From 4daea3711063c64485be3c00eaa9727404549f51 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期五, 24 二月 2023 17:55:00 +0800
Subject: [PATCH] onnx
---
funasr/export/models/predictor/cif.py | 3 +
funasr/export/export_model.py | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++++
funasr/export/models/__init__.py | 1
3 files changed, 108 insertions(+), 1 deletions(-)
diff --git a/funasr/export/export_model.py b/funasr/export/export_model.py
index 3c73152..e1d5fdb 100644
--- a/funasr/export/export_model.py
+++ b/funasr/export/export_model.py
@@ -117,6 +117,111 @@
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()
+ )
+
+
if __name__ == '__main__':
import sys
diff --git a/funasr/export/models/__init__.py b/funasr/export/models/__init__.py
index ca2c813..27a65af 100644
--- a/funasr/export/models/__init__.py
+++ b/funasr/export/models/__init__.py
@@ -1,5 +1,6 @@
from funasr.models.e2e_asr_paraformer import Paraformer
from funasr.export.models.e2e_asr_paraformer import Paraformer as Paraformer_export
+from funasr.models.e2e_uni_asr import UniASR
def get_model(model, export_config=None):
diff --git a/funasr/export/models/predictor/cif.py b/funasr/export/models/predictor/cif.py
index 5518cb8..fcfcd5f 100644
--- a/funasr/export/models/predictor/cif.py
+++ b/funasr/export/models/predictor/cif.py
@@ -109,7 +109,8 @@
frames = torch.stack(list_frames, 1)
list_ls = []
len_labels = torch.round(alphas.sum(-1)).int()
- max_label_len = len_labels.max()
+ max_label_len = len_labels.max().item()
+ print("type: {}".format(type(max_label_len)))
for b in range(batch_size):
fire = fires[b, :]
l = torch.index_select(frames[b, :, :], 0, torch.nonzero(fire >= threshold).squeeze())
--
Gitblit v1.9.1