applied PR #1152 to other recipes

This commit is contained in:
zr_jin 2023-07-24 19:19:34 +08:00 committed by jinzr
parent 3cfd75c4db
commit c03c011230
8 changed files with 10 additions and 10 deletions

View File

@ -613,7 +613,7 @@ def main():
quantize_dynamic(
model_input=decoder_filename,
model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"],
op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8,
)

View File

@ -608,7 +608,7 @@ def main():
quantize_dynamic(
model_input=decoder_filename,
model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"],
op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8,
)

View File

@ -508,7 +508,7 @@ def main():
quantize_dynamic(
model_input=decoder_filename,
model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"],
op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8,
)

View File

@ -518,7 +518,7 @@ def main():
quantize_dynamic(
model_input=decoder_filename,
model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"],
op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8,
)

View File

@ -63,8 +63,8 @@ import sentencepiece as spm
import torch
import torch.nn as nn
from conformer import Conformer
from onnxruntime.quantization import QuantType, quantize_dynamic
from decoder import Decoder
from onnxruntime.quantization import QuantType, quantize_dynamic
from scaling_converter import convert_scaled_to_non_scaled
from train import add_model_arguments, get_params, get_transducer_model
@ -662,7 +662,7 @@ def main():
quantize_dynamic(
model_input=decoder_filename,
model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"],
op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8,
)

View File

@ -60,8 +60,8 @@ import sentencepiece as spm
import torch
import torch.nn as nn
from conformer import Conformer
from onnxruntime.quantization import QuantType, quantize_dynamic
from decoder import Decoder
from onnxruntime.quantization import QuantType, quantize_dynamic
from scaling_converter import convert_scaled_to_non_scaled
from train import add_model_arguments, get_params, get_transducer_model
@ -586,7 +586,7 @@ def main():
quantize_dynamic(
model_input=decoder_filename,
model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"],
op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8,
)

View File

@ -661,7 +661,7 @@ def main():
quantize_dynamic(
model_input=decoder_filename,
model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"],
op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8,
)

View File

@ -652,7 +652,7 @@ def main():
quantize_dynamic(
model_input=decoder_filename,
model_output=decoder_filename_int8,
op_types_to_quantize=["MatMul"],
op_types_to_quantize=["MatMul", "Gather"],
weight_type=QuantType.QInt8,
)