pyVideoTrans

Appearance

Compatible AI, Local Large Language Models, and Usage Methods for Domestic AIs Compatible with the OpenAI ChatGPT API ​

In video translation and dubbing software, AI large language models can serve as efficient translation channels, significantly improving translation quality by leveraging context.

Currently, most domestic AI APIs are compatible with OpenAI's technology. Therefore, users can directly operate within the Compatible AI/Local Model channel. They can also use it locally after deploying via ollama.

Using Moonshot AI ​

  1. Menu Bar -> Translation Settings -> OpenAI ChatGPT API Settings Interface
  2. In the API Endpoint field, enter https://api.moonshot.cn/v1
  3. In the SK field, enter the API Key obtained from the Moonshot open platform. You can get it from this URL: https://platform.moonshot.cn/console/api-keys
  4. In the Model text area, enter moonshot-v1-8k,moonshot-v1-32k,moonshot-v1-128k
  5. Then, select the model you want to use from the dropdown, test it, and save if there are no issues.

Using Baichuan AI ​

  1. Menu Bar -> Translation Settings -> OpenAI ChatGPT API Settings Interface
  2. In the API Endpoint field, enter https://api.baichuan-ai.com/v1
  3. In the SK field, enter the API Key obtained from the Baichuan open platform. You can get it from this URL: https://platform.baichuan-ai.com/console/apikey
  4. In the Model text area, enter Baichuan4,Baichuan3-Turbo,Baichuan3-Turbo-128k,Baichuan2-Turbo
  5. Then, select the model you want to use from the dropdown, test it, and save if there are no issues.

01.AI (Lingyi Wanwu) ​

Official Website: https://lingyiwanwu.com

API Key Acquisition Address: https://platform.lingyiwanwu.com/apikeys

API URL: https://api.lingyiwanwu.com/v1

Available Model: yi-lightning

Notes: ​

  1. Most AI translation channels may limit requests per minute. If you encounter an error indicating the request rate is exceeded during use, you can click the "Translation Channel↓" button on the software's main interface. In the pop-up window, change the "Pause Seconds" to 10. This means waiting 10 seconds after each translation before initiating the next request, limiting it to a maximum of 6 requests per minute to prevent exceeding the frequency limit.

  2. If the selected model is not intelligent enough, especially locally deployed models which are often smaller due to hardware limitations, it may not accurately follow instructions to return translations in the required format. This could result in too many blank lines in the translation output. In this case, you can try using a larger model, or go to Menu -> Tools/Options -> Advanced Options -> "Send full subtitle content when using AI translation" and uncheck this option.

Deploying the Tongyi Qianwen Large Model Locally Using Ollama ​

If you have some technical skills, you can also deploy a large model locally and use it for translation. This section uses Tongyi Qianwen as an example to introduce the deployment and usage method.

1. Download the exe and Run it Successfully ​

Open the URL: https://ollama.com/download

Click to download. After downloading, double-click to open the installation interface and click Install to complete.

After completion, a black or blue window will automatically pop up. Enter the 3 words ollama run qwen and press Enter. This will automatically download the Tongyi Qianwen model.

Wait for the model download to finish. No proxy is needed, and the speed is quite fast.

Once the model is automatically downloaded, it will run directly. When the progress reaches 100% and displays the "Success" characters, it means the model is running successfully. At this point, the entire installation and deployment of the Tongyi Qianwen large model is complete, and you can happily use it. Isn't it super simple?

The default API endpoint address is http://localhost:11434

If the window closes, how do you reopen it? It's also very simple. Open the computer's Start menu, find "Command Prompt" or "Windows PowerShell" (or directly press Win key + Q key, type cmd to search), click to open it, enter ollama run qwen, and you're done.

2. Using it Directly in the Command Console Window ​

As shown in the image below, when this interface is displayed, you can actually start using it by directly typing text in the window.

3. Of course, this interface might not be very user-friendly, so let's get a friendly UI ​

Open the URL: https://chatboxai.app/zh and click to download.

After downloading, double-click and wait for the interface window to open automatically.

Click "Start Setup". In the pop-up layer, click on the top section for Model, select "Ollama" under AI Model Provider, fill in the API Domain with the address http://localhost:11434, select Qwen:latest from the model dropdown menu, and then save. That's it.

After saving, the usage interface is displayed. Use your imagination and use it freely.

4. Enter the API into the Video Translation and Dubbing Software ​

  1. Open Menu -> Settings -> Compatible OpenAI and Local Large Models. In the middle text box, add a model ,qwen. After adding, it should look like the following, then select this model.

  1. In the API URL field, enter http://localhost:11434/v1. For SK, enter anything, for example, 1234.

  1. Test to see if it's successful. If successful, save it and go use it.

5. What Other Models Can Be Used? ​

Besides Tongyi Qianwen, there are many other models you can use. The usage method is equally simple, just 3 words: ollama run model_name

Open this address: https://ollama.com/library to see all model names. Copy the name of the one you want to use, then execute ollama run model_name.

Remember how to open the command window? Click the Start menu, find Command Prompt or Windows PowerShell.

For example, if I want to install the openchat model:

Open Command Prompt, enter ollama run openchat, press Enter, and wait until "Success" is displayed.

Notes: ​

  1. Most AI translation channels may limit requests per minute. If you encounter an error indicating the request rate is exceeded during use, you can click the "Translation Channel↓" button on the software's main interface. In the pop-up window, change the "Pause Seconds" to 10. This means waiting 10 seconds after each translation before initiating the next request, limiting it to a maximum of 6 requests per minute to prevent exceeding the frequency limit.

  2. If the selected model is not intelligent enough, especially locally deployed models which are often smaller due to hardware limitations, it may not accurately follow instructions to return translations in the required format. This could result in too many blank lines in the translation output. In this case, you can try using a larger model, or go to Menu -> Tools/Options -> Advanced Options -> "Send full subtitle content when using AI translation" and uncheck this option.