From 8e54a42db2690e0ac7f09fd400b29ad2fb2c9c5a Mon Sep 17 00:00:00 2001 From: historia Date: Mon, 24 Aug 2026 03:08:29 -0400 Subject: chore: update backend docs with custom venv --- app/docs/backend-qwen.md | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'app/docs/backend-qwen.md') diff --git a/app/docs/backend-qwen.md b/app/docs/backend-qwen.md index 074004c..af5f63e 100644 --- a/app/docs/backend-qwen.md +++ b/app/docs/backend-qwen.md @@ -7,7 +7,8 @@ If you prefer to install the backend yourself (in your own environment, not the Install qwen-tts with pip into your environment: ```bash -conda activate audiobook +python -m venv audiobook +source audiobook/bin/activate pip install -U qwen-tts ``` @@ -32,7 +33,7 @@ Whisper (`faster_whisper` or `whisper`) is used automatically to transcribe the ## Custom voice (i.e. built-in voice) ```bash -conda activate audiobook +source audiobook/bin/activate qwen-tts-demo Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice --ip 127.0.0.1 --port 7860 [--no-flash-attn] ``` @@ -51,7 +52,7 @@ FlashAttention provides a *small* speed boost on the `qwen` backend. It is **not 1. Build from source (takes absolutely forever). If you run out of memory, lower MAX_JOBS until you don't. ```bash -conda activate audiobook +source audiobook/bin/activate pip install ninja packaging psutil MAX_JOBS=4 pip install --no-build-isolation flash-attn ``` @@ -59,7 +60,7 @@ MAX_JOBS=4 pip install --no-build-isolation flash-attn 2. pip install a prebuilt wheel matching your torch / CUDA / Python / CXX11-ABI combination: ```bash -conda activate audiobook +source audiobook/bin/activate python -c "import torch; print(torch.__version__, torch.version.cuda, torch._C._GLIBCXX_USE_CXX11_ABI)" ``` -- cgit v1.2.3