From a69e77c0366dd8e8632bef3f5feb25af672af7a8 Mon Sep 17 00:00:00 2001
From: YaoyaoChang <cyy574006791@qq.com>
Date: Mon, 26 Jan 2026 03:26:52 -0800
Subject: [PATCH] 1. unify env for TTS and ASR; 2. avoid transformers 5.0.0
 temporarily

---
 docs/vibevoice-asr.md           |  3 ++-
 docs/vibevoice-realtime-0.5b.md |  3 ++-
 finetuning-asr/README.md        |  4 ++--
 pyproject.toml                  | 28 +++++++---------------------
 4 files changed, 13 insertions(+), 25 deletions(-)

diff --git a/docs/vibevoice-asr.md b/docs/vibevoice-asr.md
index bfc4a7f..ded3b19 100644
--- a/docs/vibevoice-asr.md
+++ b/docs/vibevoice-asr.md
@@ -65,7 +65,8 @@ sudo docker run --privileged --net=host --ipc=host --ulimit memlock=-1:-1 --ulim
 ```bash
 git clone https://github.com/microsoft/VibeVoice.git
 cd VibeVoice
-pip install -e .[asr]
+
+pip install -e .
 ```
 
 ## Usages
diff --git a/docs/vibevoice-realtime-0.5b.md b/docs/vibevoice-realtime-0.5b.md
index f793e96..12f4308 100644
--- a/docs/vibevoice-realtime-0.5b.md
+++ b/docs/vibevoice-realtime-0.5b.md
@@ -97,9 +97,10 @@ sudo docker run --privileged --net=host --ipc=host --ulimit memlock=-1:-1 --ulim
 git clone https://github.com/microsoft/VibeVoice.git
 cd VibeVoice/
 
-pip install -e .[tts]
+pip install -e .
 ```
 
+
 ## Usages
 
 
diff --git a/finetuning-asr/README.md b/finetuning-asr/README.md
index 064b2b5..32768b4 100644
--- a/finetuning-asr/README.md
+++ b/finetuning-asr/README.md
@@ -5,8 +5,8 @@ This directory contains scripts for LoRA (Low-Rank Adaptation) fine-tuning of th
 ## Requirements
 
 ```bash
-# you need to install vibevoice first
-# pip install -e .[asr]
+# Install vibevoice first
+pip install -e .
 
 pip install peft
 ```
diff --git a/pyproject.toml b/pyproject.toml
index 0cc390a..7b0fe4a 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -18,6 +18,7 @@ classifiers = [
 ]
 dependencies = [
     "torch",
+    "transformers>=4.51.3,<5.0.0",
     "accelerate",
     "llvmlite>=0.40.0",
     "numba>=0.57.0",
@@ -29,27 +30,12 @@ dependencies = [
     "ml-collections",
     "absl-py",
     "gradio",
-]
-
-[project.optional-dependencies]
-tts = [
-  "transformers==4.51.3", # we develop this project on transformers==4.51.3, later version may not be compatible
-  "av",
-  "aiortc",
-  "uvicorn[standard]",
-  "fastapi"
-]
-
-asr = [
-  "transformers>=4.51.3", # the versions after 4.51.3 are all support
-  "pydub" # for visualization
-]
-
-vllm = [
-  "transformers>=4.51.3",
-  "fastapi",
-  "uvicorn[standard]",
-  "requests",
+    "av",
+    "aiortc",
+    "uvicorn[standard]",
+    "fastapi",
+    "pydub",
+    "requests",
 ]
 
 [project.entry-points."vllm.general_plugins"]