o $@
7 |
--------------------------------------------------------------------------------
/examples/whisper.android/.idea/vcs.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
--------------------------------------------------------------------------------
/examples/deprecation-warning/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | add_executable(main ./deprecation-warning.cpp)
2 | add_executable(bench ./deprecation-warning.cpp)
3 | add_executable(stream ./deprecation-warning.cpp)
4 | add_executable(command ./deprecation-warning.cpp)
5 |
--------------------------------------------------------------------------------
/examples/whisper.objc/whisper.objc.xcodeproj/project.xcworkspace/contents.xcworkspacedata:
--------------------------------------------------------------------------------
1 |
2 |
4 |
6 |
7 |
8 |
--------------------------------------------------------------------------------
/ggml/src/ggml-sycl/gla.hpp:
--------------------------------------------------------------------------------
1 | #ifndef GGML_SYCL_GLA_HPP
2 | #define GGML_SYCL_GLA_HPP
3 |
4 | #include "common.hpp"
5 |
6 | void ggml_sycl_op_gated_linear_attn(ggml_backend_sycl_context & ctx, ggml_tensor * dst);
7 |
8 | #endif // GGML_SYCL_GLA_HPP
9 |
--------------------------------------------------------------------------------
/examples/python/test_whisper_processor.py:
--------------------------------------------------------------------------------
1 | import whisper_processor
2 |
3 | try:
4 | result = whisper_processor.process_audio("./audio/wake_word_detected16k.wav", "base.en")
5 | print(result)
6 | except Exception as e:
7 | print(f"Error: {e}")
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/sum.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void sum_f32_cuda(ggml_cuda_pool & pool, const float * x, float * dst, const int64_t ne, cudaStream_t stream);
4 |
5 | void ggml_cuda_op_sum(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cpu/amx/amx.h:
--------------------------------------------------------------------------------
1 | #include "ggml-backend.h"
2 | #include "ggml-cpu-impl.h"
3 |
4 | // GGML internal header
5 |
6 | #if defined(__AMX_INT8__) && defined(__AVX512VNNI__)
7 | ggml_backend_buffer_type_t ggml_backend_amx_buffer_type(void);
8 | #endif
9 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/sumrows.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void sum_rows_f32_cuda(const float * x, float * dst, const int ncols, const int nrows, cudaStream_t stream);
4 |
5 | void ggml_cuda_op_sum_rows(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-sycl/wkv6.hpp:
--------------------------------------------------------------------------------
1 | #ifndef GGML_SYCL_WKV6_HPP
2 | #define GGML_SYCL_WKV6_HPP
3 |
4 | #include "common.hpp"
5 |
6 | void ggml_sycl_op_rwkv_wkv6(ggml_backend_sycl_context & ctx, ggml_tensor * dst);
7 |
8 |
9 | #endif // GGML_SYCL_WKV6_HPP
10 |
--------------------------------------------------------------------------------
/ggml/src/ggml-vulkan/vulkan-shaders/generic_head.comp:
--------------------------------------------------------------------------------
1 | #extension GL_EXT_shader_16bit_storage : require
2 |
3 | layout (push_constant) uniform parameter
4 | {
5 | uint KX;
6 | uint KY;
7 | float param1;
8 | float param2;
9 | } p;
10 |
--------------------------------------------------------------------------------
/examples/bench/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | set(TARGET whisper-bench)
2 | add_executable(${TARGET} bench.cpp)
3 |
4 | include(DefaultTargetOptions)
5 |
6 | target_link_libraries(${TARGET} PRIVATE whisper ${CMAKE_THREAD_LIBS_INIT})
7 |
8 | install(TARGETS ${TARGET} RUNTIME)
9 |
--------------------------------------------------------------------------------
/examples/whisper.objc/whisper.objc/Assets.xcassets/AccentColor.colorset/Contents.json:
--------------------------------------------------------------------------------
1 | {
2 | "colors" : [
3 | {
4 | "idiom" : "universal"
5 | }
6 | ],
7 | "info" : {
8 | "author" : "xcode",
9 | "version" : 1
10 | }
11 | }
12 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/WhisperCppDemoApp.swift:
--------------------------------------------------------------------------------
1 | import SwiftUI
2 |
3 | @main
4 | struct WhisperCppDemoApp: App {
5 | var body: some Scene {
6 | WindowGroup {
7 | ContentView()
8 | }
9 | }
10 | }
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/rope.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_ROPE_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_op_rope(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
7 | void ggml_cuda_op_rope_back(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
8 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(64, GGML_TYPE_F16, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(64, GGML_TYPE_F16, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/wkv.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_WKV_BLOCK_SIZE 64
4 |
5 | void ggml_cuda_op_rwkv_wkv6(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
7 | void ggml_cuda_op_rwkv_wkv7(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
8 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_F16, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs256-f16-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(256, GGML_TYPE_F16, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f16.cuh"
4 |
5 | DECL_FATTN_VEC_F16_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_F16, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_F16, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_0, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q4_1, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_0, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q5_1, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs256-f16-f16.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(256, GGML_TYPE_F16, GGML_TYPE_F16);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q4_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q4_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q5_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q5_1);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q8_0.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-vec-f32.cuh"
4 |
5 | DECL_FATTN_VEC_F32_CASE(64, GGML_TYPE_F16, GGML_TYPE_Q8_0);
6 |
--------------------------------------------------------------------------------
/ggml/src/ggml-sycl/outprod.hpp:
--------------------------------------------------------------------------------
1 | #ifndef GGML_SYCL_OUTPROD_HPP
2 | #define GGML_SYCL_OUTPROD_HPP
3 |
4 | #include "common.hpp"
5 |
6 | void ggml_sycl_op_out_prod(ggml_backend_sycl_context& ctx, ggml_tensor* dst);
7 |
8 |
9 | #endif // GGML_SYCL_OUTPROD_HPP
10 |
11 |
--------------------------------------------------------------------------------
/SDL2/docs/README-wince.md:
--------------------------------------------------------------------------------
1 | WinCE
2 | =====
3 |
4 | Windows CE is no longer supported by SDL.
5 |
6 | We have left the CE support in SDL 1.2 for those that must have it, and we
7 | have support for Windows Phone 8 and WinRT in SDL2, as of SDL 2.0.3.
8 |
9 | --ryan.
10 |
11 |
--------------------------------------------------------------------------------
/examples/cli/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | set(TARGET whisper-cli)
2 | add_executable(${TARGET} cli.cpp)
3 |
4 | include(DefaultTargetOptions)
5 |
6 | target_link_libraries(${TARGET} PRIVATE common whisper ${FFMPEG_LIBRARIES} ${CMAKE_THREAD_LIBS_INIT})
7 |
8 | install(TARGETS ${TARGET} RUNTIME)
9 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/softmax.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_SOFT_MAX_BLOCK_SIZE 1024
4 |
5 | void ggml_cuda_op_soft_max(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
7 | void ggml_cuda_op_soft_max_back(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
8 |
--------------------------------------------------------------------------------
/ggml/src/ggml-rpc/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | message(STATUS "Using RPC backend")
2 |
3 | ggml_add_backend_library(ggml-rpc
4 | ggml-rpc.cpp
5 | )
6 |
7 | if (WIN32)
8 | target_link_libraries(ggml-rpc PRIVATE ws2_32)
9 | endif()
10 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/Assets.xcassets/AccentColor.colorset/Contents.json:
--------------------------------------------------------------------------------
1 | {
2 | "colors" : [
3 | {
4 | "idiom" : "universal"
5 | }
6 | ],
7 | "info" : {
8 | "author" : "xcode",
9 | "version" : 1
10 | }
11 | }
12 |
--------------------------------------------------------------------------------
/SDL2/include/SDL_revision.h:
--------------------------------------------------------------------------------
1 | /* Generated by updaterev.sh, do not edit */
2 | #ifdef SDL_VENDOR_INFO
3 | #define SDL_REVISION "SDL-release-2.28.5-0-g15ead9a40 (" SDL_VENDOR_INFO ")"
4 | #else
5 | #define SDL_REVISION "SDL-release-2.28.5-0-g15ead9a40"
6 | #endif
7 | #define SDL_REVISION_NUMBER 0
8 |
--------------------------------------------------------------------------------
/examples/wchess/wchess.cmd/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | if (WHISPER_SDL2)
2 | set(TARGET wchess)
3 | add_executable(${TARGET} wchess.cmd.cpp)
4 |
5 | include(DefaultTargetOptions)
6 |
7 | target_link_libraries(${TARGET} PRIVATE wchess-core common-sdl ${CMAKE_THREAD_LIBS_INIT})
8 | endif ()
9 |
--------------------------------------------------------------------------------
/ggml/src/ggml-threading.h:
--------------------------------------------------------------------------------
1 | #pragma once
2 |
3 | #include "ggml.h"
4 |
5 | #ifdef __cplusplus
6 | extern "C" {
7 | #endif
8 |
9 | GGML_API void ggml_critical_section_start(void);
10 | GGML_API void ggml_critical_section_end(void);
11 |
12 | #ifdef __cplusplus
13 | }
14 | #endif
15 |
--------------------------------------------------------------------------------
/bindings/java/gradle/wrapper/gradle-wrapper.properties:
--------------------------------------------------------------------------------
1 | distributionBase=GRADLE_USER_HOME
2 | distributionPath=wrapper/dists
3 | distributionUrl=https\://services.gradle.org/distributions/gradle-8.1-bin.zip
4 | networkTimeout=10000
5 | zipStoreBase=GRADLE_USER_HOME
6 | zipStorePath=wrapper/dists
7 |
--------------------------------------------------------------------------------
/bindings/java/src/main/java/io/github/ggerganov/whispercpp/params/WhisperFilters.java:
--------------------------------------------------------------------------------
1 | package io.github.ggerganov.whispercpp.params;
2 |
3 | import java.util.List;
4 |
5 | public class WhisperFilters {
6 | int n_mel;
7 | int n_fft;
8 |
9 | List data;
10 | }
11 |
--------------------------------------------------------------------------------
/examples/sycl/ls-sycl-device.cpp:
--------------------------------------------------------------------------------
1 | /*MIT license
2 | Copyright (C) 2024 Intel Corporation
3 | SPDX-License-Identifier: MIT
4 | */
5 |
6 | #include "ggml-sycl.h"
7 |
8 | int main(int argc, char ** argv) {
9 | ggml_backend_sycl_print_sycl_devices();
10 | return 0;
11 | }
--------------------------------------------------------------------------------
/examples/whisper.objc/whisper.objc/AppDelegate.h:
--------------------------------------------------------------------------------
1 | //
2 | // AppDelegate.h
3 | // whisper.objc
4 | //
5 | // Created by Georgi Gerganov on 23.10.22.
6 | //
7 |
8 | #import
9 |
10 | @interface AppDelegate : UIResponder
11 |
12 |
13 | @end
14 |
15 |
--------------------------------------------------------------------------------
/bindings/java/src/main/java/io/github/ggerganov/whispercpp/model/EModel.java:
--------------------------------------------------------------------------------
1 | package io.github.ggerganov.whispercpp.model;
2 |
3 | public enum EModel {
4 | MODEL_UNKNOWN,
5 | MODEL_TINY,
6 | MODEL_BASE,
7 | MODEL_SMALL,
8 | MODEL_MEDIUM,
9 | MODEL_LARGE,
10 | }
11 |
--------------------------------------------------------------------------------
/examples/lsp/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | if (WHISPER_SDL2)
2 | # stream
3 | set(TARGET lsp)
4 | add_executable(${TARGET} lsp.cpp)
5 |
6 | include(DefaultTargetOptions)
7 |
8 | target_link_libraries(${TARGET} PRIVATE common json_cpp common-sdl whisper ${CMAKE_THREAD_LIBS_INIT})
9 | endif ()
10 |
--------------------------------------------------------------------------------
/examples/whisper.android/gradle/wrapper/gradle-wrapper.properties:
--------------------------------------------------------------------------------
1 | #Wed Dec 14 10:37:24 EST 2022
2 | distributionBase=GRADLE_USER_HOME
3 | distributionUrl=https\://services.gradle.org/distributions/gradle-8.2-bin.zip
4 | distributionPath=wrapper/dists
5 | zipStorePath=wrapper/dists
6 | zipStoreBase=GRADLE_USER_HOME
7 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/cross-entropy-loss.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_CROSS_ENTROPY_LOSS_BLOCK_SIZE 256
4 |
5 | void ggml_cuda_cross_entropy_loss(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
7 | void ggml_cuda_cross_entropy_loss_back(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
8 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/gradle/wrapper/gradle-wrapper.properties:
--------------------------------------------------------------------------------
1 | #Fri Oct 20 11:07:15 HST 2023
2 | distributionBase=GRADLE_USER_HOME
3 | distributionPath=wrapper/dists
4 | zipStoreBase=GRADLE_USER_HOME
5 | zipStorePath=wrapper/dists
6 | distributionUrl=https\://services.gradle.org/distributions/gradle-6.5-all.zip
7 |
--------------------------------------------------------------------------------
/examples/whisper.objc/whisper.objc/Assets.xcassets/AppIcon.appiconset/Contents.json:
--------------------------------------------------------------------------------
1 | {
2 | "images" : [
3 | {
4 | "idiom" : "universal",
5 | "platform" : "ios",
6 | "size" : "1024x1024"
7 | }
8 | ],
9 | "info" : {
10 | "author" : "xcode",
11 | "version" : 1
12 | }
13 | }
14 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/getrows.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_GET_ROWS_BLOCK_SIZE 256
4 | #define CUDA_GET_ROWS_BACK_BLOCK_SIZE 256
5 |
6 | void ggml_cuda_op_get_rows(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
7 |
8 | void ggml_cuda_op_get_rows_back(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
9 |
--------------------------------------------------------------------------------
/bindings/ruby/ext/cpu.mk:
--------------------------------------------------------------------------------
1 | ggml/src/ggml-cpu/ggml-cpu-cpp.o: \
2 | ggml/src/ggml-cpu/ggml-cpu.cpp \
3 | ggml/include/ggml-backend.h \
4 | ggml/include/ggml.h \
5 | ggml/include/ggml-alloc.h \
6 | ggml/src/ggml-backend-impl.h \
7 | ggml/include/ggml-cpu.h \
8 | ggml/src/ggml-impl.h
9 | $(CXX) $(CXXFLAGS) -c $< -o $@
10 |
--------------------------------------------------------------------------------
/examples/whisper.android/.gitignore:
--------------------------------------------------------------------------------
1 | *.iml
2 | .gradle
3 | /local.properties
4 | /.idea/caches
5 | /.idea/libraries
6 | /.idea/modules.xml
7 | /.idea/workspace.xml
8 | /.idea/navEditor.xml
9 | /.idea/assetWizardSettings.xml
10 | .DS_Store
11 | /build
12 | /captures
13 | .externalNativeBuild
14 | .cxx
15 | local.properties
16 |
--------------------------------------------------------------------------------
/ggml/src/ggml-threading.cpp:
--------------------------------------------------------------------------------
1 | #include "ggml-threading.h"
2 | #include
3 |
4 | std::mutex ggml_critical_section_mutex;
5 |
6 | void ggml_critical_section_start() {
7 | ggml_critical_section_mutex.lock();
8 | }
9 |
10 | void ggml_critical_section_end(void) {
11 | ggml_critical_section_mutex.unlock();
12 | }
13 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/.gitignore:
--------------------------------------------------------------------------------
1 | *.iml
2 | .gradle
3 | /local.properties
4 | /.idea/caches
5 | /.idea/libraries
6 | /.idea/modules.xml
7 | /.idea/workspace.xml
8 | /.idea/navEditor.xml
9 | /.idea/assetWizardSettings.xml
10 | .DS_Store
11 | /build
12 | /captures
13 | .externalNativeBuild
14 | .cxx
15 | local.properties
16 |
--------------------------------------------------------------------------------
/bindings/ruby/extsources.rb:
--------------------------------------------------------------------------------
1 | require "yaml"
2 |
3 | sources = `git ls-files -z ../..`.split("\x0")
4 | paths = YAML.load_file("../../.github/workflows/bindings-ruby.yml")[true]["push"]["paths"]
5 | paths.delete "bindings/ruby/**"
6 | EXTSOURCES = (Dir.glob(paths, base: "../..").collect {|path| "../../#{path}"} << "../../LICENSE") & sources
7 |
--------------------------------------------------------------------------------
/cmake/whisper.pc.in:
--------------------------------------------------------------------------------
1 | prefix=@CMAKE_INSTALL_PREFIX@
2 | exec_prefix=${prefix}
3 | libdir=${exec_prefix}/lib
4 | includedir=${prefix}/include
5 |
6 | Name: whisper
7 | Description: Port of OpenAI's Whisper model in C/C++
8 | Version: @PROJECT_VERSION@
9 | Libs: -L${libdir} -lggml -lggml-base -lwhisper
10 | Cflags: -I${includedir}
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-sycl/wkv.hpp:
--------------------------------------------------------------------------------
1 | #ifndef GGML_SYCL_WKV_HPP
2 | #define GGML_SYCL_WKV_HPP
3 |
4 | #include "common.hpp"
5 |
6 | void ggml_sycl_op_rwkv_wkv6(ggml_backend_sycl_context & ctx, ggml_tensor * dst);
7 |
8 | void ggml_sycl_op_rwkv_wkv7(ggml_backend_sycl_context & ctx, ggml_tensor * dst);
9 |
10 | #endif // GGML_SYCL_WKV_HPP
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | find_package (Threads REQUIRED)
2 |
3 | set(TARGET vulkan-shaders-gen)
4 | add_executable(${TARGET} vulkan-shaders-gen.cpp)
5 | install(TARGETS ${TARGET} RUNTIME)
6 | target_compile_features(${TARGET} PRIVATE cxx_std_17)
7 | target_link_libraries(vulkan-shaders-gen PUBLIC Threads::Threads)
8 |
--------------------------------------------------------------------------------
/bindings/java/gradle.properties:
--------------------------------------------------------------------------------
1 | org.gradle.jvmargs=-Xms256m -Xmx1024m
2 | system.include.dir=/usr/include
3 | #system.local.include.dir=../../include
4 | system.local.include.dir=./build/generated/sources/headers/java/main
5 | jni.include.dir=/usr/lib/jvm/java-8-openjdk-amd64/include/
6 | jni.lib.dir=/usr/lib/jvm/java-8-openjdk-amd64/lib/
7 |
--------------------------------------------------------------------------------
/examples/stream/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | if (WHISPER_SDL2)
2 | set(TARGET whisper-stream)
3 | add_executable(${TARGET} stream.cpp)
4 |
5 | include(DefaultTargetOptions)
6 |
7 | target_link_libraries(${TARGET} PRIVATE common common-sdl whisper ${CMAKE_THREAD_LIBS_INIT})
8 |
9 | install(TARGETS ${TARGET} RUNTIME)
10 | endif ()
11 |
--------------------------------------------------------------------------------
/examples/whisper.android/build.gradle:
--------------------------------------------------------------------------------
1 | // Top-level build file where you can add configuration options common to all sub-projects/modules.
2 | plugins {
3 | id 'com.android.application' version '8.1.1' apply false
4 | id 'com.android.library' version '8.1.1' apply false
5 | id 'org.jetbrains.kotlin.android' version '1.9.0' apply false
6 | }
--------------------------------------------------------------------------------
/examples/command/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | if (WHISPER_SDL2)
2 | set(TARGET whisper-command)
3 | add_executable(${TARGET} command.cpp)
4 |
5 | include(DefaultTargetOptions)
6 |
7 | target_link_libraries(${TARGET} PRIVATE common common-sdl whisper ${CMAKE_THREAD_LIBS_INIT})
8 |
9 | install(TARGETS ${TARGET} RUNTIME)
10 | endif ()
11 |
--------------------------------------------------------------------------------
/examples/whisper.android/app/src/main/res/mipmap-anydpi/ic_launcher.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
--------------------------------------------------------------------------------
/examples/whisper.objc/whisper.objc/SceneDelegate.h:
--------------------------------------------------------------------------------
1 | //
2 | // SceneDelegate.h
3 | // whisper.objc
4 | //
5 | // Created by Georgi Gerganov on 23.10.22.
6 | //
7 |
8 | #import
9 |
10 | @interface SceneDelegate : UIResponder
11 |
12 | @property (strong, nonatomic) UIWindow * window;
13 |
14 | @end
15 |
16 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-anydpi-v26/ic_launcher.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
--------------------------------------------------------------------------------
/ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.comp:
--------------------------------------------------------------------------------
1 | #extension GL_EXT_control_flow_attributes : require
2 | #extension GL_EXT_shader_16bit_storage : require
3 |
4 | layout (push_constant) uniform parameter
5 | {
6 | uint M;
7 | uint K;
8 | uint stride_a;
9 | uint stride_b;
10 | uint nel;
11 | } p;
12 |
13 | #include "types.comp"
14 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/mipmap-anydpi-v26/ic_launcher_round.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
--------------------------------------------------------------------------------
/examples/whisper.objc/whisper.objc.xcodeproj/project.xcworkspace/xcshareddata/IDEWorkspaceChecks.plist:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 | IDEDidComputeMac32BitWarning
6 |
7 |
8 |
9 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/cpy.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_CPY_BLOCK_SIZE 64
4 |
5 | void ggml_cuda_cpy(ggml_backend_cuda_context & ctx, const ggml_tensor * src0, ggml_tensor * src1);
6 |
7 | void ggml_cuda_dup(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
8 |
9 | void* ggml_cuda_cpy_fn(const ggml_tensor * src0, ggml_tensor * src1);
10 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.xcodeproj/project.xcworkspace/xcshareddata/IDEWorkspaceChecks.plist:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 | IDEDidComputeMac32BitWarning
6 |
7 |
8 |
9 |
--------------------------------------------------------------------------------
/scripts/convert-all.sh:
--------------------------------------------------------------------------------
1 | #!/bin/bash
2 |
3 | models=( "tiny.en" "tiny" "base.en" "base" "small.en" "small" "medium.en" "medium" "large-v1" "large-v2" "large-v3" "large-v3-turbo" )
4 |
5 | for model in "${models[@]}"; do
6 | python3 models/convert-pt-to-ggml.py ~/.cache/whisper/$model.pt ../whisper models/
7 | mv -v models/ggml-model.bin models/ggml-$model.bin
8 | done
9 |
--------------------------------------------------------------------------------
/examples/wchess/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | add_subdirectory(libwchess)
2 | set_target_properties(wchess-core PROPERTIES FOLDER "libs")
3 |
4 | if (EMSCRIPTEN)
5 | add_subdirectory(wchess.wasm)
6 | set_target_properties(wchess.wasm PROPERTIES FOLDER "libs")
7 | else()
8 | add_subdirectory(wchess.cmd)
9 | set_target_properties(wchess PROPERTIES FOLDER "libs")
10 | endif()
11 |
--------------------------------------------------------------------------------
/examples/addon.node/package.json:
--------------------------------------------------------------------------------
1 | {
2 | "name": "addon.node",
3 | "version": "0.0.0",
4 | "description": "",
5 | "main": "index.js",
6 | "author": "Qanhe Chen",
7 | "license": "MIT",
8 | "scripts": {
9 | "test": "jest"
10 | },
11 | "devDependencies": {
12 | "cmake-js": "^7.1.1",
13 | "jest": "^29.4.0",
14 | "node-addon-api": "^5.0.0"
15 | }
16 | }
17 |
--------------------------------------------------------------------------------
/examples/server/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | set(TARGET whisper-server)
2 | add_executable(${TARGET} server.cpp httplib.h)
3 |
4 | include(DefaultTargetOptions)
5 |
6 | target_link_libraries(${TARGET} PRIVATE common json_cpp whisper ${CMAKE_THREAD_LIBS_INIT})
7 |
8 | if (WIN32)
9 | target_link_libraries(${TARGET} PRIVATE ws2_32)
10 | endif()
11 |
12 | install(TARGETS ${TARGET} RUNTIME)
13 |
--------------------------------------------------------------------------------
/samples/README.md:
--------------------------------------------------------------------------------
1 | # Audio samples
2 |
3 | This folder contains various audio files used for testing.
4 | If you want to quickly get some more samples, simply run `make samples`. This will download several public audio files and convert them to appropriate 16-bit WAV format using `ffmpeg`
5 |
6 | https://github.com/ggerganov/whisper.cpp/blob/a09ce6e8899198015729ffc49ae10f67370906b1/Makefile#L104-L123
7 |
--------------------------------------------------------------------------------
/scripts/gen-authors.sh:
--------------------------------------------------------------------------------
1 | #!/bin/bash
2 |
3 | printf "# date: $(date)\n" > AUTHORS
4 | printf "# this file is auto-generated by scripts/gen-authors.sh\n\n" >> AUTHORS
5 |
6 | git log --format='%an <%ae>' --reverse --date=short master | awk '!seen[$0]++' | sort >> AUTHORS
7 |
8 | # if necessary, update your name here. for example: jdoe -> John Doe
9 | sed -i '' 's/^jdoe/John Doe/g' AUTHORS
10 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/java/com/litongjava/whisper/android/java/app/App.java:
--------------------------------------------------------------------------------
1 | package com.litongjava.whisper.android.java.app;
2 |
3 | import android.app.Application;
4 |
5 | import com.blankj.utilcode.util.Utils;
6 |
7 | public class App extends Application {
8 | @Override
9 | public void onCreate() {
10 | super.onCreate();
11 | Utils.init(this);
12 | }
13 | }
14 |
--------------------------------------------------------------------------------
/examples/whisper.android/app/src/main/java/com/whispercppdemo/ui/theme/Color.kt:
--------------------------------------------------------------------------------
1 | package com.whispercppdemo.ui.theme
2 |
3 | import androidx.compose.ui.graphics.Color
4 |
5 | val Purple80 = Color(0xFFD0BCFF)
6 | val PurpleGrey80 = Color(0xFFCCC2DC)
7 | val Pink80 = Color(0xFFEFB8C8)
8 |
9 | val Purple40 = Color(0xFF6650a4)
10 | val PurpleGrey40 = Color(0xFF625b71)
11 | val Pink40 = Color(0xFF7D5260)
--------------------------------------------------------------------------------
/examples/sycl/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | # MIT license
2 | # Copyright (C) 2024 Intel Corporation
3 | # SPDX-License-Identifier: MIT
4 |
5 | set(TARGET ls-sycl-device)
6 | add_executable(${TARGET} ls-sycl-device.cpp)
7 | install(TARGETS ${TARGET} RUNTIME)
8 | target_link_libraries(${TARGET} PRIVATE common whisper ${CMAKE_THREAD_LIBS_INIT})
9 | target_compile_features(${TARGET} PRIVATE cxx_std_17)
10 |
--------------------------------------------------------------------------------
/ggml/src/ggml-sycl/cpy.hpp:
--------------------------------------------------------------------------------
1 | #ifndef GGML_SYCL_CPY_HPP
2 | #define GGML_SYCL_CPY_HPP
3 |
4 | #include "common.hpp"
5 |
6 | typedef void (*cpy_kernel_t)(const char * cx, char * cdst);
7 |
8 | void ggml_sycl_cpy(ggml_backend_sycl_context & ctx, const ggml_tensor * src0, const ggml_tensor * src1);
9 | void ggml_sycl_dup(ggml_backend_sycl_context & ctx, ggml_tensor * dst);
10 |
11 | #endif // GGML_SYCL_CPY_HPP
12 |
--------------------------------------------------------------------------------
/ggml/src/ggml-sycl/sycl_hw.cpp:
--------------------------------------------------------------------------------
1 | #include "sycl_hw.hpp"
2 |
3 |
4 | sycl_hw_info get_device_hw_info(sycl::device *device_ptr) {
5 | sycl_hw_info res;
6 | int32_t id = device_ptr->get_info();
7 | res.device_id = id;
8 |
9 | syclex::architecture arch = device_ptr->get_info();
10 | res.arch = arch;
11 |
12 | return res;
13 | }
14 |
--------------------------------------------------------------------------------
/bindings/java/src/main/java/io/github/ggerganov/whispercpp/params/WhisperSamplingStrategy.java:
--------------------------------------------------------------------------------
1 | package io.github.ggerganov.whispercpp.params;
2 |
3 | /** Available sampling strategies */
4 | public enum WhisperSamplingStrategy {
5 | /** similar to OpenAI's GreedyDecoder */
6 | WHISPER_SAMPLING_GREEDY,
7 |
8 | /** similar to OpenAI's BeamSearchDecoder */
9 | WHISPER_SAMPLING_BEAM_SEARCH
10 | }
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cpu/llamafile/sgemm.h:
--------------------------------------------------------------------------------
1 | #pragma once
2 | #include
3 | #include
4 | #ifdef __cplusplus
5 | extern "C" {
6 | #endif
7 |
8 | bool llamafile_sgemm(const struct ggml_compute_params * params, int64_t, int64_t, int64_t,
9 | const void *, int64_t, const void *, int64_t, void *, int64_t,
10 | int, int, int);
11 |
12 | #ifdef __cplusplus
13 | }
14 | #endif
15 |
--------------------------------------------------------------------------------
/grammars/colors.gbnf:
--------------------------------------------------------------------------------
1 | # - red
2 | # - green
3 | # - blue
4 | #
5 | # example:
6 | #
7 | # ./command -m ./models/ggml-tiny.en.bin -t 8 --grammar ./grammars/colors.gbnf --prompt "red, green, blue," --context "green, red, blue,"
8 | #
9 |
10 | root ::= init color "."
11 | prompt ::= init "."
12 |
13 | # leading space is very important!
14 | init ::= " red, green, blue"
15 |
16 | color ::= ", " ("red" | "green" | "blue")
17 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/java/com/whispercpp/java/whisper/WhisperCpuConfig.java:
--------------------------------------------------------------------------------
1 | package com.whispercpp.java.whisper;
2 |
3 | import android.os.Build;
4 |
5 | import androidx.annotation.RequiresApi;
6 |
7 | public class WhisperCpuConfig {
8 | @RequiresApi(api = Build.VERSION_CODES.N)
9 | public static int getPreferredThreadCount() {
10 | return Math.max(CpuInfo.getHighPerfCpuCount(), 2);
11 | }
12 | }
13 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cpu/kleidiai/kleidiai.h:
--------------------------------------------------------------------------------
1 | // SPDX-FileCopyrightText: Copyright 2025 Arm Limited and/or its affiliates
2 | // SPDX-License-Identifier: MIT
3 | //
4 |
5 | #pragma once
6 |
7 | #include "ggml-alloc.h"
8 |
9 | #ifdef __cplusplus
10 | extern "C" {
11 | #endif
12 |
13 | ggml_backend_buffer_type_t ggml_backend_cpu_kleidiai_buffer_type(void);
14 |
15 | #ifdef __cplusplus
16 | }
17 | #endif
18 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_8.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 1, 8);
6 | DECL_FATTN_MMA_F16_CASE(80, 1, 8);
7 | DECL_FATTN_MMA_F16_CASE(96, 1, 8);
8 | DECL_FATTN_MMA_F16_CASE(112, 1, 8);
9 | DECL_FATTN_MMA_F16_CASE(128, 1, 8);
10 | DECL_FATTN_MMA_F16_CASE(256, 1, 8);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_4.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 2, 4);
6 | DECL_FATTN_MMA_F16_CASE(80, 2, 4);
7 | DECL_FATTN_MMA_F16_CASE(96, 2, 4);
8 | DECL_FATTN_MMA_F16_CASE(112, 2, 4);
9 | DECL_FATTN_MMA_F16_CASE(128, 2, 4);
10 | DECL_FATTN_MMA_F16_CASE(256, 2, 4);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_8.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 2, 8);
6 | DECL_FATTN_MMA_F16_CASE(80, 2, 8);
7 | DECL_FATTN_MMA_F16_CASE(96, 2, 8);
8 | DECL_FATTN_MMA_F16_CASE(112, 2, 8);
9 | DECL_FATTN_MMA_F16_CASE(128, 2, 8);
10 | DECL_FATTN_MMA_F16_CASE(256, 2, 8);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_2.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 4, 2);
6 | DECL_FATTN_MMA_F16_CASE(80, 4, 2);
7 | DECL_FATTN_MMA_F16_CASE(96, 4, 2);
8 | DECL_FATTN_MMA_F16_CASE(112, 4, 2);
9 | DECL_FATTN_MMA_F16_CASE(128, 4, 2);
10 | DECL_FATTN_MMA_F16_CASE(256, 4, 2);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_4.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 4, 4);
6 | DECL_FATTN_MMA_F16_CASE(80, 4, 4);
7 | DECL_FATTN_MMA_F16_CASE(96, 4, 4);
8 | DECL_FATTN_MMA_F16_CASE(112, 4, 4);
9 | DECL_FATTN_MMA_F16_CASE(128, 4, 4);
10 | DECL_FATTN_MMA_F16_CASE(256, 4, 4);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_8.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 4, 8);
6 | DECL_FATTN_MMA_F16_CASE(80, 4, 8);
7 | DECL_FATTN_MMA_F16_CASE(96, 4, 8);
8 | DECL_FATTN_MMA_F16_CASE(112, 4, 8);
9 | DECL_FATTN_MMA_F16_CASE(128, 4, 8);
10 | DECL_FATTN_MMA_F16_CASE(256, 4, 8);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 8, 1);
6 | DECL_FATTN_MMA_F16_CASE(80, 8, 1);
7 | DECL_FATTN_MMA_F16_CASE(96, 8, 1);
8 | DECL_FATTN_MMA_F16_CASE(112, 8, 1);
9 | DECL_FATTN_MMA_F16_CASE(128, 8, 1);
10 | DECL_FATTN_MMA_F16_CASE(256, 8, 1);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_2.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 8, 2);
6 | DECL_FATTN_MMA_F16_CASE(80, 8, 2);
7 | DECL_FATTN_MMA_F16_CASE(96, 8, 2);
8 | DECL_FATTN_MMA_F16_CASE(112, 8, 2);
9 | DECL_FATTN_MMA_F16_CASE(128, 8, 2);
10 | DECL_FATTN_MMA_F16_CASE(256, 8, 2);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_4.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 8, 4);
6 | DECL_FATTN_MMA_F16_CASE(80, 8, 4);
7 | DECL_FATTN_MMA_F16_CASE(96, 8, 4);
8 | DECL_FATTN_MMA_F16_CASE(112, 8, 4);
9 | DECL_FATTN_MMA_F16_CASE(128, 8, 4);
10 | DECL_FATTN_MMA_F16_CASE(256, 8, 4);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_8.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 8, 8);
6 | DECL_FATTN_MMA_F16_CASE(80, 8, 8);
7 | DECL_FATTN_MMA_F16_CASE(96, 8, 8);
8 | DECL_FATTN_MMA_F16_CASE(112, 8, 8);
9 | DECL_FATTN_MMA_F16_CASE(128, 8, 8);
10 | DECL_FATTN_MMA_F16_CASE(256, 8, 8);
11 |
--------------------------------------------------------------------------------
/examples/whisper.android/settings.gradle:
--------------------------------------------------------------------------------
1 | pluginManagement {
2 | repositories {
3 | gradlePluginPortal()
4 | google()
5 | mavenCentral()
6 | }
7 | }
8 | dependencyResolutionManagement {
9 | repositoriesMode.set(RepositoriesMode.FAIL_ON_PROJECT_REPOS)
10 | repositories {
11 | google()
12 | mavenCentral()
13 | }
14 | }
15 | rootProject.name = "WhisperCppDemo"
16 | include ':app'
17 | include ':lib'
18 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 16, 1);
6 | DECL_FATTN_MMA_F16_CASE(80, 16, 1);
7 | DECL_FATTN_MMA_F16_CASE(96, 16, 1);
8 | DECL_FATTN_MMA_F16_CASE(112, 16, 1);
9 | DECL_FATTN_MMA_F16_CASE(128, 16, 1);
10 | DECL_FATTN_MMA_F16_CASE(256, 16, 1);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_2.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 16, 2);
6 | DECL_FATTN_MMA_F16_CASE(80, 16, 2);
7 | DECL_FATTN_MMA_F16_CASE(96, 16, 2);
8 | DECL_FATTN_MMA_F16_CASE(112, 16, 2);
9 | DECL_FATTN_MMA_F16_CASE(128, 16, 2);
10 | DECL_FATTN_MMA_F16_CASE(256, 16, 2);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_4.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 16, 4);
6 | DECL_FATTN_MMA_F16_CASE(80, 16, 4);
7 | DECL_FATTN_MMA_F16_CASE(96, 16, 4);
8 | DECL_FATTN_MMA_F16_CASE(112, 16, 4);
9 | DECL_FATTN_MMA_F16_CASE(128, 16, 4);
10 | DECL_FATTN_MMA_F16_CASE(256, 16, 4);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_32-ncols2_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 32, 1);
6 | DECL_FATTN_MMA_F16_CASE(80, 32, 1);
7 | DECL_FATTN_MMA_F16_CASE(96, 32, 1);
8 | DECL_FATTN_MMA_F16_CASE(112, 32, 1);
9 | DECL_FATTN_MMA_F16_CASE(128, 32, 1);
10 | DECL_FATTN_MMA_F16_CASE(256, 32, 1);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_32-ncols2_2.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 32, 2);
6 | DECL_FATTN_MMA_F16_CASE(80, 32, 2);
7 | DECL_FATTN_MMA_F16_CASE(96, 32, 2);
8 | DECL_FATTN_MMA_F16_CASE(112, 32, 2);
9 | DECL_FATTN_MMA_F16_CASE(128, 32, 2);
10 | DECL_FATTN_MMA_F16_CASE(256, 32, 2);
11 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_64-ncols2_1.cu:
--------------------------------------------------------------------------------
1 | // This file has been autogenerated by generate_cu_files.py, do not edit manually.
2 |
3 | #include "../fattn-mma-f16.cuh"
4 |
5 | DECL_FATTN_MMA_F16_CASE(64, 64, 1);
6 | DECL_FATTN_MMA_F16_CASE(80, 64, 1);
7 | DECL_FATTN_MMA_F16_CASE(96, 64, 1);
8 | DECL_FATTN_MMA_F16_CASE(112, 64, 1);
9 | DECL_FATTN_MMA_F16_CASE(128, 64, 1);
10 | DECL_FATTN_MMA_F16_CASE(256, 64, 1);
11 |
--------------------------------------------------------------------------------
/examples/whisper.android.java/app/src/main/res/values/colors.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 | #FFBB86FC
4 | #FF6200EE
5 | #FF3700B3
6 | #FF03DAC5
7 | #FF018786
8 | #FF000000
9 | #FFFFFFFF
10 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cpu/amx/mmq.h:
--------------------------------------------------------------------------------
1 | #pragma once
2 | #include "common.h"
3 |
4 | size_t ggml_backend_amx_desired_wsize(const struct ggml_tensor * dst);
5 |
6 | size_t ggml_backend_amx_get_alloc_size(const struct ggml_tensor * tensor);
7 |
8 | void ggml_backend_amx_convert_weight(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size);
9 |
10 | void ggml_backend_amx_mul_mat(const struct ggml_compute_params * params, struct ggml_tensor * dst);
11 |
--------------------------------------------------------------------------------
/bindings/go/go.mod:
--------------------------------------------------------------------------------
1 | module github.com/ggerganov/whisper.cpp/bindings/go
2 |
3 | go 1.23
4 |
5 | require (
6 | github.com/go-audio/wav v1.1.0
7 | github.com/stretchr/testify v1.9.0
8 | )
9 |
10 | require (
11 | github.com/davecgh/go-spew v1.1.1 // indirect
12 | github.com/go-audio/audio v1.0.0 // indirect
13 | github.com/go-audio/riff v1.0.0 // indirect
14 | github.com/pmezard/go-difflib v1.0.0 // indirect
15 | gopkg.in/yaml.v3 v3.0.1 // indirect
16 | )
17 |
--------------------------------------------------------------------------------
/examples/wchess/libwchess/CMakeLists.txt:
--------------------------------------------------------------------------------
1 | add_library(wchess-core STATIC
2 | WChess.cpp
3 | WChess.h
4 | Chessboard.cpp
5 | Chessboard.h
6 | )
7 |
8 | target_link_libraries(wchess-core
9 | PUBLIC
10 | whisper
11 | common
12 | )
13 |
14 | target_include_directories(wchess-core
15 | PUBLIC
16 | "$"
17 | )
18 |
19 | # add_executable(test-chessboard test-chessboard.cpp Chessboard.cpp)
20 |
--------------------------------------------------------------------------------
/examples/talk-llama/llama-io.cpp:
--------------------------------------------------------------------------------
1 | #include "llama-io.h"
2 |
3 | void llama_io_write_i::write_string(const std::string & str) {
4 | uint32_t str_size = str.size();
5 |
6 | write(&str_size, sizeof(str_size));
7 | write(str.data(), str_size);
8 | }
9 |
10 | void llama_io_read_i::read_string(std::string & str) {
11 | uint32_t str_size;
12 | read_to(&str_size, sizeof(str_size));
13 |
14 | str.assign((const char *) read(str_size), str_size);
15 | }
16 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/convert.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | #define CUDA_DEQUANTIZE_BLOCK_SIZE 256
4 |
5 | template
6 | using to_t_cuda_t = void (*)(const void * __restrict__ x, T * __restrict__ y, int64_t k, cudaStream_t stream);
7 |
8 | typedef to_t_cuda_t to_fp32_cuda_t;
9 | typedef to_t_cuda_t to_fp16_cuda_t;
10 |
11 | to_fp16_cuda_t ggml_get_to_fp16_cuda(ggml_type type);
12 |
13 | to_fp32_cuda_t ggml_get_to_fp32_cuda(ggml_type type);
14 |
--------------------------------------------------------------------------------
/cmake/DefaultTargetOptions.cmake:
--------------------------------------------------------------------------------
1 | # Set the default compile features and properties for a target.
2 |
3 | if (NOT TARGET)
4 | message(FATAL_ERROR "TARGET not set before including DefaultTargetOptions")
5 | endif()
6 |
7 | target_compile_features(${TARGET}
8 | PRIVATE
9 | cxx_std_11
10 | )
11 |
12 | set_target_properties(${TARGET}
13 | PROPERTIES
14 | EXPORT_COMPILE_COMMANDS ON
15 | RUNTIME_OUTPUT_DIRECTORY "${CMAKE_BINARY_DIR}/bin"
16 | )
17 |
--------------------------------------------------------------------------------
/examples/whisper.android/app/src/test/java/com/whispercppdemo/ExampleUnitTest.kt:
--------------------------------------------------------------------------------
1 | package com.whispercppdemo
2 |
3 | import org.junit.Test
4 |
5 | import org.junit.Assert.*
6 |
7 | /**
8 | * Example local unit test, which will execute on the development machine (host).
9 | *
10 | * See [testing documentation](http://d.android.com/tools/testing).
11 | */
12 | class ExampleUnitTest {
13 | @Test
14 | fun addition_isCorrect() {
15 | assertEquals(4, 2 + 2)
16 | }
17 | }
--------------------------------------------------------------------------------
/SDL2/README-SDL.txt:
--------------------------------------------------------------------------------
1 |
2 | Please distribute this file with the SDL runtime environment:
3 |
4 | The Simple DirectMedia Layer (SDL for short) is a cross-platform library
5 | designed to make it easy to write multi-media software, such as games
6 | and emulators.
7 |
8 | The Simple DirectMedia Layer library source code is available from:
9 | https://www.libsdl.org/
10 |
11 | This library is distributed under the terms of the zlib license:
12 | http://www.zlib.net/zlib_license.html
13 |
14 |
--------------------------------------------------------------------------------
/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp:
--------------------------------------------------------------------------------
1 | #version 450
2 |
3 | #include "types.comp"
4 | #include "generic_unary_head.comp"
5 |
6 | layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
7 |
8 | void main() {
9 | const uint idx = get_idx();
10 |
11 | if (idx >= p.ne) {
12 | return;
13 | }
14 |
15 | const FLOAT_TYPE val = FLOAT_TYPE(data_a[get_aoffset() + src0_idx(idx)]);
16 | data_d[get_doffset() + dst_idx(idx)] = D_TYPE(cos(val));
17 | }
18 |
--------------------------------------------------------------------------------
/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp:
--------------------------------------------------------------------------------
1 | #version 450
2 |
3 | #include "types.comp"
4 | #include "generic_unary_head.comp"
5 |
6 | layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
7 |
8 | void main() {
9 | const uint idx = get_idx();
10 |
11 | if (idx >= p.ne) {
12 | return;
13 | }
14 |
15 | const FLOAT_TYPE val = FLOAT_TYPE(data_a[get_aoffset() + src0_idx(idx)]);
16 | data_d[get_doffset() + dst_idx(idx)] = D_TYPE(sin(val));
17 | }
18 |
--------------------------------------------------------------------------------
/ggml/src/ggml-vulkan/vulkan-shaders/square.comp:
--------------------------------------------------------------------------------
1 | #version 450
2 |
3 | #include "types.comp"
4 | #include "generic_unary_head.comp"
5 |
6 | layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
7 |
8 | void main() {
9 | const uint idx = get_idx();
10 |
11 | if (idx >= p.ne) {
12 | return;
13 | }
14 |
15 | const FLOAT_TYPE val = FLOAT_TYPE(data_a[get_aoffset() + src0_idx(idx)]);
16 | data_d[get_doffset() + dst_idx(idx)] = D_TYPE(val * val);
17 | }
18 |
--------------------------------------------------------------------------------
/ggml/src/ggml-amx/mmq.h:
--------------------------------------------------------------------------------
1 | #pragma once
2 | #include "common.h"
3 | #include
4 |
5 | #ifdef __cplusplus
6 | extern "C" {
7 | #endif
8 |
9 | size_t ggml_backend_amx_get_alloc_size(const struct ggml_tensor * tensor);
10 |
11 | void ggml_backend_amx_convert_weight(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size);
12 |
13 | void ggml_backend_amx_mul_mat(ggml_backend_amx_context * ctx, struct ggml_tensor * dst);
14 |
15 | #ifdef __cplusplus
16 | }
17 | #endif
18 |
--------------------------------------------------------------------------------
/ggml/src/ggml-cuda/norm.cuh:
--------------------------------------------------------------------------------
1 | #include "common.cuh"
2 |
3 | void ggml_cuda_op_norm(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
4 |
5 | void ggml_cuda_op_group_norm(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
6 |
7 | void ggml_cuda_op_rms_norm(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
8 |
9 | void ggml_cuda_op_rms_norm_back(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
10 |
11 | void ggml_cuda_op_l2_norm(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
12 |
--------------------------------------------------------------------------------
/examples/common-ggml.h:
--------------------------------------------------------------------------------
1 | #pragma once
2 |
3 | #include "ggml.h"
4 |
5 | #include
6 | #include
7 | #include
8 |
9 | enum ggml_ftype ggml_parse_ftype(const char * str);
10 |
11 | void ggml_print_ftypes(FILE * fp = stderr);
12 |
13 | bool ggml_common_quantize_0(
14 | std::ifstream & finp,
15 | std::ofstream & fout,
16 | const ggml_ftype ftype,
17 | const std::vector & to_quant,
18 | const std::vector & to_skip);
19 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/Supporting files/WhisperCppDemo.entitlements:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 | com.apple.security.app-sandbox
6 |
7 | com.apple.security.device.audio-input
8 |
9 | com.apple.security.files.user-selected.read-only
10 |
11 |
12 |
13 |
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/Utils/RiffWaveUtils.swift:
--------------------------------------------------------------------------------
1 | import Foundation
2 |
3 | func decodeWaveFile(_ url: URL) throws -> [Float] {
4 | let data = try Data(contentsOf: url)
5 | let floats = stride(from: 44, to: data.count, by: 2).map {
6 | return data[$0..<$0 + 2].withUnsafeBytes {
7 | let short = Int16(littleEndian: $0.load(as: Int16.self))
8 | return max(-1.0, min(Float(short) / 32767.0, 1.0))
9 | }
10 | }
11 | return floats
12 | }
13 |
--------------------------------------------------------------------------------
/.github/workflows/bindings-go.yml:
--------------------------------------------------------------------------------
1 | name: Bindings Tests (Go)
2 | on:
3 | push:
4 | paths:
5 | - bindings/go/**
6 | - whisper.h
7 | pull_request:
8 | paths:
9 | - bindings/go/**
10 | - whisper.h
11 |
12 | jobs:
13 | ubuntu-22:
14 | runs-on: ubuntu-22.04
15 | steps:
16 | - uses: actions/setup-go@v5
17 | with:
18 | go-version: '^1.23'
19 | - uses: actions/checkout@v4
20 | - run: |
21 | cd bindings/go
22 | make test
23 |
--------------------------------------------------------------------------------
/examples/whisper.android/.idea/misc.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
--------------------------------------------------------------------------------
/examples/sycl/run-whisper.sh:
--------------------------------------------------------------------------------
1 | #!/bin/bash
2 |
3 | # MIT license
4 | # Copyright (C) 2024 Intel Corporation
5 | # SPDX-License-Identifier: MIT
6 |
7 | INPUT2="Building a website can be done in 10 simple steps:\nStep 1:"
8 | source /opt/intel/oneapi/setvars.sh
9 |
10 | if [ $# -gt 0 ]; then
11 | export GGML_SYCL_DEVICE=$1
12 | else
13 | export GGML_SYCL_DEVICE=0
14 | fi
15 | echo GGML_SYCL_DEVICE=$GGML_SYCL_DEVICE
16 | #export GGML_SYCL_DEBUG=1
17 | ./build/bin/main -m models/ggml-base.en.bin -f samples/jfk.wav
--------------------------------------------------------------------------------
/examples/whisper.swiftui/whisper.swiftui.demo/Models/Model.swift:
--------------------------------------------------------------------------------
1 | import Foundation
2 |
3 | struct Model: Identifiable {
4 | var id = UUID()
5 | var name: String
6 | var info: String
7 | var url: String
8 |
9 | var filename: String
10 | var fileURL: URL {
11 | FileManager.default.urls(for: .documentDirectory, in: .userDomainMask)[0].appendingPathComponent(filename)
12 | }
13 |
14 | func fileExists() -> Bool {
15 | FileManager.default.fileExists(atPath: fileURL.path)
16 | }
17 | }
18 |
--------------------------------------------------------------------------------
/ggml/src/ggml-sycl/sycl_hw.hpp:
--------------------------------------------------------------------------------
1 | #ifndef SYCL_HW_HPP
2 | #define SYCL_HW_HPP
3 |
4 | #include
5 | #include
6 | #include
7 | #include