resolve conflicts

docs/source/recipes/Streaming-ASR/librispeech/lstm_pruned_stateless_transducer.rst

@@ -531,16 +531,36 @@ First, let us install a modified version of ``ncnn``:
   git clone https://github.com/csukuangfj/ncnn
   cd ncnn
   git submodule update --recursive --init
-  python3 setup.py bdist_wheel
+
-  ls -lh dist/
+  # Note: We don't use "python setup.py install" or "pip install ." here
-  pip install ./dist/*.whl
+
+  mkdir -p build-wheel
+  cd build-wheel
+
+  cmake \
+    -DCMAKE_BUILD_TYPE=Release \
+    -DNCNN_PYTHON=ON \
+    -DNCNN_BUILD_BENCHMARK=OFF \
+    -DNCNN_BUILD_EXAMPLES=OFF \
+    -DNCNN_BUILD_TOOLS=OFF \
+    ..
+
+  make -j4
+
+  cd ..
+
+  # Note: $PWD here is /path/to/ncnn
+
+  export PYTHONPATH=$PWD/python:$PYTHONPATH
+  export PATH=$PWD/tools/pnnx/build/src:$PATH
+  export PATH=$PWD/build/tools/quantize:$PATH
 
   # now build pnnx
   cd tools/pnnx
   mkdir build
   cd build
+  cmake ..
   make -j4
-  export PATH=$PWD/src:$PATH
 
   ./src/pnnx
 
@@ -549,6 +569,9 @@ First, let us install a modified version of ``ncnn``:
    We assume that you have added the path to the binary ``pnnx`` to the
    environment variable ``PATH``.
 
+   We also assume that you have added ``build/tools/quantize`` to the environment
+   variable ``PATH`` so that you are able to use ``ncnn2int8`` later.
+
 Second, let us export the model using ``torch.jit.trace()`` that is suitable
 for ``pnnx``:
 

egs/librispeech/ASR/conv_emformer_transducer_stateless2/emformer2.py

@@ -1512,24 +1512,6 @@ class EmformerEncoder(nn.Module):
             )
         return states
 
-        attn_caches = [
-            [
-                torch.zeros(self.memory_size, self.d_model, device=device),
-                torch.zeros(self.left_context_length, self.d_model, device=device),
-                torch.zeros(self.left_context_length, self.d_model, device=device),
-            ]
-            for _ in range(self.num_encoder_layers)
-        ]
-        conv_caches = [
-            torch.zeros(self.d_model, self.cnn_module_kernel - 1, device=device)
-            for _ in range(self.num_encoder_layers)
-        ]
-        states: Tuple[List[List[torch.Tensor]], List[torch.Tensor]] = (
-            attn_caches,
-            conv_caches,
-        )
-        return states
-
 
 class Emformer(EncoderInterface):
     def __init__(

egs/librispeech/ASR/pruned_transducer_stateless7_ctc/ctc_decode.py

@@ -44,7 +44,7 @@ Usage:
     --exp-dir ./pruned_transducer_stateless7_ctc/exp \
     --max-duration 600 \
     --hlg-scale 0.8 \
-    --decoding-method 1best
+    --decoding-method nbest
 
 (4) nbest-rescoring
 ./pruned_transducer_stateless7_ctc/ctc_decode.py \

egs/librispeech/ASR/pruned_transducer_stateless7_ctc_bs/ctc_decode.py

@@ -42,7 +42,7 @@ Usage:
     --exp-dir ./pruned_transducer_stateless7_ctc_bs/exp \
     --max-duration 600 \
     --hlg-scale 0.8 \
-    --decoding-method 1best
+    --decoding-method nbest
 (4) nbest-rescoring
 ./pruned_transducer_stateless7_ctc_bs/ctc_decode.py \
     --epoch 30 \