Fetching metadata from the HF Docker repository...
Mar 1 15:22 CRITICAL FIX: Use correct Wav2Vec2 model (lv-60-espeak-cv-ft) matching training Running - app.py inference was using wav2vec2-large-xlsr-53 but model was trained
on wav2vec2-lv-60-espeak-cv-ft embeddings -> always wrong predictions
- Dockerfile: pre-download correct model (lv-60-espeak-cv-ft)
- Language dropdown now optional for audio/URL classification
- Add model download endpoint for admin panel (ephemeral HF storage)
- Add Download Model button in admin results 13:37 Add admin panel for model training with SSE progress 12:36 Retrain model on input_data.csv embeddings with PCA(256d), threshold=0.4 12:11 Fix model architecture: add PCA transform (2051->256d) to match training pipeline 11:58 Add Git LFS tracking and update gitignore for HF Spaces 11:56 Deploy to Hugging Face Spaces 11:45 added model 11:20 added dockerfile 11:05 added web interface 09:40 Merge branch 'Utkarsharma7:main' into main 09:32 deleted some files 09:29 Merge pull request #2 from Utkarsharma7/revert-1-main Revert "added web interface and a pre trained model" 09:29 Revert "added web interface and a pre trained model" 09:24 Final change 08:42 Merge pull request #1 from Neev-Rathod/main added web interface and a pre trained model 08:41 Merge branch 'main' into main 08:37 added web interface 08:32 Updated with cnn and comparision with cnn,mlp done 03:27 cnn but not tested Sarthak-gupta-cpp
02:51 less accuracy better graph Sarthak-gupta-cpp
02:10 changes Sarthak-gupta-cpp
00:00 Updated Feb 28 18:23 added pretrained models 16:44 Updated lowerd overfitting 16:34 Arabic-only: prosody layers, remove English, add test metrics, anti-overfitting, FLOW.md Made-with: Cursor Feb 27 23:55 Added classification for english also 10:58 Changed to XLSR model Feb 19 12:02 Created a git ignore file 11:20 Initial commit 
