Add meta data to NeMo canary ONNX models (#2351)

@@ -62,22 +62,7 @@ jobs:
           d=sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8
           d=sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8
           mkdir -p $d
           mkdir -p $d
           cp encoder.int8.onnx $d
           cp encoder.int8.onnx $d
-          cp decoder.fp16.onnx $d
-          cp tokens.txt $d
-
-          mkdir $d/test_wavs
-          cp de.wav $d/test_wavs
-          cp en.wav $d/test_wavs
-
-          tar cjfv $d.tar.bz2 $d
-
-      - name: Collect files (fp16)
-        shell: bash
-        run: |
-          d=sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-fp16
-          mkdir -p $d
-          cp encoder.fp16.onnx $d
-          cp decoder.fp16.onnx $d
+          cp decoder.int8.onnx $d
           cp tokens.txt $d
           cp tokens.txt $d
 
 
           mkdir $d/test_wavs
           mkdir $d/test_wavs
@@ -101,7 +86,6 @@ jobs:
             models=(
             models=(
               sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr
               sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr
               sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8
               sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8
-              sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-fp16
             )
             )
 
 
             for m in ${models[@]}; do
             for m in ${models[@]}; do

 #!/usr/bin/env python3
 #!/usr/bin/env python3
 # Copyright      2025  Xiaomi Corp.        (authors: Fangjun Kuang)
 # Copyright      2025  Xiaomi Corp.        (authors: Fangjun Kuang)
 
 
+"""
+<|en|>
+<|pnc|>
+<|noitn|>
+<|nodiarize|>
+<|notimestamp|>
+"""
+
 import os
 import os
-from typing import Tuple
+from typing import Dict, Tuple
 
 
 import nemo
 import nemo
-import onnxmltools
+import onnx
 import torch
 import torch
 from nemo.collections.common.parts import NEG_INF
 from nemo.collections.common.parts import NEG_INF
-from onnxmltools.utils.float16_converter import convert_float_to_float16
 from onnxruntime.quantization import QuantType, quantize_dynamic
 from onnxruntime.quantization import QuantType, quantize_dynamic
 
 
 """
 """
@@ -64,10 +71,25 @@ nemo.collections.common.parts.form_attention_mask = fixed_form_attention_mask
 from nemo.collections.asr.models import EncDecMultiTaskModel
 from nemo.collections.asr.models import EncDecMultiTaskModel
 
 
 
 
-def export_onnx_fp16(onnx_fp32_path, onnx_fp16_path):
-    onnx_fp32_model = onnxmltools.utils.load_model(onnx_fp32_path)
-    onnx_fp16_model = convert_float_to_float16(onnx_fp32_model, keep_io_types=True)
-    onnxmltools.utils.save_model(onnx_fp16_model, onnx_fp16_path)
+def add_meta_data(filename: str, meta_data: Dict[str, str]):
+    """Add meta data to an ONNX model. It is changed in-place.
+
+    Args:
+      filename:
+        Filename of the ONNX model to be changed.
+      meta_data:
+        Key-value pairs.
+    """
+    model = onnx.load(filename)
+    while len(model.metadata_props):
+        model.metadata_props.pop()
+
+    for key, value in meta_data.items():
+        meta = model.metadata_props.add()
+        meta.key = key
+        meta.value = str(value)
+
+    onnx.save(model, filename)
 
 
 
 
 def lens_to_mask(lens, max_length):
 def lens_to_mask(lens, max_length):
@@ -222,7 +244,7 @@ def export_decoder(canary_model):
         ),
         ),
         "decoder.onnx",
         "decoder.onnx",
         dynamo=True,
         dynamo=True,
-        opset_version=18,
+        opset_version=14,
         external_data=False,
         external_data=False,
         input_names=[
         input_names=[
             "decoder_input_ids",
             "decoder_input_ids",
@@ -269,6 +291,29 @@ def export_tokens(canary_model):
 @torch.no_grad()
 @torch.no_grad()
 def main():
 def main():
     canary_model = EncDecMultiTaskModel.from_pretrained("nvidia/canary-180m-flash")
     canary_model = EncDecMultiTaskModel.from_pretrained("nvidia/canary-180m-flash")
+    canary_model.eval()
+
+    preprocessor = canary_model.cfg["preprocessor"]
+    sample_rate = preprocessor["sample_rate"]
+    normalize_type = preprocessor["normalize"]
+    window_size = preprocessor["window_size"]  # ms
+    window_stride = preprocessor["window_stride"]  # ms
+    window = preprocessor["window"]
+    features = preprocessor["features"]
+    n_fft = preprocessor["n_fft"]
+    vocab_size = canary_model.tokenizer.vocab_size  # 5248
+
+    subsampling_factor = canary_model.cfg["encoder"]["subsampling_factor"]
+
+    assert sample_rate == 16000, sample_rate
+    assert normalize_type == "per_feature", normalize_type
+    assert window_size == 0.025, window_size
+    assert window_stride == 0.01, window_stride
+    assert window == "hann", window
+    assert features == 128, features
+    assert n_fft == 512, n_fft
+    assert subsampling_factor == 8, subsampling_factor
+
     export_tokens(canary_model)
     export_tokens(canary_model)
     export_encoder(canary_model)
     export_encoder(canary_model)
     export_decoder(canary_model)
     export_decoder(canary_model)
@@ -280,7 +325,32 @@ def main():
             weight_type=QuantType.QUInt8,
             weight_type=QuantType.QUInt8,
         )
         )
 
 
-        export_onnx_fp16(f"{m}.onnx", f"{m}.fp16.onnx")
+    meta_data = {
+        "vocab_size": vocab_size,
+        "normalize_type": normalize_type,
+        "subsampling_factor": subsampling_factor,
+        "model_type": "EncDecMultiTaskModel",
+        "version": "1",
+        "model_author": "NeMo",
+        "url": "https://huggingface.co/nvidia/canary-180m-flash",
+        "feat_dim": features,
+    }
+
+    add_meta_data("encoder.onnx", meta_data)
+    add_meta_data("encoder.int8.onnx", meta_data)
+
+    """
+    To fix the following error with onnxruntime 1.17.1 and 1.16.3:
+
+    onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 :FAIL : Load model from ./decoder.int8.onnx failed:/Users/runner/work/1/s/onnxruntime/core/graph/model.cc:150 onnxruntime::Model::Model(onnx::ModelProto &&, const onnxruntime::PathString &, const onnxruntime::IOnnxRuntimeOpSchemaRegistryList *, const logging::Logger &, const onnxruntime::ModelOptions &)
+    Unsupported model IR version: 10, max supported IR version: 9
+    """
+    for filename in ["./decoder.onnx", "./decoder.int8.onnx"]:
+        model = onnx.load(filename)
+        print("old", model.ir_version)
+        model.ir_version = 9
+        print("new", model.ir_version)
+        onnx.save(model, filename)
 
 
     os.system("ls -lh *.onnx")
     os.system("ls -lh *.onnx")
 
 

@@ -19,8 +19,8 @@ pip install \
   kaldi-native-fbank \
   kaldi-native-fbank \
   librosa \
   librosa \
   onnx==1.17.0 \
   onnx==1.17.0 \
-  onnxmltools \
   onnxruntime==1.17.1 \
   onnxruntime==1.17.1 \
+  onnxscript \
   soundfile
   soundfile
 
 
 python3 ./export_onnx_180m_flash.py
 python3 ./export_onnx_180m_flash.py
@@ -66,7 +66,7 @@ log "-----int8------"
 
 
 python3 ./test_180m_flash.py \
 python3 ./test_180m_flash.py \
   --encoder ./encoder.int8.onnx \
   --encoder ./encoder.int8.onnx \
-  --decoder ./decoder.fp16.onnx \
+  --decoder ./decoder.int8.onnx \
   --source-lang en \
   --source-lang en \
   --target-lang en \
   --target-lang en \
   --tokens ./tokens.txt \
   --tokens ./tokens.txt \
@@ -74,7 +74,7 @@ python3 ./test_180m_flash.py \
 
 
 python3 ./test_180m_flash.py \
 python3 ./test_180m_flash.py \
   --encoder ./encoder.int8.onnx \
   --encoder ./encoder.int8.onnx \
-  --decoder ./decoder.fp16.onnx \
+  --decoder ./decoder.int8.onnx \
   --source-lang en \
   --source-lang en \
   --target-lang de \
   --target-lang de \
   --tokens ./tokens.txt \
   --tokens ./tokens.txt \
@@ -82,7 +82,7 @@ python3 ./test_180m_flash.py \
 
 
 python3 ./test_180m_flash.py \
 python3 ./test_180m_flash.py \
   --encoder ./encoder.int8.onnx \
   --encoder ./encoder.int8.onnx \
-  --decoder ./decoder.fp16.onnx \
+  --decoder ./decoder.int8.onnx \
   --source-lang de \
   --source-lang de \
   --target-lang de \
   --target-lang de \
   --tokens ./tokens.txt \
   --tokens ./tokens.txt \
@@ -90,41 +90,7 @@ python3 ./test_180m_flash.py \
 
 
 python3 ./test_180m_flash.py \
 python3 ./test_180m_flash.py \
   --encoder ./encoder.int8.onnx \
   --encoder ./encoder.int8.onnx \
-  --decoder ./decoder.fp16.onnx \
-  --source-lang de \
-  --target-lang en \
-  --tokens ./tokens.txt \
-  --wav ./de.wav
-
-log "-----fp16------"
-
-python3 ./test_180m_flash.py \
-  --encoder ./encoder.fp16.onnx \
-  --decoder ./decoder.fp16.onnx \
-  --source-lang en \
-  --target-lang en \
-  --tokens ./tokens.txt \
-  --wav ./en.wav
-
-python3 ./test_180m_flash.py \
-  --encoder ./encoder.fp16.onnx \
-  --decoder ./decoder.fp16.onnx \
-  --source-lang en \
-  --target-lang de \
-  --tokens ./tokens.txt \
-  --wav ./en.wav
-
-python3 ./test_180m_flash.py \
-  --encoder ./encoder.fp16.onnx \
-  --decoder ./decoder.fp16.onnx \
-  --source-lang de \
-  --target-lang de \
-  --tokens ./tokens.txt \
-  --wav ./de.wav
-
-python3 ./test_180m_flash.py \
-  --encoder ./encoder.fp16.onnx \
-  --decoder ./decoder.fp16.onnx \
+  --decoder ./decoder.int8.onnx \
   --source-lang de \
   --source-lang de \
   --target-lang en \
   --target-lang en \
   --tokens ./tokens.txt \
   --tokens ./tokens.txt \

@@ -79,8 +79,7 @@ class OnnxModel:
         )
         )
 
 
         meta = self.encoder.get_modelmeta().custom_metadata_map
         meta = self.encoder.get_modelmeta().custom_metadata_map
-        #  self.normalize_type = meta["normalize_type"]
-        self.normalize_type = "per_feature"
+        self.normalize_type = meta["normalize_type"]
         print(meta)
         print(meta)
 
 
     def init_decoder(self, decoder):
     def init_decoder(self, decoder):
@@ -267,7 +266,7 @@ def main():
 
 
     for pos, decoder_input_id in enumerate(decoder_input_ids):
     for pos, decoder_input_id in enumerate(decoder_input_ids):
         logits, decoder_mems_list = model.run_decoder(
         logits, decoder_mems_list = model.run_decoder(
-            np.array([[decoder_input_id,pos]], dtype=np.int32),
+            np.array([[decoder_input_id, pos]], dtype=np.int32),
             decoder_mems_list,
             decoder_mems_list,
             enc_states,
             enc_states,
             enc_masks,
             enc_masks,