From 89a8492178ce163c4f47884ddebf4785e3ae2f53 Mon Sep 17 00:00:00 2001
From: wizardeur <57131776+wizardeur@users.noreply.github.com>
Date: Tue, 27 May 2025 12:15:52 +0200
Subject: [PATCH] Update README.md with information about ROCm support

---
 README.md | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/README.md b/README.md
index 867c46e..6c6bf4d 100644
--- a/README.md
+++ b/README.md
@@ -101,7 +101,7 @@ Orpheus-FastAPI/
 ### Prerequisites
 
 - Python 3.8-3.11 (Python 3.12 is not supported due to removal of pkgutil.ImpImporter)
-- CUDA-compatible GPU (recommended: RTX series for best performance)
+- CUDA-compatible or ROCm-compatible GPU (recommended: RTX series for best performance)
 - Using docker compose or separate LLM inference server running the Orpheus model (e.g., LM Studio or llama.cpp server)
 
 ### 🐳 Docker compose
@@ -122,11 +122,16 @@ ORPHEUS_MODEL_NAME=Orpheus-3b-French-FT-Q8_0.gguf  # Example for French
 
 Then start the services:
 
-For GPU support run
+For CUDA GPU support run
 ```bash
 docker compose -f docker-compose-gpu.yml up
 ```
 
+For ROCm GPU support run
+```bash
+docker compose -f docker-compose-gpu-rocm.yml up
+```
+
 For CPU support run:
 ```bash
 docker compose -f docker-compose-cpu.yml up
@@ -157,6 +162,11 @@ conda activate orpheus-tts
 ```bash
 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu124
 ```
+or
+Install PyTorch with ROCm support:
+```bash
+pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.4/
+```
 
 4. Install other dependencies:
 ```bash