GPT-SoVITS
1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
Runssource: GitHubPythonMITcommit 48b1a0169a28
Python, MIT licensed. The project labels itself: text to speech, tts, vits, voice clone, voice cloneai and voice cloning.
GPT-SoVITS runs. An Argusic agent installed it in 35 minutes and hit 4 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
GPT-SoVITS API server starts and responds to TTS inference requests: HTTP 200 with a valid synthesized WAV audio file returned for English text-to-speech using a default reference audio on CPU.
What the agent ended up with on a clean machine, in its own words. How this is measured
Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 100 of 100 (the mean of this project's run scores).
At a glance
Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.
What was measured
From the run that decided this verdict. Every line is a recorded value, not an estimate.
What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- pyopenjtalk failed - Python.h not found8 minutes
- jieba_fast failed - Python.h not found3 minutes
- opencc build from source failed (missing C dependencies)2 minutes
- NLTK averaged_perceptron_tagger_eng missing1 minute
- Install time
- 35 minutes
- Cold machine to finish
- 23 minutes
- Errors hit and fixed
- 4 hit, 4 fixed with no human help
- How the result was proved
- curl -s -m 60 'http://127.0.0.1:9886/?text=hello+world&text_language=en' -o /tmp/output.wav -w '%{http_code}' returned 200; output was a valid 137KB, 2.14s WAV file verified by ffprobe and ffmpeg
- Model tokens used
- 840,898
- Exact commit tested
- 48b1a0169a28
- Container image
- sha256:cdd920bce783
What was not measured
The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.
- Peak GPU memory
- this ran on a CPU machine, so there was no GPU to sample
- Behaviour over a long run
- each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
launch_succeeded: Yes, the API server started and responded to requests install_succeeded: YesHere is the final summary of everything I observed and verified: **Install**: Created a Python 3.12 venv, installed PyTorch CPU, and all 80+ pip packages from `requ… **Dependencies verified**: torch, opencc, pyopenjtalk, jieba_fast, jieba, gradio, fastapi, onnxrunt… **Pretrained models**: Downloaded from HuggingFace (chinese-hubert-base, chinese-roberta-wwm-ext-la… **G2PW ONNX inference**: `G2PWPinyin.pinyin(['你好', '世界'])` returned `[['nǐ'], ['hǎo'], ['shì'], ['j… **API server**: Started on port 9886, accepted HTTP requests, performed full TTS inference end-to-e… **Errors encountered**: - opencc build from source failed → fixed via `opencc-1.4.2` prebuilt wheel - pyopenjtalk failed due to missing `Python.h` → extracted and installed `libpython3.12-dev` header… - jieba_fast failed due to same missing headers → same fix - NLTK `averaged_perceptron_tagger_eng` not found → downloaded via `nltk.download()`
Replay the whole session, every command from a clean machine to this point.
How it was tested
One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.
Strengths and limits
Measured facts, not opinions. How this is written.
What went well
- Reached a running state on a clean machine, with the session recorded.
- Installed in 35 minutes.
- Recovered from all 4 errors without a human stepping in, which says the failures are documented well enough to solve.
- Ran without a GPU, so it does not need one to start.
- Licensed MIT, as reported by its host.
What did not, or is not known
- Measured once so far. A single attempt shows it can run, not that it runs reliably.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
Also tested, in the same area
Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.
Run history
Topics (from GitHub)
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/gpt-sovits)Questions
- Does GPT-SoVITS run?
- Yes. GPT-SoVITS runs. Argusic installed and launched it on a clean machine in 35 minutes, hitting 4 errors on the way, and recorded the session.
- How did Argusic test GPT-SoVITS?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 48b1a0169a28. One attempt is recorded, and the full method is on the methodology page.
- What did it cost to test GPT-SoVITS?
- The run that produced this verdict cost $0.16: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does GPT-SoVITS take to install?
- 35 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
- Does GPT-SoVITS need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing GPT-SoVITS?
- 4 things broke in the recorded run, and 4 were fixed without human help. Each one, and the time it cost, is listed on this page.
- Where is the evidence for GPT-SoVITS?
- The recorded run is on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.