--- language: - tw - ak - en license: apache-2.0 library_name: transformers pipeline_tag: automatic-speech-recognition base_model: openai/whisper-small tags: - ak - akan - en - ghana - ghana-health-ai - serendepify - tw - twi datasets: - google/WaxalNLP - fsicoli/common_voice_22_0 - ghananlpcommunity/twi-speech-text-multispeaker-16k model-index: - name: gha-whisper-small-twi-en-balanced-v7-lite results: - task: type: automatic-speech-recognition dataset: type: google/WaxalNLP name: google/WaxalNLP (aka_asr) metrics: - type: wer value: 0.423143 name: WER - type: cer value: 0.161800 name: CER --- # teckedd/gha-whisper-small-twi-en-balanced-v7-lite Balanced Twi/Akan + English-retention Whisper small proof checkpoint for Ghana Health AI. Validation WER 42.31% on the capped Waxal validation subset; not a production promotion candidate. - **Base model:** `openai/whisper-small` - **Task:** `automatic-speech-recognition` - **Languages:** `tw`, `ak`, `en` - **Card generated:** 2026-08-09 (UTC) - **Product:** [ghanahealth.serendepify.com](https://ghanahealth.serendepify.com) > Not a medical device. Outputs support community health guidance only. ## Metrics | Metric | Value | |--------|-------| | `wer` | 0.4231 | | `cer` | 0.1618 | ## Training data - `google/WaxalNLP (aka_asr)` - `fsicoli/common_voice_22_0 (tw)` - `fsicoli/common_voice_22_0 (en)` - `ghananlpcommunity/twi-speech-text-multispeaker-16k (default)` ## Intended use - In-product ASR / TTS / chat for Twi (Akan) and English health conversations in Ghana. - Research on low-resource Ghanaian language speech and health dialogue. ## Out of scope - Clinical diagnosis or autonomous medical decisions. - Claiming near-native quality without reporting held-out WER/CER or human A/B scores. ## How to load ```python from transformers import pipeline # or AutoModel + processor per task # repo: teckedd/gha-whisper-small-twi-en-balanced-v7-lite ``` ## Citation If you use this checkpoint, please credit **Ghana Health AI / Serendepify** and the upstream base model authors plus any listed datasets.