Yesterday
-
v1.2: Pan-African fine-tuned recognition engine (rec.onnx)
-
v1.2: Add African language tags and benchmarks (Fongbe, Yoruba, Peul, Biali, Bariba, Dendi)
-
v1.2: Unified Pan-African 5-gram LM (138k sentences: Fon, Yoruba, Peul, Biali, Bariba, Dendi, Fr)
-
v1.2: Pan-African fine-tuned recognition engine (rec.onnx)
-
v1.1: Update documentation with Fongbé and Yorùbá specialization tags and benchmarks
-
v1.1: Retrain unified 5-gram language model on Fon + Yoruba + French corpus (lm.npz)
-
v1.1: Fine-tune Latin/African text recognition engine (rec.onnx)
-
Update visuals: fix banner UTF-8 accents and update throughput benchmark (Finanfa OCR 45 p/s)
-
Initial release: Finanfa OCR - State of the Art Offline CPU Document Intelligence
-
initial commit