Release v1.14.0

Author: julianmack

.clang-format

@@ -1,10 +1,9 @@
----
 AccessModifierOffset: -1
 AlignAfterOpenBracket: AlwaysBreak
 AlignConsecutiveAssignments: false
 AlignConsecutiveDeclarations: false
 AlignEscapedNewlinesLeft: true
-AlignOperands:   false
+AlignOperands: false
 AlignTrailingComments: true
 AllowAllParametersOfDeclarationOnNextLine: false
 AllowShortBlocksOnASingleLine: false
@@ -18,46 +17,46 @@ AlwaysBreakTemplateDeclarations: true
 BinPackArguments: false
 BinPackParameters: false
 BraceWrapping:
-  AfterClass:      false
+  AfterClass: false
   AfterControlStatement: false
-  AfterEnum:       false
-  AfterFunction:   false
-  AfterNamespace:  false
+  AfterEnum: false
+  AfterFunction: false
+  AfterNamespace: false
   AfterObjCDeclaration: false
-  AfterStruct:     false
-  AfterUnion:      false
-  BeforeCatch:     false
-  BeforeElse:      false
-  IndentBraces:    false
+  AfterStruct: false
+  AfterUnion: false
+  BeforeCatch: false
+  BeforeElse: false
+  IndentBraces: false
 BreakBeforeBinaryOperators: None
 BreakBeforeBraces: Attach
 BreakBeforeTernaryOperators: true
 BreakConstructorInitializersBeforeComma: false
 BreakAfterJavaFieldAnnotations: false
 BreakStringLiterals: false
-ColumnLimit:     100
-CommentPragmas:  '^ IWYU pragma:'
+ColumnLimit: 100
+CommentPragmas: '^ IWYU pragma:'
 CompactNamespaces: false
 ConstructorInitializerAllOnOneLineOrOnePerLine: true
 ConstructorInitializerIndentWidth: 4
 ContinuationIndentWidth: 4
 Cpp11BracedListStyle: true
 DerivePointerAlignment: false
-DisableFormat:   false
-ForEachMacros:   [ FOR_EACH_RANGE, FOR_EACH, ]
+DisableFormat: false
+ForEachMacros: [FOR_EACH_RANGE, FOR_EACH]
 IncludeCategories:
-  - Regex:           '^<.*\.h(pp)?>'
-    Priority:        1
-  - Regex:           '^<.*'
-    Priority:        2
-  - Regex:           '.*'
-    Priority:        3
+  - Regex: '^<.*\.h(pp)?>'
+    Priority: 1
+  - Regex: '^<.*'
+    Priority: 2
+  - Regex: '.*'
+    Priority: 3
 IndentCaseLabels: true
-IndentWidth:     2
+IndentWidth: 2
 IndentWrappedFunctionNames: false
 KeepEmptyLinesAtTheStartOfBlocks: false
 MacroBlockBegin: ''
-MacroBlockEnd:   ''
+MacroBlockEnd: ''
 MaxEmptyLinesToKeep: 1
 NamespaceIndentation: None
 PenaltyBreakBeforeFirstCallParameter: 1
@@ -67,18 +66,18 @@ PenaltyBreakString: 1000
 PenaltyExcessCharacter: 1000000
 PenaltyReturnTypeOnItsOwnLine: 2000000
 PointerAlignment: Left
-ReflowComments:  true
-SortIncludes:    true
+ReflowComments: true
+SortIncludes: true
 SpaceAfterCStyleCast: false
 SpaceBeforeAssignmentOperators: true
 SpaceBeforeParens: ControlStatements
 SpaceInEmptyParentheses: false
 SpacesBeforeTrailingComments: 1
-SpacesInAngles:  false
+SpacesInAngles: false
 SpacesInContainerLiterals: true
 SpacesInCStyleCastParentheses: false
 SpacesInParentheses: false
 SpacesInSquareBrackets: false
-Standard:        Cpp11
-TabWidth:        8
-UseTab:          Never
+Standard: Cpp11
+TabWidth: 8
+UseTab: Never

.github/workflows/mdbook.yaml

@@ -3,27 +3,22 @@
 # To get started with mdBook see: https://rust-lang.github.io/mdBook/index.html
 #
 name: Deploy mdBook site to Pages
-
 on:
   # Runs on pushes targeting the default branch
   push:
     branches: ["main"]
-
   # Allows you to run this workflow manually from the Actions tab
   workflow_dispatch:
-
 # Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages
 permissions:
   contents: read
   pages: write
   id-token: write
-
 # Allow only one concurrent deployment, skipping runs queued between the run in-progress and latest queued.
 # However, do NOT cancel in-progress runs as we want to allow these production deployments to complete.
 concurrency:
   group: "pages"
   cancel-in-progress: false
-
 jobs:
   # Build job
   build:
@@ -51,7 +46,6 @@ jobs:
       - name: Remove book artifact
         if: always()
         run: docker run -v $(realpath -e docs):/root/docs ubuntu:22.04 rm -rf /root/docs/book
-
   # Deployment job
   deploy:
     environment:

.gitignore

@@ -26,3 +26,6 @@ compile_commands.json
 
 # logs
 *.log
+
+# env
+*/.env

.mergify.yml

@@ -0,0 +1,20 @@
+queue_rules:
+  - name: default
+    # When auto-updating PRs for the base branch, rebase.
+    update_method: rebase
+    merge_method: merge
+pull_request_rules:
+  - name: Automatically rebase and merge pull requests
+    actions:
+      queue:
+        name: default
+    conditions:
+      - label=ready-for-merge
+      - label!=WIP
+      - "#approved-reviews-by>=1"
+      - "#changes-requested-reviews-by=0"
+  - name: Delete head branch after merge
+    actions:
+      delete_head_branch: {}
+    conditions:
+      - merged

.pre-commit-config.yaml

@@ -1,46 +1,58 @@
 # See https://pre-commit.com for more information
 # See https://pre-commit.com/hooks.html for more hooks
 repos:
--   repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v3.2.0
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v5.0.0
     hooks:
-    -   id: check-added-large-files
-    -   id: check-merge-conflict
-    -   id: end-of-file-fixer
+      - id: check-added-large-files
+      - id: check-merge-conflict
+      - id: end-of-file-fixer
         exclude_types: ["svg"]
-    -   id: requirements-txt-fixer
-    -   id: trailing-whitespace
--   repo: https://github.com/pycqa/isort
-    rev: 5.12.0
-    hooks:
-    - id: isort
-      args: [--profile=black, --project=caiman_asr_train, --project=myrtle_asr_train]
--   repo: https://github.com/psf/black
-    rev: 23.3.0
-    hooks:
-    - id: black
-      language_version: python3
-      args: [--line-length=88]
-- repo: https://github.com/astral-sh/ruff-pre-commit
-  # Ruff version.
-  rev: v0.1.5
-  hooks:
-    # Fix unused imports
-    - id: ruff
-      args: [--select, F401, --fix]
-      # This also excludes the smoke tests in myrtle-training
-      exclude: training/tests/test_smoke_tests.py
-- repo: https://github.com/PyCQA/flake8
-  rev: 7.0.0
-  hooks:
-  - id: flake8
-    args: [--max-line-length=92, "--extend-ignore=E203,F401,F722"]
-- repo: https://github.com/executablebooks/mdformat
-  rev: 0.7.17
-  hooks:
-  - id: mdformat
-    args: ["--number"]
-- repo: https://github.com/scop/pre-commit-shfmt
-  rev: v3.9.0-1
-  hooks:
-    - id: shfmt         # prebuilt upstream executable
+      - id: requirements-txt-fixer
+      - id: trailing-whitespace
+  - repo: https://github.com/pycqa/isort
+    rev: 6.0.0
+    hooks:
+      - id: isort
+        args: [--profile=black, --project=caiman_asr_train, --project=myrtle_asr_train]
+  - repo: https://github.com/psf/black
+    rev: 25.1.0
+    hooks:
+      - id: black
+        language_version: python3
+        args: [--line-length=88]
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    # Ruff version.
+    rev: v0.9.4
+    hooks:
+      # Fix unused imports
+      - id: ruff
+        args: [--select, F401, --fix]
+        # This also excludes the smoke tests in myrtle-training
+        exclude: training/tests/test_smoke_tests.py
+  - repo: https://github.com/PyCQA/flake8
+    rev: 7.1.1
+    hooks:
+      - id: flake8
+        args: [--max-line-length=92, "--extend-ignore=E203,F401,F722"]
+  - repo: https://github.com/executablebooks/mdformat
+    rev: 0.7.22
+    hooks:
+      - id: mdformat
+        additional_dependencies:
+          - mdformat-myst
+        args: ["--number"]
+  - repo: https://github.com/scop/pre-commit-shfmt
+    rev: v3.10.0-2
+    hooks:
+      - id: shfmt # prebuilt upstream executable
+  - repo: https://github.com/codespell-project/codespell
+    rev: v2.4.1
+    hooks:
+      - id: codespell
+        exclude: training/caiman_asr_train/data/text/english.json
+        args: ["-L", "nd,patter,crate,caf,strng,AAS"]
+  - repo: https://github.com/google/yamlfmt
+    rev: v0.15.0
+    hooks:
+      - id: yamlfmt

README.md

@@ -6,9 +6,9 @@ The solution is designed for **low-latency realtime streaming ASR workloads**.
 
 ## See next
 
-- Performance figures (Real-time streams, latency): [\[docs\]](https://caiman-asr.myrtle.ai/key_features.html#model-configurations) [\[markdown\]](docs/src/key_features.md#model-configurations)
-- WERs: [\[docs\]](https://caiman-asr.myrtle.ai/key_features.html#word-error-rates-wers) [\[markdown\]](docs/src/key_features.md#word-error-rates-wers)
-- Training speeds: [\[docs\]](https://caiman-asr.myrtle.ai/training/training_times.html) [\[markdown\]](docs/src/training/training_times.md)
-- Getting started: [\[docs\]](https://caiman-asr.myrtle.ai/introduction.html) [\[markdown\]](docs/src/introduction.md)
+- Performance figures (Real-time streams, latency): [[docs]](https://caiman-asr.myrtle.ai/key_features.html#model-configurations) [[markdown]](docs/src/key_features.md#model-configurations)
+- WERs: [[docs]](https://caiman-asr.myrtle.ai/key_features.html#word-error-rates-wers) [[markdown]](docs/src/key_features.md#word-error-rates-wers)
+- Training speeds: [[docs]](https://caiman-asr.myrtle.ai/training/training_times.html) [[markdown]](docs/src/training/training_times.md)
+- Getting started: [[docs]](https://caiman-asr.myrtle.ai/introduction.html) [[markdown]](docs/src/introduction.md)
 
 ![CAIMAN-ASR Logo](docs/src/assets/CAIMAN-ASR-wordmark.png)

docs/book.toml

@@ -9,6 +9,8 @@ title = "CAIMAN-ASR"
 
 mathjax-support = true
 additional-css = ["./mdbook-admonish.css"]
+curly-quotes = true
+git-repository-url = "https://github.com/MyrtleSoftware/caiman-asr"
 
 [preprocessor]
 

docs/src/SUMMARY.md

@@ -32,6 +32,7 @@
     - [Changing the character set](training/changing_the_character_set.md)
     - [Delay penalty](training/delay_penalty.md)
     - [Endpointing](training/endpointing.md)
+    - [Checkpoint averaging](training/checkpoint_averaging.md)
   - [Validation](training/validation.md)
     - [WER calculation](training/wer_calculation.md)
     - [State resets](training/state_resets.md)
@@ -41,6 +42,8 @@
     - [Saving predictions](training/saving_predictions.md)
     - [User-perceived latency](inference/user_perceived_latency.md)
     - [Emission latency](training/emission_latency.md)
+    - [Timestamps](training/timestamps.md)
+    - [Keyword boosting](training/keyword_boosting.md)
   - [Export inference checkpoint](training/export_inference_checkpoint.md)
 
 - [Inference flow](./inference/inference_flow.md)
@@ -54,5 +57,8 @@
     - [Live audio demonstration](./inference/live_audio_demonstration.md)
     - [Hardware requirements](./inference/hardware_requirements.md)
   - [CAIMAN-ASR demo](./inference/caiman-asr_demo.md)
+  - [CAIMAN-ASR benchmark](./inference/caiman-asr_benchmark.md)
+    - [Benchmark Installation](./inference/benchmark_installation.md)
+    - [Benchmarking custom datasets](./inference/benchmarking_custom_datasets.md)
 
 - [Versions](./versions.md)

docs/src/inference/benchmark_installation.md

@@ -0,0 +1,45 @@
+# Benchmark installation
+
+Follow these installation instructions to run the benchmark scripts in a Linux-based OS.
+
+## Install using a virtual environment
+
+(1) Install the basic dependencies:
+
+```bash
+cd benchmark
+sudo ./install_basic_deps.bash
+```
+
+(2) The following command will install the Python dependencies in a virtual environment:
+
+```bash
+./install_python_deps.bash
+```
+
+Note this requires Python >= 3.10.
+
+If you don't have the correct version of Python, you can
+use [uv](https://docs.astral.sh/uv/) to create
+the virtual environment with the correct version of Python:
+
+```bash
+USE_UV=true ./install_python_deps.bash
+```
+
+(3) Launch the virtual environment. All commands are meant to be run inside the environment:
+
+```bash
+source .venv/bin/activate
+```
+
+## Install using Docker
+
+(1) If you do not have docker installed, see [here](https://caiman-asr.myrtle.ai/training/installation.html).
+
+(2) Build the benchmarking docker image (takes about 10 minutes):
+
+```bash
+cd benchmark
+./build.bash
+```