Fix Cell 14: replace evaluate.load('wer') with jiwer directly

Fix Cell 13: drop bitsandbytes, load whisper-small in fp16

Fix Cell 2: restart kernel after torchvision reinstall using sentinel file

Cell 12: never raise, let Cell 12b handle no-data case gracefully

Cell 4: remove Common Voice (moved to Mozilla Data Collective, off HF)

Fix Cell 12: add diagnostics + synthetic fallback cell (12b)

Fix Cell 11: bypass WhisperProcessor entirely to avoid torchvision crash

Fix Cell 2: reinstall torchvision to match Kaggle CUDA version at runtime

Fix Cells 2+11: unpin torch, use direct WhisperProcessor import

Fix Cell 10: correct r3 assertion to 'visit now!' (URL strips link not preceding word)

Fix Cell 10: add ! and ? to allowlist, fix smoke test assertions

Fix Cell 8: remove FLEURS fallback + strip emoji (charmap safe)

Fix Kaggle push: title slug + Windows UTF-8 encoding

Update push_to_kaggle.sh: accept KAGGLE_API_TOKEN env var (new API tokens)

Add Kaggle auto-sync: kernel-metadata.json + post-commit hook

Fix Cell 8: correct WaxalNLP subset names + drop fleurs fallback

Fix Cell 6: replace get_last_checkpoint import with inline scanner

Add Kaggle Master Trainer — Deep Sleep fine-tuning notebook

Remove myshell-openvoice from build — librosa==0.9.1 conflict

Fix OpenVoice build: package name is myshell-openvoice, not openvoice

Add multi-user speaker profiles, collective voice, and mode toggle

Add confidence loop, curiosity engine, and lightweight TTS

Fix Bambara TTS red state + surface detailed errors in UI

Fix: install maliba-ai lazily at runtime, not at build time

Fix: unpin soundfile and audiomentations to resolve maliba-ai conflicts

Fix: unpin librosa to resolve maliba-ai dependency conflict (needs 0.11.0)

Phase 2: Waxal TTS — Bambara voice output + Fula training notebook

Fix: switch LLM to Qwen2.5-72B-Instruct (Gemma not on HF free tier)

Phase 1: Sahel-Voice-Lab — The Memory Loop

Fix: replace YouTube download with audio upload (HF Spaces blocks outbound HTTP)

Add phrase library + fuzzy matcher for general Bambara/Fula conversation

Improve Bambara/Fula quality + add full correction UI

Fix: disable SSR mode and use port 7860 to unblock HF Spaces startup

Fix: use direct WhisperProcessor/WhisperForConditionalGeneration imports

Add setup.sh to pre-cache Whisper model during build

Fix: robust Whisper import + upgrade Gradio to fix API schema crash

Switch training dataset from google/waxal to google/fleurs

Fix: use AutoModelForSpeechSeq2Seq for transformers 5.x; disable buggy API schema endpoint

Fix: upgrade to Gradio 5 (HfFolder removed from huggingface-hub 1.x)

Initial commit: Sahel-Agri Voice AI