Spaces:
Runtime error
Runtime error
Commit History
-
Store full IP address without anonymization
Ko-TTS-Arena Contributors +1: Claude Opus 4.5Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>d9af61c -
Use access_route for IP detection in HF Spaces
Ko-TTS-Arena Contributors +1: Claude Opus 4.5- Add access_route fallback since HF Spaces doesn't pass X-Forwarded-For - Remove debug-headers endpoint Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>b2e640a -
Add debug-headers endpoint to check proxy headers
Ko-TTS-Arena Contributorsbb088b9 -
Add ProxyFix middleware to get real client IP
Ko-TTS-Arena Contributors +1: Claude Opus 4.5HF Spaces uses reverse proxy, so X-Forwarded-For header needs to be properly handled via ProxyFix middleware. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>c7fb090 -
Change preferences repo to channelcorp/ko-tts-arena-db
Ko-TTS-Arena Contributors +1: Claude Opus 4.5Use same dataset repo for both database and preferences backup Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>0d8b012 -
Improve UI design with modern sophisticated styling
Ko-TTS-Arena Contributors +1: Claude Opus 4.5- Add Pretendard font for better Korean typography - Update color palette with gradient primary colors - Add glassmorphism effect to sidebar and main content - Improve button styles with hover animations - Add smooth transitions and shadows - Update dark mode with better colors - Improve player card and vote button styling - Add subtle backdrop blur effects Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>58b72ea -
Add ElevenLabs v3 model and change voice to Olivia
Ko-TTS-Arena Contributors +1: Claude Opus 4.5- Change default voice ID from Rachel to Olivia (spzDnDXYekkOPfsxp0U3) - Add eleven-v3 model (eleven_v3) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>34ec563 -
Revert "Fix database path fallback to use /tmp instead of instance/"
Ko-TTS-Arena ContributorsThis reverts commit 6087e14dcf2e79806fa154293f1463fcae620df1.2e3325e -
Fix database path fallback to use /tmp instead of instance/
Ko-TTS-Arena Contributors +1: Claude Opus 4.5HF Spaces may not have write permission to instance/ directory. Using /tmp/tts_arena/ as fallback which is always writable. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>6087e14
-
Fix unterminated string literal in typecast API URL
Ko-TTS-Arena Contributors +1: Claude Opus 4.5Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>6300890 -
Update Typecast TTS API to v1 endpoint (#3)
- Update Typecast TTS API to v1 endpoint (5eb1871ffc70226caa6fa7f6fec016d708f5c24b) Co-authored-by: SunghyunJun <SunghyunJun@users.noreply.huggingface.co>a2154fc
-
Fix fallback to prevent duplicate A/B models
Ko-TTS-Arena Contributors +1: Claude Opus 4.5Added assigned_model_ids tracking to ensure fallback never selects a model already used by another slot. Also checks if initial model is already taken and picks alternative. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>4924063 -
Add TTS fallback: retry with different model on API error
Ko-TTS-Arena Contributors +1: Claude Opus 4.5When a TTS provider fails (e.g., Typecast 404), automatically select and try another model instead of failing completely. Supports up to 3 retries per slot with weighted model selection. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>7cb6386
-
Increase exposure for low-vote models
Ko-TTS-Arena Contributors +1: Claude Sonnet 4.5- Reduce SMOOTHING_FACTOR from 500 to 100 - New models now 5x more likely to appear vs established models - Helps newly added models get voted on faster Example weights: - 0 votes: 1/(0+100) = 0.01 (was 0.002) - 100 votes: 1/(100+100) = 0.005 (was 0.00167) - Ratio: 2x difference (was 1.2x) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>33b5896 -
Remove all Hugging Face login buttons and prompts
Ko-TTS-Arena Contributors +1: Claude Sonnet 4.5- Removed login link from base.html navigation - Removed login tip overlay and mobile banner - Removed login prompt from leaderboard.html - Removed all login-related CSS and JavaScript - Users can still use the app without login UI clutter Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>cb03ebc -
Merge remote changes and resolve conflicts
Ko-TTS-Arena Contributors887a0d8 -
Improve model pairing: prioritize new and low-vote models
Ko-TTS-Arena Contributors +1: Claude Sonnet 4.5- Remove similarity-based weighting for model pairs - New models now match against both new and established models - All models weighted by vote count (fewer votes = higher probability) - Better exposure for newly added models in voting Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>540021e
-
Add Channel Hana 1.1 TTS model
Ko-TTS-Arena Contributors +1: Claude Sonnet 4.5- Add channel-hana-1.1 model to models.py - Add API endpoint and handler for Hana 1.1 TTS - Configure to use https://ch-hana-tts.channel.io Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>57d9f94 -
Fix JSON formatting: remove trailing comma
Ko-TTS-Arena Contributors +1: Claude Sonnet 4.5Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>1089319 -
Expand Korean TTS examples with more customer service scenarios
Ko-TTS-Arena Contributors +1: Claude Sonnet 4.5Add additional examples for hotels, restaurants, airlines, and facilities. Total 18 diverse customer service prompts. Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>fdbb1a7 -
Update Korean TTS example prompts with customer service scenarios
Ko-TTS-Arena Contributors +1: Claude Sonnet 4.5Replace existing prompts with 12 new customer service examples including pricing, contact info, orders, facilities, and transportation announcements. Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>6c1b55b
-
Sync from GitHub repo
GitHub Actionscf17ad1
-
Improve model pairing: match models with similar vote counts
Ko-TTS-Arena Contributors8167848
-
Add Humelo DIVE 나나 voice and retire 리아 voice
Ko-TTS-Arena Contributorsa998bdb -
Humelo: DIVE API Response Field명 변경 (#2)
Verified signature- Humelo: DIVE API Response Field명 변경 (fcbff51c70072c3bf86b6cf7e95516e59bead2a5) Co-authored-by: itsowavys <itsowavy@users.noreply.huggingface.co>c86fd8b -
Add Gemini 2.5 Flash Preview TTS (API Key based)
Ko-TTS-Arena Contributors14b59fc -
Disable Humelo TTS (API temporarily down)
Ko-TTS-Arena Contributorsa05825a -
chore: Humelo 목소리 업데이트 요청 (#1)
Verified signature- chore: Humelo 목소리 업데이트 요청 (c7ea15112d244b7f85bc1aa65297ecbe30e1c292) Co-authored-by: itsowavys <itsowavy@users.noreply.huggingface.co>79cbd6c
-
Update tts.py
Verified signature2f4c8f4
-
feat: Add category filters (한영혼합, 숫자혼합, 긴문장) to leaderboard
Ko-TTS-Arena Contributorse1cb33d -
fix: Hide inactive models from leaderboard
Ko-TTS-Arena Contributorsaebf092 -
feat: Keep only Typecast v3 (자인), deactivate 재선 and legacy models
Ko-TTS-Arena Contributors26dd2ec -
chore: Trigger space restart for new Typecast models
Ko-TTS-Arena Contributors1baa590 -
feat: Re-enable Typecast v3.0
Ko-TTS-Arena Contributorsa347681 -
fix: Update Typecast default model to ssfm-v30
Ko-TTS-Arena Contributors06db955 -
fix: Temporarily disable Typecast v3.0 (503 errors from HF Space)
Ko-TTS-Arena Contributors58cbf9c -
feat: Add retry logic for Typecast API (3 retries with exponential backoff)
Ko-TTS-Arena Contributorsee45344 -
feat: Upgrade Typecast to v3.0 API (ssfm-v30), add Jaesun & Jain voices, deactivate legacy GeumHee
Ko-TTS-Arena Contributorsf54c69b -
fix: Gemini TTS requires service account JSON (API Key not supported), disable until JSON provided
Ko-TTS-Arena Contributors515ea95 -
fix: Use REST API v1beta1 for Gemini TTS, disable google-wavenet/neural2
Ko-TTS-Arena Contributorsc82ad44 -
fix: Use GEMINI_TTS_API_KEY -> GOOGLE_API_KEY for Gemini TTS (client lib reads this env var)
Ko-TTS-Arena Contributors50a02f7 -
fix: Require service account for Gemini TTS and disable legacy Typecast model
Ko-TTS-Arena Contributors45eb847 -
fix: Use REST API v1beta1 for Gemini TTS (supports API Key)
Ko-TTS-Arena Contributors6662253 -
fix: Gemini TTS requires service account JSON (API key not supported by client lib), disable until JSON provided
Ko-TTS-Arena Contributors0458096 -
fix: Use GOOGLE_API_KEY env var for Gemini TTS (simpler than service account)
Ko-TTS-Arena Contributorsa1a6334 -
feat: Enable Gemini TTS with service account JSON authentication
Ko-TTS-Arena Contributors673d09e -
fix: Add ffmpeg to Docker, disable Gemini TTS (requires OAuth2), keep REST API code for future
Ko-TTS-Arena Contributorsb0bdfc9 -
feat: Add Gemini TTS (Aoede voice) - Google Cloud Text-to-Speech
Ko-TTS-Arena Contributors407795c -
feat: Change Typecast voice from JaeYi to GeumHee
Ko-TTS-Arena Contributors9c1f76a -
feat: Add Typecast TTS (JaeYi voice) - https://typecast.ai/
Ko-TTS-Arena Contributors1af3f4d