add IPXE Support

.github/workflows/ci-shell.yml

@@ -6,9 +6,11 @@ on:
   push:
     branches:
       - master
+      - intel-gpu
   pull_request:
     branches:
       - master
+      - intel-gpu
 
 jobs:
   test-shell-script:

.github/workflows/ci.yml

@@ -6,9 +6,11 @@ on:
   push:
     branches:
       - master
+      - intel-gpu
   pull_request:
     branches:
       - master
+      - intel-gpu
 
 jobs:
   build:

.gitignore

@@ -10,4 +10,4 @@ outputs/
 modules/__pycache__/
 models/
 modules/yt_tmp.wav
-configs/default_parameters.yaml
+configs/default_parameters.yaml

README.md

@@ -34,13 +34,13 @@ If you wish to try this on Colab, you can do it in [here](https://colab.research
 # Installation and Running
 ### Prerequisite
 To run this WebUI, you need to have `git`, `python` version 3.8 ~ 3.10, `FFmpeg`. <br>
-And if you're not using an Nvida GPU, or using a different `CUDA` version than 12.4,  edit the [`requirements.txt`](https://github.com/jhj0517/Whisper-WebUI/blob/master/requirements.txt) to match your environment.
+And if you're using an Intel GPU,  view the [`requirements-ipxe.txt`](https://github.com/jhj0517/Whisper-WebUI/blob/master/requirements.txt) to match your environment.
 
 Please follow the links below to install the necessary software:
 - git : [https://git-scm.com/downloads](https://git-scm.com/downloads)
 - python : [https://www.python.org/downloads/](https://www.python.org/downloads/) **( If your python version is too new, torch will not install properly.)**
 - FFmpeg :  [https://ffmpeg.org/download.html](https://ffmpeg.org/download.html)
-- CUDA : [https://developer.nvidia.com/cuda-downloads](https://developer.nvidia.com/cuda-downloads)
+
 
 After installing FFmpeg, **make sure to add the `FFmpeg/bin` folder to your system PATH!**
 

modules/diarize/diarizer.py

@@ -1,5 +1,11 @@
 import os
 import torch
+try:
+    import intel_extension_for_pytorch as ipex
+    if torch.xpu.is_available():
+        xpu_available = True
+except:
+    pass
 from typing import List, Union, BinaryIO, Optional
 import numpy as np
 import time
@@ -120,6 +126,8 @@ def get_device():
             return "cuda"
         elif torch.backends.mps.is_available():
             return "mps"
+        elif torch.xpu.is_available():
+            return "xpu"
         else:
             return "cpu"
 
@@ -130,4 +138,6 @@ def get_available_device():
             devices.append("cuda")
         elif torch.backends.mps.is_available():
             devices.append("mps")
-        return devices
+        elif torch.xpu.is_available():
+            devices.append("xpu")
+        return devices

modules/translation/translation_base.py

@@ -1,5 +1,11 @@
 import os
 import torch
+try:
+    import intel_extension_for_pytorch as ipex
+    if torch.xpu.is_available():
+        xpu_available = True
+except:
+    pass
 import gradio as gr
 from abc import ABC, abstractmethod
 from typing import List
@@ -134,22 +140,27 @@ def translate_file(self,
         except Exception as e:
             print(f"Error: {str(e)}")
         finally:
-            self.release_cuda_memory()
+            self.release_gpu_memory()
 
     @staticmethod
     def get_device():
         if torch.cuda.is_available():
             return "cuda"
         elif torch.backends.mps.is_available():
             return "mps"
+        elif torch.xpu.is_available():
+            return "xpu"
         else:
             return "cpu"
 
     @staticmethod
-    def release_cuda_memory():
+    def release_gpu_memory():
         if torch.cuda.is_available():
             torch.cuda.empty_cache()
             torch.cuda.reset_max_memory_allocated()
+        elif torch.xpu.is_available():
+            torch.xpu.empty_cache()
+            torch.xpu.reset_peak_memory_stats()
 
     @staticmethod
     def remove_input_files(file_paths: List[str]):

modules/whisper/whisper_base.py

@@ -1,5 +1,11 @@
 import os
 import torch
+try:
+    import intel_extension_for_pytorch as ipex
+    if torch.xpu.is_available():
+        xpu_available = True
+except:
+    pass
 import whisper
 import gradio as gr
 import torchaudio
@@ -250,7 +256,7 @@ def transcribe_file(self,
         except Exception as e:
             print(f"Error transcribing file: {e}")
         finally:
-            self.release_cuda_memory()
+            self.release_gpu_memory()
 
     def transcribe_mic(self,
                        mic_audio: str,
@@ -305,7 +311,7 @@ def transcribe_mic(self,
         except Exception as e:
             print(f"Error transcribing file: {e}")
         finally:
-            self.release_cuda_memory()
+            self.release_gpu_memory()
 
     def transcribe_youtube(self,
                            youtube_link: str,
@@ -369,7 +375,7 @@ def transcribe_youtube(self,
         except Exception as e:
             print(f"Error transcribing file: {e}")
         finally:
-            self.release_cuda_memory()
+            self.release_gpu_memory()
 
     @staticmethod
     def generate_and_write_file(file_name: str,
@@ -459,6 +465,8 @@ def get_device():
                 # Device `SparseMPS` is not supported for now. See : https://github.com/pytorch/pytorch/issues/87886
                 return "cpu"
             return "mps"
+        elif torch.xpu.is_available():
+            return "xpu"
         else:
             return "cpu"
 
@@ -480,11 +488,14 @@ def is_sparse_api_supported():
             return False
 
     @staticmethod
-    def release_cuda_memory():
+    def release_gpu_memory():
         """Release memory"""
         if torch.cuda.is_available():
             torch.cuda.empty_cache()
             torch.cuda.reset_max_memory_allocated()
+        elif torch.xpu.is_available():
+            torch.xpu.empty_cache()
+            torch.xpu.reset_peak_memory_stats()
 
     @staticmethod
     def remove_input_files(file_paths: List[str]):

requirements-ipex.txt

@@ -0,0 +1,11 @@
+#torch @ https://github.com/Nuullll/intel-extension-for-pytorch/releases/download/v2.1.10%2Bxpu/torch-2.1.0a0+cxx11.abi-cp310-cp310-win_amd64.whl
+#torchvision @ https://github.com/Nuullll/intel-extension-for-pytorch/releases/download/v2.1.10%2Bxpu/torchvision-0.16.0a0+cxx11.abi-cp310-cp310-win_amd64.whl
+#torchaudio @ https://github.com/Nuullll/intel-extension-for-pytorch/releases/download/v2.1.10%2Bxpu/torchaudio-2.1.0a0+cxx11.abi-cp310-cp310-win_amd64.whl
+#intel_extension_for_pytorch @ https://github.com/Nuullll/intel-extension-for-pytorch/releases/download/v2.1.10%2Bxpu/intel_extension_for_pytorch-2.1.10+xpu-cp310-cp310-win_amd64.whl
+
+git+https://github.com/jhj0517/jhj0517-whisper.git
+faster-whisper==1.0.3
+transformers==4.42.3
+gradio==4.29.0
+pytubefix
+pyannote.audio==3.1.1

requirements.txt

@@ -2,7 +2,7 @@
 # If you're using it, update url to your CUDA version (CUDA 12.1 is minimum requirement):
 # For CUDA 12.1, use : https://download.pytorch.org/whl/cu121
 # For CUDA 12.4, use : https://download.pytorch.org/whl/cu124
---extra-index-url https://download.pytorch.org/whl/cu121
+#--extra-index-url https://download.pytorch.org/whl/cu121
 
 
 torch==2.3.1

start-webui-ipex.bat

@@ -0,0 +1,7 @@
+@echo off
+
+call venv\scripts\activate
+python app.py  --whisper_type whisper
+
+echo "launching the app"
+pause