參考以下步驟完成加分作業,加總分(2分)
錄製或剪裁聲音檔案,大約15秒的聲音檔案(可以使用線上工具):YoutubeToMP3、Mp3Cut
登入Kaggle(https://www.kaggle.com/) ,並使用這個連結 https://www.kaggle.com/code/a24998667/breezyvoice-playground
點選 copy and edit
在右邊 Notebook -> upload 上傳剪裁好的聲音檔案,設定標題
上傳完成後,複製這個聲音檔的連結
修改倒數第三個程式碼區塊的內容(紅色粗體的部分)
!python3 single_inference.py \
--speaker_prompt_audio_path "/kaggle/input/ptcounty/15sec.mp3" \
--speaker_prompt_text_transcription "" \
--content_to_synthesize "枋 寮 高 中 的同學大家好 , 我是周春米縣長 , 枋 寮 高 中 是我們屏東高中第一首府 , 枋 寮 高 中 的學生每天都在做火箭 , 真的很辛苦,所以我特別來跟大家勉勵 , 希望每個人做的火箭都會飛 , 會飛的都會飛很遠" \
--output_path results/out.wav 2>/dev/null
回到最上面的設定列 點選 settings -> Accelerator -> GPU 100
Run -> Run All
最後等待完成
Record or trim an audio file, approximately 15 seconds long (you can use online tools): YoutubeToMP3, Mp3Cut.
Log in to Kaggle (https://www.kaggle.com/) and use this link: https://www.kaggle.com/code/a24998667/breezyvoice-playground.
Click "copy and edit."
Upload the trimmed audio file in Notebook -> Upload on the right, and set a title.
After uploading, copy the audio file's link. Modify the content of the third-to-last code block (the part in bold red).
!python3 single_inference.py \
--speaker_prompt_audio_path "/kaggle/input/ptcounty/15sec.mp3" \
--speaker_prompt_text_transcription "" \
--content_to_synthesize "枋 寮 高 中 的同學大家好 , 我是周春米縣長 , 枋 寮 高 中 是我們屏東高中第一首府 , 枋 寮 高 中 的學生每天都在做火箭 , 真的很辛苦,所以我特別來跟大家勉勵 , 希望每個人做的火箭都會飛 , 會飛的都會飛很遠" \
--output_path results/out.wav 2>/dev/null
Return to the top settings menu, select settings -> Accelerator -> GPU 100, then run -> Run All.
Finally, wait for it to complete.
BreezyVoice is a voice-cloning text-to-speech system specifically adapted for Taiwanese Mandarin, highlighting phonetic control abilities via auxiliary 注音 inputs.
About BreezyVoice (MediaTek-Traditonal Chinese)
Try in Huggingface https://huggingface.co/spaces/Splend1dchan/BreezyVoice-Playground
Take out ur mobile phone and open voice recoder(or other tool...),please record a voice about 15 seconds.
then attach the voice file to ur email....
Login to ur email and dwonload the file
Conver the voice file to mp3 https://online-audio-converter.com/tw/ and download the file
You depoly BreezyVoice in ur computer. But prepare a computer with a Grapic card for example RTC 4060 or higher level...
Another way to depoly BreezyVoice ....kaggle https://www.kaggle.com/
First, Regist with ur gmail account.
After login kaggle , please finsh ur id veritfy. Prepare ur mobile phone and u'll recieve a code(four number).
Setting -> Phone verification
click this link https://www.kaggle.com/code/a24998667/breezyvoice-playground
click Copy & Edit, then it will open another tab/page.
click "Upload"
click "New datasheet"
upload ur voice file
Your dataset was created successfully.
paste the path link
content_to _synthesize 打上你的文字,接下來會製作語音
go back to Top, click run-> run all
Click Here.
Congradulation , u can use GPU in kaggle. Click "Copy and edit" and follow my
follow my step,plz.
Napkin https://www.napkin.ai/
Words to Flow, table,graph...etc
音訊、影片檔案可以轉逐字稿(只有英文),可以定義講者名字,改寫逐字稿內容,重新再輸出音訊
MultiTalk in Hugginface https://huggingface.co/spaces/fffiloni/Meigen-MultiTalk
Nano Banana pro