[zipvoice] Add requirements.txt and pinyin.txt, remove k2 from pretrained model inference. (#1965)

* Add requirements.txt and pinyin.txt needed by zipvoice

* simplify the requirements for pretrained model inference

egs/zipvoice/README.md

@@ -39,15 +39,6 @@ source venv/bin/activate
 * Install the required packages:
 
 ```bash
-# Install pytorch and k2.
-# If you want to use different versions, please refer to https://k2-fsa.org/get-started/k2/ for details.
-# For users in China mainland, please refer to https://k2-fsa.org/zh-CN/get-started/k2/
-
-pip install torch==2.5.1 torchaudio==2.5.1 --index-url https://download.pytorch.org/whl/cu121
-pip install k2==1.24.4.dev20250208+cuda12.1.torch2.5.1 -f https://k2-fsa.github.io/k2/cuda.html
-
-# Install other dependencies.
-pip install piper_phonemize -f https://k2-fsa.github.io/icefall/piper_phonemize.html
 pip install -r requirements.txt
 ```
 
@@ -97,6 +88,16 @@ The following steps show how to train a model from scratch on Emilia and LibriTT
 ### 0. Install dependencies for training
 
 ```bash
+# Install pytorch and k2.
+# If you want to use different versions, please refer to https://k2-fsa.org/get-started/k2/ for details.
+# For users in China mainland, please refer to https://k2-fsa.org/zh-CN/get-started/k2/
+
+# Note: Make sure you have installed the correct version of PyTorch and k2 that matches your CUDA version.
+# For example, if want to use pytorch 2.5.1 and you are using CUDA 12.1, you can install PyTorch and k2 as follows:
+
+pip install torch==2.5.1 torchaudio==2.5.1 --index-url https://download.pytorch.org/whl/cu121
+pip install k2==1.24.4.dev20250208+cuda12.1.torch2.5.1 -f https://k2-fsa.github.io/k2/cuda.html
+
 pip install -r ../../requirements.txt
 ```
 
@@ -403,7 +404,7 @@ on three test sets, i.e., LibriSpeech-PC test-clean, Seed-TTS test-en and Seed-T
 
 ```bibtex
 @article{zhu-2025-zipvoice,
-      title={ZipVoice: Fast and High-Quality Zero-Shot Text-to-Speech with Flow Matching}, 
+      title={ZipVoice: Fast and High-Quality Zero-Shot Text-to-Speech with Flow Matching},
       author={Han Zhu and Wei Kang and Zengwei Yao and Liyong Guo and Fangjun Kuang and Zhaoqing Li and Weiji Zhuang and Long Lin and Daniel Povey}
       journal={arXiv preprint arXiv:2506.13053},
       year={2025},

egs/zipvoice/local/compute_fbank.py

@@ -19,11 +19,13 @@
 import argparse
 import logging
 import os
+from concurrent.futures import ProcessPoolExecutor as Pool
 from pathlib import Path
 from typing import Optional
-from concurrent.futures import ProcessPoolExecutor as Pool
 
+import lhotse
 import torch
+from feature import TorchAudioFbank, TorchAudioFbankConfig
 from lhotse import (
     CutSet,
     LilcomChunkyWriter,
@@ -31,9 +33,6 @@ from lhotse import (
     set_audio_duration_mismatch_tolerance,
 )
 
-from feature import TorchAudioFbank, TorchAudioFbankConfig
-import lhotse
-
 # Torch's multithreaded behavior needs to be disabled or
 # it wastes a lot of CPU and slow things down.
 # Do this outside of main() in case it needs to take effect

egs/zipvoice/local/preprocess_emilia.py

@@ -24,15 +24,14 @@ This file reads the texts in given manifest and save the cleaned new cuts.
 """
 
 import argparse
-import logging
 import glob
+import logging
 import os
+from concurrent.futures import ProcessPoolExecutor as Pool
 from pathlib import Path
 from typing import List
 
 from lhotse import CutSet, load_manifest_lazy
-from concurrent.futures import ProcessPoolExecutor as Pool
-
 from tokenizer import (
     is_alphabet,
     is_chinese,

egs/zipvoice/requirements.txt

@@ -0,0 +1,17 @@
+--find-links https://k2-fsa.github.io/icefall/piper_phonemize.html
+
+torch
+torchaudio
+huggingface_hub
+lhotse
+safetensors
+vocos
+
+# Normalization
+cn2an
+inflect
+
+# Tokenization
+jieba
+piper_phonemize
+pypinyin

egs/zipvoice/zipvoice/scaling.py

@@ -18,9 +18,17 @@
 import logging
 import math
 import random
+import sys
 from typing import Optional, Tuple, Union
 
-import k2
+try:
+    import k2
+except Exception as ex:
+    logging.warning(
+        "k2 is not installed correctly. Swoosh functions will fallback to "
+        "pytorch implementation."
+    )
+
 import torch
 import torch.nn as nn
 from torch import Tensor
@@ -1398,7 +1406,11 @@ class SwooshLFunction(torch.autograd.Function):
 class SwooshL(torch.nn.Module):
     def forward(self, x: Tensor) -> Tensor:
         """Return Swoosh-L activation."""
-        if torch.jit.is_scripting() or torch.jit.is_tracing():
+        if (
+            torch.jit.is_scripting()
+            or torch.jit.is_tracing()
+            or "k2" not in sys.modules
+        ):
             zero = torch.tensor(0.0, dtype=x.dtype, device=x.device)
             return logaddexp(zero, x - 4.0) - 0.08 * x - 0.035
         if not x.requires_grad:
@@ -1472,7 +1484,11 @@ class SwooshRFunction(torch.autograd.Function):
 class SwooshR(torch.nn.Module):
     def forward(self, x: Tensor) -> Tensor:
         """Return Swoosh-R activation."""
-        if torch.jit.is_scripting() or torch.jit.is_tracing():
+        if (
+            torch.jit.is_scripting()
+            or torch.jit.is_tracing()
+            or "k2" not in sys.modules
+        ):
             zero = torch.tensor(0.0, dtype=x.dtype, device=x.device)
             return logaddexp(zero, x - 1.0) - 0.08 * x - 0.313261687
         if not x.requires_grad:
@@ -1636,7 +1652,11 @@ class ActivationDropoutAndLinear(torch.nn.Module):
         self.dropout_shared_dim = dropout_shared_dim
 
     def forward(self, x: Tensor):
-        if torch.jit.is_scripting() or torch.jit.is_tracing():
+        if (
+            torch.jit.is_scripting()
+            or torch.jit.is_tracing()
+            or "k2" not in sys.modules
+        ):
             if self.activation == "SwooshL":
                 x = SwooshLForward(x)
             elif self.activation == "SwooshR":

egs/zipvoice/zipvoice/tokenizer.py

@@ -321,7 +321,8 @@ def tokenize_ZH(text: str) -> List[str]:
             if final != "":
                 phones.append(final)
         return phones
-    except:
+    except Exception as ex:
+        logging.warning(f"Tokenize ZH failed: {ex}")
         return []
 
 
@@ -332,7 +333,8 @@ def tokenize_EN(text: str) -> List[str]:
         tokens = phonemize_espeak(text, "en-us")
         tokens = reduce(lambda x, y: x + y, tokens)
         return tokens
-    except:
+    except Exception as ex:
+        logging.warning(f"Tokenize EN failed: {ex}")
         return []
 
 
@@ -561,7 +563,7 @@ class TokenizerLibriTTS(object):
 
 if __name__ == "__main__":
     text = "我们是5年小米人,是吗? Yes I think so! mr king, 5 years, from 2019 to 2024. 霍...啦啦啦超过90%的人咯...?!9204"
-    tokenizer = Tokenizer()
+    tokenizer = TokenizerEmilia()
     tokens = tokenizer.texts_to_tokens([text])
     print(f"tokens : {tokens}")
     tokens2 = "|".join(tokens[0])