From f98c4bf6d2bb5202488cd4243efdbca65288c313 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期一, 27 二月 2023 14:26:32 +0800
Subject: [PATCH] onnx export
---
funasr/export/export_model.py | 110 ++-----------------------------------------------------
1 files changed, 4 insertions(+), 106 deletions(-)
diff --git a/funasr/export/export_model.py b/funasr/export/export_model.py
index 933a927..8d41462 100644
--- a/funasr/export/export_model.py
+++ b/funasr/export/export_model.py
@@ -7,10 +7,12 @@
import logging
import torch
-from funasr.bin.asr_inference_paraformer import Speech2Text
from funasr.export.models import get_model
import numpy as np
import random
+
+torch_version = float(".".join(torch.__version__.split(".")[:2]))
+assert torch_version > 1.9
class ASRModelExportParaformer:
def __init__(self, cache_dir: Union[Path, str] = None, onnx: bool = True):
@@ -30,7 +32,7 @@
def _export(
self,
- model: Speech2Text,
+ model,
tag_name: str = None,
verbose: bool = False,
):
@@ -112,110 +114,6 @@
os.path.join(path, f'{model.model_name}.onnx'),
verbose=verbose,
opset_version=14,
- input_names=model.get_input_names(),
- output_names=model.get_output_names(),
- dynamic_axes=model.get_dynamic_axes()
- )
-
-
-class ASRModelExport:
- def __init__(self, cache_dir: Union[Path, str] = None, onnx: bool = True):
- assert check_argument_types()
- self.set_all_random_seed(0)
- if cache_dir is None:
- cache_dir = Path.home() / ".cache" / "export"
-
- self.cache_dir = Path(cache_dir)
- self.export_config = dict(
- feats_dim=560,
- onnx=False,
- )
- print("output dir: {}".format(self.cache_dir))
- self.onnx = onnx
-
- def _export(
- self,
- model: Speech2Text,
- tag_name: str = None,
- verbose: bool = False,
- ):
-
- export_dir = self.cache_dir / tag_name.replace(' ', '-')
- os.makedirs(export_dir, exist_ok=True)
-
- # export encoder1
- self.export_config["model_name"] = "model"
- model = get_model(
- model,
- self.export_config,
- )
- model.eval()
- # self._export_onnx(model, verbose, export_dir)
- if self.onnx:
- self._export_onnx(model, verbose, export_dir)
- else:
- self._export_torchscripts(model, verbose, export_dir)
-
- print("output dir: {}".format(export_dir))
-
- def _export_torchscripts(self, model, verbose, path, enc_size=None):
- if enc_size:
- dummy_input = model.get_dummy_inputs(enc_size)
- else:
- dummy_input = model.get_dummy_inputs_txt()
-
- # model_script = torch.jit.script(model)
- model_script = torch.jit.trace(model, dummy_input)
- model_script.save(os.path.join(path, f'{model.model_name}.torchscripts'))
-
- def set_all_random_seed(self, seed: int):
- random.seed(seed)
- np.random.seed(seed)
- torch.random.manual_seed(seed)
-
- def export(self,
- tag_name: str = 'damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch',
- mode: str = 'paraformer',
- ):
-
- model_dir = tag_name
- if model_dir.startswith('damo/'):
- from modelscope.hub.snapshot_download import snapshot_download
- model_dir = snapshot_download(model_dir, cache_dir=self.cache_dir)
- asr_train_config = os.path.join(model_dir, 'config.yaml')
- asr_model_file = os.path.join(model_dir, 'model.pb')
- cmvn_file = os.path.join(model_dir, 'am.mvn')
- json_file = os.path.join(model_dir, 'configuration.json')
- if mode is None:
- import json
- with open(json_file, 'r') as f:
- config_data = json.load(f)
- mode = config_data['model']['model_config']['mode']
- if mode.startswith('paraformer'):
- from funasr.tasks.asr import ASRTaskParaformer as ASRTask
- elif mode.startswith('uniasr'):
- from funasr.tasks.asr import ASRTaskUniASR as ASRTask
-
- model, asr_train_args = ASRTask.build_model_from_file(
- asr_train_config, asr_model_file, cmvn_file, 'cpu'
- )
- self._export(model, tag_name)
-
- def _export_onnx(self, model, verbose, path, enc_size=None):
- if enc_size:
- dummy_input = model.get_dummy_inputs(enc_size)
- else:
- dummy_input = model.get_dummy_inputs()
-
- # model_script = torch.jit.script(model)
- model_script = model # torch.jit.trace(model)
-
- torch.onnx.export(
- model_script,
- dummy_input,
- os.path.join(path, f'{model.model_name}.onnx'),
- verbose=verbose,
- opset_version=12,
input_names=model.get_input_names(),
output_names=model.get_output_names(),
dynamic_axes=model.get_dynamic_axes()
--
Gitblit v1.9.1