mirror of
https://github.com/modelscope/FunASR
synced 2025-09-15 14:48:36 +08:00
f2d8ded57f
* qwenaudio qwenaudiochat * qwenaudio qwenaudiochat * whisper * whisper * llm * llm * llm * llm * llm * llm * llm * llm * export onnx
74 lines
2.0 KiB
Python
74 lines
2.0 KiB
Python
import os
|
|
import torch
|
|
|
|
def export_onnx(model,
|
|
data_in=None,
|
|
type: str = "onnx",
|
|
quantize: bool = False,
|
|
fallback_num: int = 5,
|
|
calib_num: int = 100,
|
|
opset_version: int = 14,
|
|
**kwargs):
|
|
model_scripts = model.export(**kwargs)
|
|
export_dir = kwargs.get("output_dir", os.path.dirname(kwargs.get("init_param")))
|
|
os.makedirs(export_dir, exist_ok=True)
|
|
|
|
if not isinstance(model_scripts, (list, tuple)):
|
|
model_scripts = (model_scripts,)
|
|
for m in model_scripts:
|
|
m.eval()
|
|
_onnx(m,
|
|
data_in=data_in,
|
|
type=type,
|
|
quantize=quantize,
|
|
fallback_num=fallback_num,
|
|
calib_num=calib_num,
|
|
opset_version=opset_version,
|
|
export_dir=export_dir,
|
|
**kwargs
|
|
)
|
|
print("output dir: {}".format(export_dir))
|
|
|
|
return export_dir
|
|
|
|
def _onnx(model,
|
|
data_in=None,
|
|
quantize: bool = False,
|
|
opset_version: int = 14,
|
|
export_dir:str = None,
|
|
**kwargs):
|
|
|
|
dummy_input = model.export_dummy_inputs()
|
|
|
|
verbose = kwargs.get("verbose", False)
|
|
|
|
export_name = model.export_name() if hasattr(model, "export_name") else "model.onnx"
|
|
model_path = os.path.join(export_dir, export_name)
|
|
torch.onnx.export(
|
|
model,
|
|
dummy_input,
|
|
model_path,
|
|
verbose=verbose,
|
|
opset_version=opset_version,
|
|
input_names=model.export_input_names(),
|
|
output_names=model.export_output_names(),
|
|
dynamic_axes=model.export_dynamic_axes()
|
|
)
|
|
|
|
if quantize:
|
|
from onnxruntime.quantization import QuantType, quantize_dynamic
|
|
import onnx
|
|
quant_model_path = model_path.replace(".onnx", "_quant.onnx")
|
|
if not os.path.exists(quant_model_path):
|
|
onnx_model = onnx.load(model_path)
|
|
nodes = [n.name for n in onnx_model.graph.node]
|
|
nodes_to_exclude = [m for m in nodes if 'output' in m or 'bias_encoder' in m or 'bias_decoder' in m]
|
|
quantize_dynamic(
|
|
model_input=model_path,
|
|
model_output=quant_model_path,
|
|
op_types_to_quantize=['MatMul'],
|
|
per_channel=True,
|
|
reduce_range=False,
|
|
weight_type=QuantType.QUInt8,
|
|
nodes_to_exclude=nodes_to_exclude,
|
|
) |