Add fp16 variant: encoder + static/dyn decoders (verified packer; fp32 transcript parity; ~3.5x faster wall on DML) f879e68 verified Masterx commited on Jul 11
Upload onnx/decoder_model_merged.onnx_data with huggingface_hub a895b7e verified Masterx commited on Jul 7
Upload onnx/decoder_model_merged_q4.onnx_data with huggingface_hub 5cea9f8 verified Masterx commited on Jul 7
Upload onnx/encoder_model_q4.onnx_data with huggingface_hub 3181076 verified Masterx commited on Jul 7