voice-sage95m-int8

INT8 ONNX export of ai-forever/sage-fredt5-distilled-95m (SAGE Russian spelling correction, FRED-T5 distilled 95M) for on-device transcript polishing in the «Диктофон» voice-recorder app.

Exported with optimum (ONNX, encoder + no-past decoder) and dynamically quantized to INT8 (arm64). The no-past decoder trades ~1.7x generation speed for a 74 MB smaller download - fine for sentence-sized chunks.

Files:

• encoder_model_quantized.onnx (~45 MB)
• decoder_model_quantized.onnx (~77 MB, no KV cache)
• vocab.json, merges.txt - GPT-2 byte-level BPE (FRED-T5)

Special tokens: BOS <LM>=50357 (prepended to encoder input), EOS </s>=2, decoder start <pad>=0.

License: MIT (inherited from the base model).