mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-08-08 17:42:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

Support int8 quantization in decoder (#1152)

Browse Source
This commit is contained in:
Wei Kang 2023-06-29 16:48:59 +08:00 committed by GitHub
parent 9c2172c1c4
commit db71b03026
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
egs/librispeech/ASR/zipformer/export-onnx-streaming.py
View File

@ -757,7 +757,7 @@ def main():
quantize_dynamic( quantize_dynamic(
model_input=decoder_filename, model_input=decoder_filename,
model_output=decoder_filename_int8, model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"], op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8, weight_type=QuantType.QInt8,
) )

2
egs/librispeech/ASR/zipformer/export-onnx.py
View File

@ -602,7 +602,7 @@ def main():
quantize_dynamic( quantize_dynamic(
model_input=decoder_filename, model_input=decoder_filename,
model_output=decoder_filename_int8, model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"], op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8, weight_type=QuantType.QInt8,
) )