This is edited (and corrected version) for installation of Spark-TTS for Windows

[AcTePuKc]

There were several issues, but mainly misleading explanation and the way the files get download - if it crash it didn't resume, also you did not mention installation of Cuda/Pytorch at all.

Spark-TTS Installation (Windows Guide)

1. Install Conda (if you haven’t already)

• Download Miniconda and install it.
• Make sure to check "Add Conda to PATH" during installation.

Download Spark-TTS

You have two options to get the files:

Option 1 (Recommended for Windows): Download ZIP manually

• Go to Spark-TTS GitHub
• Click "Code" > "Download ZIP", then extract it.

Option 2: Use Git (Optional)

• If you prefer using Git, install Git and run:

  git clone https://github.com/SparkAudio/Spark-TTS.git

---

2. Create a Conda Environment

Open Command Prompt (cmd) and run:

conda create -n sparktts python=3.12 -y
conda activate sparktts

This creates and activates a Python 3.12 environment for Spark-TTS.

---

3. Install Dependencies

Inside the Spark-TTS folder (whether from ZIP or Git), run:

pip install -r requirements.txt

---

4. Install PyTorch (Auto-Detect CUDA or CPU)

pip install torch torchvision torchaudio --index-url https://pytorch.org/get-started/previous-versions/

# OR Manually install a specific CUDA version (if needed)
# pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
# pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118  # Older GPUs

---

5. Download the Model

There are two ways to get the model files. Pick one:

Option 1 (Recommended): Using Python
Create a new file in the Spark-TTS folder called download_model.py, paste this inside, and run it:

from huggingface_hub import snapshot_download
import os

# Set download path
model_dir = "pretrained_models/Spark-TTS-0.5B"

# Check if model already exists
if os.path.exists(model_dir) and len(os.listdir(model_dir)) > 0:
    print("Model files already exist. Skipping download.")
else:
    print("Downloading model files...")
    snapshot_download(
        repo_id="SparkAudio/Spark-TTS-0.5B",
        local_dir=model_dir,
        resume_download=True  # Resumes partial downloads
    )
    print("Download complete!")

Run it with:

python download_model.py

✅ Option 2: Using Git (If You Installed It)

mkdir pretrained_models
git clone https://huggingface.co/SparkAudio/Spark-TTS-0.5B pretrained_models/Spark-TTS-0.5B

Either method works—choose whichever is easier for you.

---

6. Run Spark-TTS

Web UI (Recommended)

For an interactive browser-based interface, run:

python webui.py

This launches a local web server where you can enter text and generate speech or clone a voice.

---

7. Troubleshooting & Common Questions

🔎 Before Asking for Help
Many common issues are already covered in existing discussions, documentation, or online resources. Please:

• Search GitHub issues first 🕵️‍♂️
• Check the documentation 📖
• Google or use AI tools (ChatGPT, DeepSeek, etc.)

If you still need help, please explain what you’ve already tried so we can assist you better!

---

Now you’re good to go! 🚀🔥

Happy TTS-ing.

[xinshengwang]

@AcTePuKc Thanks for pointing out these issues and sharing such detailed instructions! We’ve revised the instructions and updated the README to link to this Windows installation guide. If you have any other questions or suggestions, feel free to let us know. Your feedback is always appreciated!

[AcTePuKc]

Writing it in few minutes will post GUI and CLI examples for lazy people like me :D

[fragsun]

followed the documentation step by step for deployment, but encountered several errors when executing webui.py.

1. Initially, it prompted:

   OMP: Error #15: Initializing libiomp5md.dll, but found libiomp5md.dll already initialized.
   

   After deleting the \Library\bin\libiomp5md.dll file, the program continued to execute.

2. Then I encountered the following error:

   Traceback (most recent call last): 
       File "D:\python practice\Spark-TTS\webui.py", line 250, in <module>  
       demo.launch(  
       File "C:\Users\fragsun\.conda\envs\sparktts\Lib\site-packages\gradio\blocks.py", line 2674, in launch  
       raise ValueError(  
       ValueError: When localhost is not accessible, a shareable link must be created. Please set share=True or check your proxy settings to allow access to localhost.  
   

   After modifying the launch() function in webui.py, the error disappeared.

3. However, a warning still appears:

   \.conda\envs\sparktts\Lib\site-packages\torch\nn\utils\weight_norm.py:143: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.  
       WeightNorm.apply(module, name, dim)
       Missing tensor: mel_transformer.spectrogram.window
       Missing tensor: mel_transformer.mel_scale.fb
       信息: 用提供的模式无法找到文件。
   

   Research indicates this issue is beyond my current understanding, as I have limited knowledge of programming code. Currently, I can access the local SparkTTS page, but cannot generate voice files.

How can I resolve this problem? Is it possible to fix this via a Git update?

[AcTePuKc]

Using webui.py requires a lot of VRAM, so the high memory usage is expected. Some of the messages you're seeing are just warnings, not errors—so the app should still be able to generate audio.

If you have a 3090 or a 40-series GPU, you can continue using the web UI. However, if you’re running into performance issues, you might want to try the CLI version or the alternative UI I’ve built for better memory efficiency.

• SparkTTS (CLI) → Uses around 4-5 GB VRAM (on my 3060).
• SparkTTS (GUI) → Uses 8-12 GB VRAM, depending on settings.
• SparkTTS (WebUI) → Uses 8-18 GB VRAM (as much as you have).

💡 Most of the warnings (like torch.nn.utils.weight_norm is deprecated) do not affect functionality and can be ignored.

🔹 Fix for libiomp5md.dll issue: This is not a SparkTTS-related issue but a general library conflict. Instead of deleting the file, set this environment variable to allow multiple OpenMP libraries to load:

set KMP_DUPLICATE_LIB_OK=TRUE

This should prevent any OpenMP conflicts without breaking dependencies.

Note: Issues related to missing or conflicting libraries (libiomp5md.dll, torch, gradio, etc.) are not specific to SparkTTS but are due to your system setup or dependencies. If audio generation still doesn’t work after applying the fix, the best approach is to delete the conda environment and reinstall everything from scratch.

[jackqdldd]

RuntimeError: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from http://www.nvidia.com/Download/index.aspx

Must with an NVIDIA GPU?

[xzw168]

这是我用过效果最好的tts，p106-100占用5.5GB VRAM。合成10秒需要40秒，要是合成速度能优化下就完美了。

[deedcao]

When I run the command "pip install -r requirements.txt", the erros as below

please check

[A-I-Dev]

Any support for AMD cards, because I hit the

    raise AssertionError("Torch not compiled with CUDA enabled")
AssertionError: Torch not compiled with CUDA enabled

wall.

[Idkanythingfuck]

line 18, in <module> import soundfile as sf ModuleNotFoundError: No module named 'soundfile'

help!! i am really bad at it

[keyserjaya]

@deedcao Have you used python 3.12?
if yes... try:

python -m pip upgrade
pip install scipy
pip install -r requirements.txt

*I'm not an expert, but i think it will fix your problem.

[yao-charlie]

Some issues...

On step 4, running
pip install torch torchvision torchaudio --index-url https://pytorch.org/get-started/previous-versions/

Yields:

Requirement already satisfied: torch in c:\users\charl\anaconda3\envs\sparktts\lib\site-packages (2.5.1)
ERROR: Could not find a version that satisfies the requirement torchvision (from versions: none)
ERROR: No matching distribution found for torchvision```

Ok, fine, go direct with:

```pip install torch torchvision torchaudio```

Ok, installs but when running:

```C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\nn\utils\weight_norm.py:143: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
  WeightNorm.apply(module, name, dim)
Missing tensor: mel_transformer.spectrogram.window
Missing tensor: mel_transformer.mel_scale.fb
Traceback (most recent call last):
  File "C:\Users\charl\Github\Spark-TTS\webui.py", line 244, in <module>
    demo = build_ui(
           ^^^^^^^^^
  File "C:\Users\charl\Github\Spark-TTS\webui.py", line 81, in build_ui
    model = initialize_model(model_dir, device=device)
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\charl\Github\Spark-TTS\webui.py", line 31, in initialize_model
    model = SparkTTS(model_dir, device)
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\charl\Github\Spark-TTS\cli\SparkTTS.py", line 44, in __init__
    self._initialize_inference()
  File "C:\Users\charl\Github\Spark-TTS\cli\SparkTTS.py", line 50, in _initialize_inference
    self.audio_tokenizer = BiCodecTokenizer(self.model_dir, device=self.device)
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\charl\Github\Spark-TTS\sparktts\models\audio_tokenizer.py", line 42, in __init__
    self._initialize_model()
  File "C:\Users\charl\Github\Spark-TTS\sparktts\models\audio_tokenizer.py", line 46, in _initialize_model
    self.model = BiCodec.load_from_checkpoint(f"{self.model_dir}/BiCodec").to(
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\nn\modules\module.py", line 1343, in to
    return self._apply(convert)
           ^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\nn\modules\module.py", line 903, in _apply
    module._apply(fn)
  File "C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\nn\modules\module.py", line 903, in _apply
    module._apply(fn)
  File "C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\nn\modules\module.py", line 903, in _apply
    module._apply(fn)
  File "C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\nn\modules\module.py", line 930, in _apply
    param_applied = fn(param)
                    ^^^^^^^^^
  File "C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\nn\modules\module.py", line 1329, in convert
    return t.to(
           ^^^^^
  File "C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\cuda\__init__.py", line 310, in _lazy_init
    raise AssertionError("Torch not compiled with CUDA enabled")
AssertionError: Torch not compiled with CUDA enabled

Alright, specify an older install with CUDA:

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 # Older GPUs

Ok, no CUDA error. Running still yields:

C:\Users\charl\anaconda3\envs\sparktts\Lib\site-packages\torch\nn\utils\weight_norm.py:143: FutureWarning: 'torch.nn.utils.weight_norm' is deprecated in favor of 'torch.nn.utils.parametrizations.weight_norm'.
  WeightNorm.apply(module, name, dim)
Missing tensor: mel_transformer.spectrogram.window
Missing tensor: mel_transformer.mel_scale.fb

Running on local URL:  http://0.0.0.0:7860

And on following the link"

This site can’t be reached

Haven't been able to find something reliable re: missing tensors as noted yet but thought I'd comment with progress.

[heldenby]

Although the startup says:
Running on local URL: http://0.0.0.0:7860

I think the URL should be http://127.0.0.1:7860/

Give it a try.

[yao-charlie]

Although the startup says: Running on local URL: http://0.0.0.0:7860

I think the URL should be http://127.0.0.1:7860/

Give it a try.

Lol that would make way more sense... seems to work after all. Despite how embarrassing my comment is to not check localhost, I'll leave it up for anyone else who might need some troubleshooting.

[deedcao]

@deedcao Have you used python 3.12? if yes... try:

python -m pip upgrade
pip install scipy
pip install -r requirements.txt

*I'm not an expert, but i think it will fix your problem.

Hi @keyserjaya
Thanks for your support. I've checked my python, the version is 3.12