git clone https://github.com/pyenv-win/pyenv-win.git "$HOME\.pyenv"

[System.Environment]::SetEnvironmentVariable('PYENV',$env:USERPROFILE + "\.pyenv\pyenv-win\","User")

[System.Environment]::SetEnvironmentVariable('PYENV_ROOT',$env:USERPROFILE + "\.pyenv\pyenv-win\","User")

[System.Environment]::SetEnvironmentVariable('PYENV_HOME',$env:USERPROFILE + "\.pyenv\pyenv-win\","User")

[System.Environment]::SetEnvironmentVariable('path', $env:USERPROFILE + "\.pyenv\pyenv-win\bin;" + $env:USERPROFILE + "\.pyenv\pyenv-win\shims;" + [System.Environment]::GetEnvironmentVariable('path', "User"),"User")

pyenv install -l
pyenv install 3.10.6
pyenv global 3.10.6

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git

webui-user.bat

git clone https://github.com/lllyasviel/stable-diffusion-webui-forge.git

git clone https://github.com/reriiasu/speech-to-text.git
cd speech-to-text

python3 -m venv venv
.\venv\Scripts\activate.bat
pip install -r requirements.txt

python -m speech_to_text

git clone https://github.com/litagin02/Style-Bert-VITS2.git
cd Style-Bert-VITS2
python3 -m venv venv
venv\Scripts\activate
pip install torch torchaudio --index-url https://download.pytorch.org/whl/cu118
pip install -r requirements.txt
python initialize.py

python server_editor.py --inbrowser

python server_fastapi.py

python app.py

venv\Scripts\activate
git clone https://github.com/reazon-research/ReazonSpeech
pip install setuptools
pip install Cython
pip install ReazonSpeech/pkg/nemo-asr

import os
from flask import Flask, request, jsonify, make_response
from flask_cors import CORS
import tempfile
import logging
import json
from reazonspeech.nemo.asr import load_model, transcribe, audio_from_path

app = Flask(__name__)
CORS(app)  # どこからでもリクエストを受け付けるようにCORSを設定

# ログ設定
logging.basicConfig(level=logging.INFO)

# reazonspeech
model = load_model()

@app.route('/transcribe', methods=['POST'])
def handle():
    if 'audio' not in request.files:
        return jsonify({'error': 'No audio file provided'}), 400
    
    audio_file = request.files['audio']
    
    # ファイルを一時的に保存
    temp_dir = tempfile.gettempdir()
    audio_path = os.path.join(temp_dir, audio_file.filename)
    audio_file.save(audio_path)

    # 文字起こし前のログ
    logging.info(f"Received audio file: {audio_file.filename}")
    
    # reazonspeech
    audio = audio_from_path(audio_path)
    transcription_text = transcribe(model, audio)
    
    # 一時ファイルを削除
    os.remove(audio_path)

    # 文字起こし後のログ
    logging.info("Transcription completed.")
    logging.info(f"Transcription result: {transcription_text.text}")
    
    # 結果をJSON形式で返す
    response = make_response(json.dumps({'transcription': transcription_text.text}, ensure_ascii=False))
    response.mimetype = "application/json"
    return response

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5001)

 // 参考: https://zenn.dev/himanushi/articles/99579cf407c30b

const callDifyByStreaming = async (input) => {
  const completion = await fetch(DIFY_URL, {
    headers: {
      "Content-Type": "application/json",
      Authorization: AUTHORIZATION_HEADER,
    },
    method: "POST",
    body: JSON.stringify({
      inputs: input,
      response_mode: "streaming",
      user: USER,
    }),
  });

  const reader = completion.body?.getReader();

  if (completion.status !== 200 || !reader) {
    return "error";
  }

  let result = "";

  const decoder = new TextDecoder("utf-8");
  try {
    let texts = "";
    let allOutput = [];
    const read = async () => {
      const { done, value } = await reader.read();
      if (done) return reader.releaseLock();
      const chunk = decoder.decode(value, { stream: true });

      texts += chunk;
      const json = texts
        .split("data:")
        .map((data) => {
          if (data) {
            try {
              const event = JSON.parse(data);
              allOutput.push(event);
              fs.writeFileSync("outputs.json", JSON.stringify(allOutput));
              if (event.event === "workflow_finished") {
                const result = event.data;
                return result.outputs;
              }
            } catch (e) {
              texts = `data:${data}`;
            }
          }
          return "";
        })
        .filter((data) => data)[0];
      if (json) {
        result = json;
      }
      return read();
    };
    await read();
  } catch (e) {
    console.error(e);
  }
  reader.releaseLock();

  return result;
};

async def callDifyByStreaming(input):
    timeout = aiohttp.ClientTimeout(total=None)
    async with aiohttp.ClientSession(timeout=timeout) as session:
        async with session.post(
            DIFY_URL,
            headers={
                "Content-Type": "application/json",
                "Authorization": AUTHORIZATION_HEADER,
            },
            json={
                "inputs": input,
                "response_mode": "streaming",
                "user": USER,
            }
        ) as response:
            if response.status != 200:
                return "error"

            result = ""
            try:
                texts = ""
                allOutput = []
                async for value in response.content.iter_chunked(1024):
                    chunk = value.decode('utf-8')
                    texts += chunk
                    parts = texts.split('data:')
                    texts = ''
                    for i, data in enumerate(parts):
                        data = data.strip()
                        if not data:
                            continue
                        try:
                            event = json.loads(data)
                            allOutput.append(event)
                            with open("outputs.json", "w", encoding="utf-8") as f:
                                json.dump(allOutput, f, ensure_ascii=False)
                            if event.get('event') == 'workflow_finished':
                                result = event['data']['outputs']
                                return result
                        except json.JSONDecodeError:
                            # JSONのパースに失敗した場合、未処理のデータをテキストに戻す
                            texts = 'data:' + 'data:'.join(parts[i:])
                            break  # 次のチャンクを待つためにループを抜ける
                # ループ終了後、未処理のデータが残っている可能性があるため処理する
                data = texts.strip()
                if data:
                    parts = data.split('data:')
                    for data_part in parts:
                        data_part = data_part.strip()
                        if not data_part:
                            continue
                        try:
                            event = json.loads(data_part)
                            allOutput.append(event)
                            with open("outputs.json", "w", encoding="utf-8") as f:
                                json.dump(allOutput, f, ensure_ascii=False)
                            if event.get('event') == 'workflow_finished':
                                result = event['data']['outputs']
                                return result
                        except json.JSONDecodeError:
                            pass
                return result
            except Exception as e:
                print(e)
                return result

import AnthropicBedrock from "@anthropic-ai/bedrock-sdk";
import dotenv from "dotenv";

dotenv.config();

const client = new AnthropicBedrock({
  awsSessionToken: process.env.AWS_SESSION_TOKEN,
  awsRegion: process.env.AWS_REGION,
});

async function main() {
  const message = await client.messages.create({
    model: "anthropic.claude-3-5-sonnet-20240620-v1:0",
    max_tokens: 256,
    messages: [{ role: "user", content: "Hello, world" }],
  });
  console.log(message);
}
main().catch(console.error);

python3 -m venv venv
venv\Scripts\activate
pip install torch --index-url https://download.pytorch.org/whl/cu121
pip install transformers
pip install 'accelerate>=0.26.0'

def quick_sort(arr):
    if len(arr) <= 1:
        return arr
    else:
        pivot = arr[len(arr) // 2]
        left = [x for x in arr if x < pivot]
        middle = [x for x in arr if x == pivot]
        right = [x for x in arr if x > pivot]
        return quick_sort(left) + middle + quick_sort(right)

# Example usage:
arr = [3, 6, 8, 10, 1, 2, 1]
sorted_arr = quick_sort(arr)
print(sorted_arr)

llama-server.exe -m モデルパス\DeepSeek-R1-Distill-Qwen-14B-Q4_K_M-00001-of-00001.gguf -c 16384 -ngl 100 --no-mmap --cache-type-k q8_0 --cache-type-v q8_0 --flash-attn

python3 -m venv venv
venv\Scripts\activate
git clone https://github.com/fishaudio/fish-speech
cd fish-speech
pip install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/cu121
pip install -e .
pip install https://github.com/AnyaCoder/fish-speech/releases/download/v0.1.0/triton_windows-0.1.0-py3-none-any.whl
python .\tools\download_models.py
python .\tools\run_webui.py 

python3 -m venv venv
venv\Scripts\activate
git clone https://github.com/microsoft/OmniParser.git
cd OmniParser

cd omnitool/omniparserserver
python -m omniparserserver

const fs = require("fs");
const axios = require("axios");

// FastAPIサーバーのURL（必要に応じて変更してください）
const apiUrl = "http://localhost:8000/parse/";

// 同一ディレクトリの image.jpg を読み込み、Base64に変換する関数
function getBase64Image(filePath) {
  const imageBuffer = fs.readFileSync(filePath);
  return imageBuffer.toString("base64");
}

async function callParseAPI() {
  try {
    // image.jpg を読み込み、Base64エンコード
    const base64_image = getBase64Image("./image.png");

    // リクエストボディのデータを作成
    const data = { base64_image };

    // POSTリクエストを送信
    const response = await axios.post(apiUrl, data);
    console.log("Response data:", response.data);

    // レスポンスからBase64の画像データを取得
    const somImageBase64 = response.data.som_image_base64;
    // Base64文字列をバッファに変換
    const imageBuffer = Buffer.from(somImageBase64, "base64");
    // バッファの内容を output.jpg として保存
    fs.writeFileSync("output.png", imageBuffer);
    console.log("Base64画像を output.jpg として保存しました。");
  } catch (error) {
    console.error("リクエスト送信中にエラーが発生しました:", error);
  }
}

callParseAPI();

git clone https://github.com/Wan-Video/Wan2.1.git
cd Wan2.1
python3 -m venv venv
venv\Scripts\activate

torch>=2.4.0
torchvision>=0.19.0
opencv-python>=4.9.0.80
diffusers>=0.31.0
transformers>=4.49.0
tokenizers>=0.20.3
accelerate>=1.1.1
tqdm
imageio
easydict
ftfy
dashscope
imageio-ffmpeg
flash_attn　←これを削除⭐️
gradio>=5.0.0
numpy>=1.23.5,<2

pip install torch torchvision --index-url https://download.pytorch.org/whl/cu124
pip install -r requirements.txt
pip install flash-attn==2.6.1 --no-build-isolation
pip install "huggingface_hub[cli]"
huggingface-cli download Wan-AI/Wan2.1-T2V-1.3B --local-dir ./Wan2.1-T2V-1.3B

- args.save_file = f"{args.task}_{args.size}_{args.ulysses_size}_{args.ring_size}_{formatted_prompt}_{formatted_time}" + suffix
+ args.save_file = f"{formatted_time}" + suffix

python generate.py  --task t2v-1.3B --size 832*480 --ckpt_dir ./Wan2.1-T2V-1.3B --offload_model True --t5_cpu --sample_shift 8 --sample_guide_scale 6 --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage."

python -m venv myenv
myenv\Scripts\Activate
pip install notebook
pip install ipykernel
python -m ipykernel install --user --name=myenv --display-name "Python (myenv)"
jupyter notebook

// notebookで実行
!git clone https://huggingface.co/spaces/ASLP-lab/DiffRhythm

// notebookで実行
%cd DiffRhythm

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu124    

pip install triton-3.0.0-cp310-cp310-win_amd64.whl

// notebookで実行
!pip install -r requirements.txt

// notebookで実行
import os
os.environ["PHONEMIZER_ESPEAK_LIBRARY"] = "C:\\Program Files\\eSpeak NG\\libespeak-ng.dll"

diff --git a/diffrhythm/g2p/g2p/__init__.py b/diffrhythm/g2p/g2p/__init__.py
index a926893..5f35e62 100755
--- a/diffrhythm/g2p/g2p/__init__.py
+++ b/diffrhythm/g2p/g2p/__init__.py
@@ -15,21 +15,16 @@ class PhonemeBpeTokenizer:

     def __init__(self, vacab_path="./diffrhythm/g2p/g2p/vocab.json"):
         self.lang2backend = {
-            "zh": "cmn",
-            "ja": "ja",
             "en": "en-us",
-            "fr": "fr-fr",
-            "ko": "ko",
-            "de": "de",
         }
         self.text_tokenizers = {}
         self.int_text_tokenizers()
 
-        with open(vacab_path, "r") as f:
+        with open(vacab_path, "r", encoding="utf-8") as f:
             json_data = f.read()
         data = json.loads(json_data)
         self.vocab = data["vocab"]
-        LangSegment.setfilters(["en", "zh", "ja", "ko", "fr", "de"])
+        LangSegment.setfilters(["en"])

     def int_text_tokenizers(self):
         for key, value in self.lang2backend.items():
diff --git a/diffrhythm/g2p/g2p_generation.py b/diffrhythm/g2p/g2p_generation.py
index e78f3e7..4b73936 100755
--- a/diffrhythm/g2p/g2p_generation.py
+++ b/diffrhythm/g2p/g2p_generation.py
@@ -114,7 +114,7 @@ def chn_eng_g2p(text: str):


 text_tokenizer = PhonemeBpeTokenizer()
-with open("./diffrhythm/g2p/g2p/vocab.json", "r") as f:
+with open("./diffrhythm/g2p/g2p/vocab.json", "r", encoding="utf-8") as f:
     json_data = f.read()
 data = json.loads(json_data)
 vocab = data["vocab"]
diff --git a/diffrhythm/g2p/utils/g2p.py b/diffrhythm/g2p/utils/g2p.py
index aba5e2f..ee1e1d4 100755
--- a/diffrhythm/g2p/utils/g2p.py
+++ b/diffrhythm/g2p/utils/g2p.py
@@ -14,11 +14,6 @@ import sys
 # separator=Separator(phone=' ', word=' _ ', syllable='|'),
 separator = Separator(word=" _ ", syllable="|", phone=" ")

-phonemizer_zh = EspeakBackend(
-    "cmn", preserve_punctuation=False, with_stress=False, language_switch="remove-flags"
-)
-# phonemizer_zh.separator = separator
-
 phonemizer_en = EspeakBackend(
     "en-us",
     preserve_punctuation=False,
@@ -27,40 +22,11 @@ phonemizer_en = EspeakBackend(
 )
 # phonemizer_en.separator = separator

-phonemizer_ja = EspeakBackend(
-    "ja", preserve_punctuation=False, with_stress=False, language_switch="remove-flags"
-)
-# phonemizer_ja.separator = separator
-
-phonemizer_ko = EspeakBackend(
-    "ko", preserve_punctuation=False, with_stress=False, language_switch="remove-flags"
-)
-# phonemizer_ko.separator = separator
-
-phonemizer_fr = EspeakBackend(
-    "fr-fr",
-    preserve_punctuation=False,
-    with_stress=False,
-    language_switch="remove-flags",
-)
-# phonemizer_fr.separator = separator
-
-phonemizer_de = EspeakBackend(
-    "de", preserve_punctuation=False, with_stress=False, language_switch="remove-flags"
-)
-# phonemizer_de.separator = separator
-
-
 lang2backend = {
-    "zh": phonemizer_zh,
-    "ja": phonemizer_ja,
     "en": phonemizer_en,
-    "fr": phonemizer_fr,
-    "ko": phonemizer_ko,
-    "de": phonemizer_de,
 }

-with open("./diffrhythm/g2p/utils/mls_en.json", "r") as f:
+with open("./diffrhythm/g2p/utils/mls_en.json", "r", encoding="utf-8") as f:
     json_data = f.read()
 token = json.loads(json_data)

diff --git a/diffrhythm/infer/infer_utils.py b/diffrhythm/infer/infer_utils.py
index e403235..89093d6 100755
--- a/diffrhythm/infer/infer_utils.py
+++ b/diffrhythm/infer/infer_utils.py
@@ -99,7 +99,7 @@ def parse_lyrics(lyrics: str):

 class CNENTokenizer():
     def __init__(self):
-        with open('./diffrhythm/g2p/g2p/vocab.json', 'r') as file:
+        with open('./diffrhythm/g2p/g2p/vocab.json', 'r', encoding="utf-8") as file:
             self.phone2id:dict = json.load(file)['vocab']
         self.id2phone = {v:k for (k, v) in self.phone2id.items()}
         # from f5_tts.g2p.g2p_generation import chn_eng_g2p

// notebookで実行
!python main.py --ref_audio "example/サンプルの音源.wav" --output "出力ファイル名.wav"

python3 -m venv venv
venv\Scripts\activate
pip install torch --index-url https://download.pytorch.org/whl/cu124
pip install transformers
pip install 'accelerate>=0.26.0'

import transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
import torch
import uuid
import time

# FastAPI アプリの作成
app = FastAPI()

# CORS ミドルウェアの設定（必要なら制限可能）
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],  # 全オリジン許可
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# モデルのロード（API 起動時に1回だけ）
print("Loading model... (This may take some time)")
model_id = "microsoft/Phi-4-mini-instruct"
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
    torch_dtype="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_id)

# テキスト生成パイプラインの作成（API 起動時に1回だけ）
text_generator = transformers.pipeline("text-generation", model=model, tokenizer=tokenizer)
print("Model loaded successfully!")

# OpenAI互換のテキスト補完リクエストスキーマ
class CompletionRequest(BaseModel):
    model: str
    prompt: str
    max_tokens: int = 12800
    temperature: float = 0.7
    top_p: float = 0.9

@app.post("/v1/chat/completions")
async def completions(request: CompletionRequest):
    try:
        # デバッグ: 入力された prompt を標準出力に出力
        print("Request Received")
        #print(request.prompt)

        # 補完生成
        outputs = text_generator(
            request.prompt,
            max_new_tokens=request.max_tokens,
            do_sample=True,
            temperature=request.temperature,
            top_p=request.top_p
        )
        print(outputs[0]["generated_text"])

        # OpenAI互換のレスポンス形式（UUIDでidを生成）
        response = {
            "id": str(uuid.uuid4()),
            "object": "text_completion",
            "created": int(time.time()),
            "model": request.model,
            "choices": [
                {
                    "text": outputs[0]["generated_text"],
                    "index": 0,
                    "finish_reason": "stop"
                }
            ]
        }

        return response

    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

@app.get("/")
def read_root():
    return {"message": "OpenAI-compatible API for VSCode continue is running!"}

llama-server.exe -m モデルパス -c 16384 -ngl 100 --no-mmap --cache-type-k q8_0 --cache-type-v q8_0 --flash-attn

import os
from dotenv import load_dotenv
from agents import Agent, Runner, WebSearchTool, set_tracing_export_api_key

load_dotenv()
set_tracing_export_api_key(os.getenv("OPENAI_API_KEY"))


child_agent1 = Agent(
    name="Assistant",
    instructions="あなたはQiitaやGitHubなどの情報を発信内容を要約するエージェントです。スキル判定に必要を重点的に収集してください。",
    tools=[
        WebSearchTool(),
    ]
)

child_agent2 = Agent(
    name="Assistant",
    instructions="スキルを判断し、年収レンジを推測するエージェントです。",
)

agent = Agent(
    name="Assistant",
    instructions="年収判定AIです。情報収集エージェントと年収推測エージェントを活用してユーザーの年収レンジを算出します。各エージェントの調査結果に不足があれば、追加で調査指示を出してください。評価の高かった具体的なポイントや記事の名前、活動なども出力してください。",
    model="o3-mini-2025-01-31",
    handoffs=[child_agent1,child_agent2],
)

result = Runner.run_sync(agent, "私はソフトウェアエンジニアのmegumu-uです。私の情報についてWebで確認し、Qiitaの記事(https://qiita.com/megumu-u)の評価や、QiitaからたどれるGitHub(https://github.com/uemegu)の実装内容等からスキルを推測し、私がもらうべき年収はどのようなレンジがふさわしいか判断してください。今後の活動の参考にします。。")
print(result.final_output)

python3 -m venv venv
venv\Scripts\activate
git clone https://github.com/joanrod/star-vector.git
cd star-vector
pip install packaging wheel
pip install -e .

venv\Scripts\activate
pip uninstall torch torchvision
pip install torch==2.5.1 torchvision==0.20.1 --index-url https://download.pytorch.org/whl/cu124
pip install fairscale
set DISTUTILS_USE_SDK=1
pip install flash-attn==2.7.3 --no-build-isolation
pip install deepspeed

from PIL import Image
from transformers import AutoModelForCausalLM, AutoTokenizer, AutoProcessor
from starvector.data.util import process_and_rasterize_svg
import torch
 
model_name = "starvector/starvector-8b-im2svg"
 
starvector = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16, trust_remote_code=True)
processor = starvector.model.processor
tokenizer = starvector.model.svg_transformer.tokenizer
 
starvector.cuda()
starvector.eval()
 
image_pil = Image.open('assets/examples/sample-18.png')
 
image = processor(image_pil, return_tensors="pt")['pixel_values'].cuda()
if not image.shape[0] == 1:
    image = image.squeeze(0)
batch = {"image": image}
 
raw_svg = starvector.generate_im2svg(batch, max_length=4000)[0]
svg, raster_image = process_and_rasterize_svg(raw_svg)

wsl --install
wsl.exe -d Ubuntu

sudo apt update
sudo apt install -y nvidia-cuda-toolkit
nvidia-smi
sudo apt install python3-pip python3.12-venv

python3 -m venv nanogpt-env
source nanogpt-env/bin/activate
pip install torch --index-url https://download.pytorch.org/whl/cu118
pip install torch numpy transformers datasets tiktoken wandb tqdm
git clone https://github.com/karpathy/nanoGPT
cd nanoGPT/
python data/shakespeare/prepare.py

code .

diff --git a/config/finetune_shakespeare.py b/config/finetune_shakespeare.py
index 148a4c4..eab4306 100644
--- a/config/finetune_shakespeare.py
+++ b/config/finetune_shakespeare.py
@@ -8,7 +8,7 @@ wandb_project = 'shakespeare'
 wandb_run_name = 'ft-' + str(time.time())

 dataset = 'shakespeare'
-init_from = 'gpt2-xl' # this is the largest GPT-2 model
+init_from = 'gpt2' # this is the largest GPT-2 model

python3 train.py config/finetune_shakespeare.py
python sample.py   --out_dir=out-shakespeare   --start="ROMEO:"   --num_samples=3   --max_new_tokens=100

ROMEO:

Mentos.

Mentos sojourn in that house;
As,
He is his father's lord.

Mentos:
So he is, for that.

LARRY:
O, my lord!

MENTOS:
O, what is it, sir?

LARRY:
That it is, sir.

LARRY:
How so?

O,
Say,
---------------
ROMEO: we are given in our faces to say, 'It is settled,' 'it is what it is,' 'it is what it is,' 'it is what it is.'

CURTIS: then, though the three of you were, as I know, at the first meeting of the Council of Trent, yet they were in this city, and were, as I am, dead; but in the old town, which was, in all probability, the body of the Council,
---------------
ROMEO:
Boy, be it with you, I shall think it best to call upon thee of this matter.

DON:
You know how this man is now, son of the king.

JAN:
And, son of that king, thou art in his
wiring:
Do not resist, I'd like to stay, but bet on such
as thou art to hear.

JAN:
I'll grant it, and therefore go to the King's

brew install espeak-ng
 python -m venv venv
source venv/bin/activate
pip install kokoro soundfile IPython
pip install pyopenjtalk-plus
pip install mojimoji
pip install jaconv
pip install fugashi unidic-lite

from kokoro import KPipeline
from IPython.display import display, Audio
import soundfile as sf
import torch
pipeline = KPipeline(lang_code='ja')
text = '''‘
こんにちは！これは音声合成で作られた声だよ！よろしくね！
'''
generator = pipeline(text, voice='jf_alpha')
for i, (gs, ps, audio) in enumerate(generator):
    print(i, gs, ps)
    display(Audio(data=audio, rate=24000, autoplay=i==0))
    sf.write(f'{i}.wav', audio, 24000)

git clone https://github.com/VAST-AI-Research/TripoSG.git
cd TripoSG
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu124
pip install -r requirements.txt
pip uninstall numpy -y
pip install numpy --upgrade

pip install torch torchvision --index-url https://download.pytorch.org/whl/cu124
pip install -r requirements.txt
pip install "flash_attn-2.7.4.post1+cu124torch2.6.0cxx11abiFALSE-cp311-cp311-win_amd64.whl"
pip install triton-3.0.0-cp311-cp311-win_amd64.whl

git clone https://github.com/edwko/OuteTTS.git
python -m venv venv
source venv/bin/activate
cd OuteTTS
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu124
set CMAKE_ARGS=-DGGML_CUDA=on
pip install outetts --upgrade

import outetts

# Initialize the interface
interface = outetts.Interface(
    config=outetts.ModelConfig.auto_config(
        model=outetts.Models.VERSION_1_0_SIZE_1B,
        # For llama.cpp backend
        backend=outetts.Backend.LLAMACPP,
        quantization=outetts.LlamaCppQuantization.FP16
        # For transformers backend
        # backend=outetts.Backend.HF,
    )
)

# デフォルト音声で作る場合
# speaker = interface.load_default_speaker("EN-FEMALE-1-NEUTRAL")

# カスタム音声で作る場合
speaker = interface.create_speaker("sample.wav")
interface.save_speaker(speaker, "speaker.json")
speaker = interface.load_speaker("speaker.json")

# Generate speech
output = interface.generate(
    config=outetts.GenerationConfig(
        text="こんにちは！元気ですか？",
        generation_type=outetts.GenerationType.CHUNKED,
        speaker=speaker,
        sampler_config=outetts.SamplerConfig(
            temperature=0.4
        ),
    )
)

# Save to file
output.save("output.wav")

git clone https://github.com/lllyasviel/FramePack.git
python -m venv venv
venv\Scripts\activate
cd FramePack
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu126
pip install -r requirements.txt
python demo_gradio.py

python -m venv venv
venv\Scripts\activate
pip install torch torchaudio torchvision --index-url https://download.pytorch.org/whl/cu124
pip install git+https://github.com/getuka/canary-tts.git
pip install triton-3.0.0-cp311-cp311-win_amd64.whl

import torch, torchaudio
from transformers import AutoModelForCausalLM, AutoTokenizer
from canary_tts.xcodec2.modeling_xcodec2 import XCodec2Model
from rubyinserter import add_ruby

tokenizer = AutoTokenizer.from_pretrained("2121-8/canary-tts-0.5b")
model = AutoModelForCausalLM.from_pretrained(
    "2121-8/canary-tts-0.5b",
    device_map="auto",
    torch_dtype=torch.bfloat16
)
codec = XCodec2Model.from_pretrained("HKUSTAudio/xcodec2")

description = "A cheerful girl with a high-pitched, lively voice speaks with energetic and dynamic intonation. Her excitement comes through clearly, and the recording feels bright and close, with very little noise."
prompt = 'ローカル環境で完結してアニメ作りたいんです!!'
prompt = add_ruby(prompt)

chat = [
    {"role": "system", "content": description},
    {"role": "user", "content": prompt}
]

tokenized_input = tokenizer.apply_chat_template(
    chat,
    add_generation_prompt=True,
    tokenize=True,
    return_tensors="pt"
).to(model.device)

with torch.no_grad():
    output = model.generate(
        tokenized_input,
        max_new_tokens=256,
        top_p=0.8, 
        temperature=0.3, 
        repetition_penalty=1.1,
    )[0]

audio_tokens = output[len(tokenized_input[0]):]
decoded_audio = codec.decode_code(
    audio_tokens.unsqueeze(0).unsqueeze(0).cpu()
)

torchaudio.save(
    "sample2.wav",
    src=decoded_audio[0].cpu(),
    sample_rate=16000
)

git clone https://huggingface.co/Respair/Tsukasa_Speech
cd Tsukasa_Speech
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt
pip install gradio sentencepiece konoha openai

diff --git a/app_tsuka.py b/app_tsuka.py
index 1c216e7..91e0e69 100644
--- a/app_tsuka.py
+++ b/app_tsuka.py
@@ -78,7 +78,7 @@ theme = gr.themes.Base(
-voicelist = [v for v in os.listdir("/home/ubuntu/Kanade_Project/gradio/Tsukasa_Speech/reference_sample_wavs")]
+voicelist = [v for v in os.listdir("./reference_sample_wavs")]

python app_tsuka.py

git clone https://github.com/ace-step/ACE-Step.git
cd .\ACE-Step\
python -m venv venv
venv\Scripts\activate
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu126
pip install -e .

acestep --port 7865