Add auto to model/beam size to whisper #3859
Conversation
📝 Walkthrough📝 WalkthroughWalkthroughThe changes in this pull request involve updates to the Whisper add-on's documentation and configuration files. The default model selection has been modified to Changes
Sequence Diagram(s)sequenceDiagram
participant User
participant WhisperAddOn
participant Configuration
User->>WhisperAddOn: Request model selection
WhisperAddOn->>Configuration: Check default model
Configuration-->>WhisperAddOn: Return model (auto)
WhisperAddOn->>User: Provide selected model (auto)
User->>WhisperAddOn: Request beam size
WhisperAddOn->>Configuration: Check default beam size
Configuration-->>WhisperAddOn: Return beam size (0)
WhisperAddOn->>User: Provide selected beam size (0)
Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media? 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
CodeRabbit Configuration File (
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Outside diff range and nitpick comments (1)
whisper/DOCS.md (1)
74-74: Fix typo in beam size documentation.There's a small typo in the sentence.
-The default value of `0` will automatically select `1` of ARM devices like the Raspberry Pi 4 and `5` otherwise. +The default value of `0` will automatically select `1` for ARM devices like the Raspberry Pi 4 and `5` otherwise.
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
📒 Files selected for processing (3)
whisper/DOCS.md(2 hunks)whisper/build.yaml(1 hunks)whisper/config.yaml(2 hunks)
✅ Files skipped from review due to trivial changes (1)
- whisper/build.yaml
🧰 Additional context used
📓 Path-based instructions (2)
whisper/DOCS.md (6)
Pattern */**(html|markdown|md): - For instructional content in documentation, use a direct and authoritative tone. Avoid expressions of politeness such as 'may' or 'please', and ensure the goal of the instruction is fronted.
- Apply the Microsoft Style Guide to ensure documentation maintains clarity and conciseness.
- In step-by-step instructions, front the location phrase in the instructional sentence.
- In step-by-step instructions, front the 'goal' in the instructional sentence.
- In step-by-step instructions, if in doubt what to front, front the 'goal' before the location phrase in the instructional sentence.
- do not hyphenate terms like 'top-right' or 'bottom-left' with 'corner'
Pattern */**(html|markdown|md): - Use bold to mark UI strings.
- If "" are used to mark UI strings, replace them by bold.
Pattern */**(html|markdown|md): - Be brief in your replies and don't add fluff like "thank you for..." and "Please let me know if"
Pattern */**(html|markdown|md): - Use sentence-style capitalization also in headings.
Pattern */**(html|markdown|md): do not comment on HTML used for icons
Pattern */**(html|markdown|md): Avoid flagging inline HTML for embedding videos in future reviews for this repository.
whisper/config.yaml (6)
Pattern */**(html|markdown|md): - For instructional content in documentation, use a direct and authoritative tone. Avoid expressions of politeness such as 'may' or 'please', and ensure the goal of the instruction is fronted.
- Apply the Microsoft Style Guide to ensure documentation maintains clarity and conciseness.
- In step-by-step instructions, front the location phrase in the instructional sentence.
- In step-by-step instructions, front the 'goal' in the instructional sentence.
- In step-by-step instructions, if in doubt what to front, front the 'goal' before the location phrase in the instructional sentence.
- do not hyphenate terms like 'top-right' or 'bottom-left' with 'corner'
Pattern */**(html|markdown|md): - Use bold to mark UI strings.
- If "" are used to mark UI strings, replace them by bold.
Pattern */**(html|markdown|md): - Be brief in your replies and don't add fluff like "thank you for..." and "Please let me know if"
Pattern */**(html|markdown|md): - Use sentence-style capitalization also in headings.
Pattern */**(html|markdown|md): do not comment on HTML used for icons
Pattern */**(html|markdown|md): Avoid flagging inline HTML for embedding videos in future reviews for this repository.
🔇 Additional comments (4)
whisper/config.yaml (3)
2-2: LGTM: Version update matches PR objective.
Version updated to 2.4.0 as intended.
16-18: LGTM: Default values updated for auto-selection.
The new defaults enable automatic selection based on device architecture:
model: auto- selects between tiny-int8 and base-int8beam_size: 0- selects between 1 and 5
22-22: LGTM: Model schema updated to include auto option.
The schema has been properly updated to include the new auto option while maintaining all existing model choices.
whisper/DOCS.md (1)
39-44: LGTM: Clear documentation of auto model selection.
The documentation clearly explains the auto-selection behavior and provides good context about model types and their trade-offs.
|
Please take a look at the requested changes, and use the Ready for review button when you are done, thanks 👍 |
|
@frenck Changelog updated |
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Outside diff range and nitpick comments (1)
whisper/CHANGELOG.md (1)
5-5: Enhance the change description with specific default values.While the description captures the key changes, it would be more helpful to explicitly mention the new default values and their behavior.
-Add "auto" for model and beam size (0) to select values based on CPU +Add "auto" as default model (tiny-int8 for ARM, base-int8 for others) and beam size (0) for CPU-based optimization
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
📒 Files selected for processing (1)
whisper/CHANGELOG.md(1 hunks)
🧰 Additional context used
📓 Path-based instructions (1)
whisper/CHANGELOG.md (6)
Pattern */**(html|markdown|md): - For instructional content in documentation, use a direct and authoritative tone. Avoid expressions of politeness such as 'may' or 'please', and ensure the goal of the instruction is fronted.
- Apply the Microsoft Style Guide to ensure documentation maintains clarity and conciseness.
- In step-by-step instructions, front the location phrase in the instructional sentence.
- In step-by-step instructions, front the 'goal' in the instructional sentence.
- In step-by-step instructions, if in doubt what to front, front the 'goal' before the location phrase in the instructional sentence.
- do not hyphenate terms like 'top-right' or 'bottom-left' with 'corner'
Pattern */**(html|markdown|md): - Use bold to mark UI strings.
- If "" are used to mark UI strings, replace them by bold.
Pattern */**(html|markdown|md): - Be brief in your replies and don't add fluff like "thank you for..." and "Please let me know if"
Pattern */**(html|markdown|md): - Use sentence-style capitalization also in headings.
Pattern */**(html|markdown|md): do not comment on HTML used for icons
Pattern */**(html|markdown|md): Avoid flagging inline HTML for embedding videos in future reviews for this repository.
🔇 Additional comments (1)
whisper/CHANGELOG.md (1)
3-4: LGTM!
Version entry follows the established format and correctly reflects the version update.
Bump
wyoming-faster-whisperto 2.4.0: rhasspy/wyoming-faster-whisper@v2.3.0...v2.4.0Adds
autofor model and beam size (0), which are set as the new defaults.These will select different models/beam sizes depending on the platform.
Summary by CodeRabbit
New Features
auto, optimizing performance based on device architecture.autooption for model selection in the configuration.Bug Fixes
beam_sizeconfiguration for improved decoding performance.Documentation
Chores