videos and more

Hi,

besides the few videos in the Videos section, would there be the possibility to add some videos showing how LWE integrates and works within Trados Studio 2022?

I also miss information about if there are rules that can be set  up and if so, how? How are the results finetuned? If you have a look at Opus-CAT MT’s website, they have a lot of information with screenshots about how everyhting works. Unfortunately for LWE there is no visible similar information. Disappointed If you could implement this in a video too, that would be great. Or is there the possibility to have a quick chat via Skype to show how it works via shared screen?

EDIT: I forgot to add the questions I also had for Opus:

- Is it possible to implement grammar rules in Opus-CAT MT? In Luxembourgish for example we have a quite complex rule for dropping final n which would make most translations useless with simple search and replace. I got it narrowed down to 130 lines of code and besides the rules and exceptions it requires a db with the words it applies to in the background to work correctly.

- Are genders of words considered when matching against terminology? To me it seems it only searches and replaces terms and this again leads to many grammar errors.

- Are categories considered when translating, say you select a field of translation before you start translating with Opus? (e.g. medical, HR, Technical, or even more fine tuned subcategories?)

- is it possible to use German as target language for Luxembourgish and work with search and replace to get Luxembourgish texts for those parts where there is no sufficient data to translate to LB directly?

and a few regarding LWE:

- what are the costs for using LWE on your own server?

- Right now I only have a Windows Server, so when will the version for Windows be available approximately?

Best regards,

Pascal



I forgot to add the questions
[edited by: Pascal Zotto at 12:17 PM (GMT 1) on 14 Oct 2022]
emoji
Parents Reply Children
  • Hello Pascal,

    Please see below the steps to connect Trados Studio to Language Weaver Edge.

    Pre-requisites to use connect Trados Studio to Language Weaver Edge:

    1. Have a Language Weaver Edge account.
    2. Obtain you <API Key>: This is a credential equivalent to a user and password. To find out the API Key, log into the Language Weaver Edge Web GUI as an admin user and click to [Username] > My Account. The API Key will be something like u_user@example.com_NM2ba8j9pHQ2.
    3. Install the Language Weaver Edge for Trados Studio (https://appstore.rws.com/plugin/40/).

     

    Once you have installed the app, you can follow the steps below to connect Trados Studio to Language Weaver Edge.

    1. From Trados Studio, go to the Project Settings.
    2. Go to Language Pairs > All Language Pairs > Translation Memory and Automated Translation and click on SDL Machine Translation Edge.

      Trados Studio Project Settings window showing the selection of SDL Machine Translation Edge under Translation Memory and Automated Translation.
    3. Enter your connection details as follows:

      SDL Machine Translation Edge settings window with fields for Host, Port, and API Key. The 'Requires secure protocol' option is checked and 'Use API Key' is selected.
    4. Click on OK.  You should now get the MT for each segment in your translation file.
    emoji


    Generated Image Alt-Text
    [edited by: Trados AI at 10:51 AM (GMT 0) on 4 Mar 2024]
  • Hi ,

    sorry, but this does not even answer my question to the slightest. I did not ask how to connect or use it in Trados but I asked for explanations and videos to the above questions.

    Furthermore, I don’t have any credentials yet as I did not book any subscription as I’m still trying to find out if that NMT even suits my needs. Which I can decide on, if I get answers to my questions.

    Br,

    Pascal

    emoji
  • Hi  ,

    I'm not aware of videos showing how Studio 2022 works with Language Weaver Edge, and I'm sorry if we don't have any, but there are videos showing how Studio is working with Language Weaver (our Saas offering). This one in particular should provide more details on the integration and how it is working in Studio: https://www.youtube.com/watch?v=AhP-Q1m9gyk

    Note that Language Weaver (Saas) and Language Weaver Edge (self hosted) have the same feature scope and capabilities - they both support adaptation capabilities, starting with Dictionaries and Brands, up to model adaptation with Adaptable Language Pair training. With the upcoming Edge 8.6, we will also support Auto adaptive language pairs, which will use user feedback to adapt the models transparently. Note that this specific feature will require an update of the Studio Edge plugin - and an internal feature request has already been placed to allow user feedback to be sent from Studio to Edge.

    With the current studio plugin, you can already reference a specific Edge language pair and specify a dictionary that needs to be applied on top of the language pair. 

    Multiple dictionaries are supported per Language Pair, so you can define a dictionary for each category/domain you're working on and select the appropriate dictionary based on the specific project. This can be enforced for each Language Pair Mapping in the Studio Plugin.

    If you already have domain specific translation memories, you can also train multiple adapted Language Pairs, one for each domain and select the appropriate one in the Studio template. And Dictionaries can work on top of generic or adapted language pairs.

    Dictionaries do not support inflection at the moment (when the source segment is found, it is tagged and then replaced in the output with the corresponding target segment); but we are working on an improved terminology support which will provide better support for inflection and will produce more natural output. I'm expecting to see the first improved models in Q1 2023 and then we'll release support for more languages throughout the course of 2023.

    We don't support Grammar Rules, but Edge 8.6 (our upcoming release) will support pre and post-processing scripts that can be called before the content is processed and after the content has been processed. You would be able to apply specific rules to prepare your content and then apply them on the output. This would work best for structured content or plain text.

    We don't have Luxembourgish Language pairs, but if you have Luxembourgish content in translation memories, the adaptation process would already steer the German model in producing an output which would be closer to Luxembourgish.

    Regarding Edge specifically, the product supports both Windows and Linux, so it can be deployed on any of the supported Operating systems listed in the documentation at https://docs.rws.com/833475/397044/language-weaver-edge-documentation/compatible-operating-systems

    Note that the upcoming 8.6 will also support Windows Server 2022.

    Regarding a demo and pricing, our Sales team will be happy to get in touch and provide more details about the product capabilities and provide a quote to meet your needs. You can contact them directly by sending an email to sales@languageweaver.com or we can ask a member of our sales team to reach out to you. 

    Thank you

    emoji