• State Not Answered
  • Replies 1 reply
  • Subscribers 181 subscribers
  • Views 51 views
  • Users 0 members are here
Options
Related

Refine translations with Azure open AI provider not working

T T

I have a project using my translation memory and the Microsoft translator provider which I use for pre-translate (Microsoft translation provider is used for auto translating).

I can only process 4-6 files in one time, if I process all files together I have errors.

Table showing translation memory and providers, including Microsoft Translator Provider and OpenAI provider for Trados, with enabled and lookup options checked.

Afterwards I refine the translations with 'Pre-translate Files with OpenAI provider for Trados'. 

Edit Model Connection window for Azure OpenAI, showing fields for name, provider, endpoint, API key, model, model type, and temperature slider.

Settings window showing connections, prompts, and search options for refining translations, with options like ignoring translated segments and enabling terminology-aware suggestions.

Table showing translation memory and providers, with OpenAI provider for Trados selected and lookup option checked.

When I process files with about 80%-90% translation progress, everythings works fine.  Files with 10%-50% provide following error:

Error message stating 'Unexpected exception when configuring file multiFileConverter for task Pre-translate Files with OpenAI provider for Trados: Only Exception type is allowed.'

This is the error in the file editor view:

Error message stating 'An error has occurred while using the translation provider OpenAI provider for Trados: Unexpected end when deserializing object. Path error, line 1, position 442.'

When I perform a verify file batch task, I don't have any errors.

I have the impression that a small file with about 400-500 words and 20% still works, but a larger file with 2000-3000 words doesn't work.  Sometimes a second attempt works.

Is there a threshold or a limit or a time out? Is a log file created when the error occurs?

Many thanks!

Generated Image Alt-Text
[edited by: RWS Community AI at 9:22 PM (GMT 1) on 15 Jul 2025]
emoji
Rate translation
×
Suggest better translation
×
Moderator UI
×
  • Get AI Suggestion
Parents
  • 0

    Probability score: 0.7  

    Question improvement suggestion: Provide specific error messages or codes encountered when processing files, as well as the exact version of the Azure OpenAI provider being used.  

    In the meantime, based on the information you have provided, here is a suggested answer to your question:  

    Hello,  

    It seems you are experiencing issues with refining translations using the Azure OpenAI provider in Trados Studio. Here are some steps and considerations that may help you resolve the problem:

    **Step 1:** **Check for Error Messages**  

    Make sure to note any specific error messages or codes that appear when you attempt to process files with 10%-50% translation progress. This information can be crucial for diagnosing the issue.

    **Step 2:** **Review File Size and Content**  

    Since you mentioned that smaller files (400-500 words) seem to work while larger files (2000-3000 words) do not, consider splitting larger files into smaller segments to see if that alleviates the issue. Additionally, ensure that the translation units do not exceed the character limits set by the Azure OpenAI provider.

    **Step 3:** **Adjust Timeout Settings**  

    If you suspect that timeouts may be causing the issue, you can increase the Server Farm Proxy Timeout in IIS. Set it to a higher value (e.g., 60, 120, or 180 seconds) to see if that resolves the problem.

    **Step 4:** **Use Workarounds**  

    As a workaround, you can pre-translate the project using the Language Weaver Cloud plugin instead of the Azure OpenAI provider. This has been noted as a successful method in similar cases.

    **Step 5:** **Verify Files Before Pre-translation**  

    Run a 'Verify Files' task before attempting to pre-translate. This can help identify any potential issues that might cause exceptions during the pre-translation process.

    **Step 6:** **Check Translation Memory Settings**  

    Ensure that your Translation Memory does not contain overly long translation units or tags that could interfere with processing. You may want to clean up your TMX file by removing any translation units that exceed a certain character limit.

    If the problem persists after trying these steps, please provide any specific error messages you encounter, as well as the exact version of the Azure OpenAI provider you are using. This information will help in further diagnosing the issue.

    I hope this answer addressed your question. If not, please reject the answer, add the question improvement suggestion information, and click the RWS Community AI checkbox on the right to try again.  

    RWS Community AI
    emoji
Reply
  • 0

    Probability score: 0.7  

    Question improvement suggestion: Provide specific error messages or codes encountered when processing files, as well as the exact version of the Azure OpenAI provider being used.  

    In the meantime, based on the information you have provided, here is a suggested answer to your question:  

    Hello,  

    It seems you are experiencing issues with refining translations using the Azure OpenAI provider in Trados Studio. Here are some steps and considerations that may help you resolve the problem:

    **Step 1:** **Check for Error Messages**  

    Make sure to note any specific error messages or codes that appear when you attempt to process files with 10%-50% translation progress. This information can be crucial for diagnosing the issue.

    **Step 2:** **Review File Size and Content**  

    Since you mentioned that smaller files (400-500 words) seem to work while larger files (2000-3000 words) do not, consider splitting larger files into smaller segments to see if that alleviates the issue. Additionally, ensure that the translation units do not exceed the character limits set by the Azure OpenAI provider.

    **Step 3:** **Adjust Timeout Settings**  

    If you suspect that timeouts may be causing the issue, you can increase the Server Farm Proxy Timeout in IIS. Set it to a higher value (e.g., 60, 120, or 180 seconds) to see if that resolves the problem.

    **Step 4:** **Use Workarounds**  

    As a workaround, you can pre-translate the project using the Language Weaver Cloud plugin instead of the Azure OpenAI provider. This has been noted as a successful method in similar cases.

    **Step 5:** **Verify Files Before Pre-translation**  

    Run a 'Verify Files' task before attempting to pre-translate. This can help identify any potential issues that might cause exceptions during the pre-translation process.

    **Step 6:** **Check Translation Memory Settings**  

    Ensure that your Translation Memory does not contain overly long translation units or tags that could interfere with processing. You may want to clean up your TMX file by removing any translation units that exceed a certain character limit.

    If the problem persists after trying these steps, please provide any specific error messages you encounter, as well as the exact version of the Azure OpenAI provider you are using. This information will help in further diagnosing the issue.

    I hope this answer addressed your question. If not, please reject the answer, add the question improvement suggestion information, and click the RWS Community AI checkbox on the right to try again.  

    RWS Community AI
    emoji
Children
No Data