- Dec 29, 2017
- 2,709
- 4,899
I think I will pause my efforts on the "user friendly" SLR system for now.
I need to know the resources of an up to date Translator++ package so I can actually make this run properly.
Basically I need to know what kind of ruby package it uses, what version of psych, if there were modifications to the rvdata gem, and whether or not the rmrgss module includes different code compared to 5.6.3, so I can build from there.
Currently I'm either missing a bunch of stuff, or Translator++ is even more broken than I remember.
Otherwise working on this approach is a waste of time because it's just going to crash when using it on Ace.
I doubt a system that only works on MV/MZ is particularly useful, because most "hopeless" games are Ace. (Because Ace is annoying to work with.)
I'm going to wait a month in case someone can provide that stuff to me, and if not I will scrap the Idea and try to take a different approach.
It's probably not going to be very easy to use then though.
Edit: Nvm Dreamsavior did not fix anything in the newer versions.
Basically I need to know what kind of ruby package it uses, what version of psych, if there were modifications to the rvdata gem, and whether or not the rmrgss module includes different code compared to 5.6.3, so I can build from there.
Currently I'm either missing a bunch of stuff, or Translator++ is even more broken than I remember.
Otherwise working on this approach is a waste of time because it's just going to crash when using it on Ace.
I doubt a system that only works on MV/MZ is particularly useful, because most "hopeless" games are Ace. (Because Ace is annoying to work with.)
It's probably not going to be very easy to use then though.
Edit: Nvm Dreamsavior did not fix anything in the newer versions.
Last edited:
- Dec 29, 2017
- 2,709
- 4,899
Here you go:
https://f95zone.to/threads/translation-request-ghost-house-gunsweeper-t-enta-p.186521/post-12665681
Sorry for the missing choice boxes, but I cannot automate them and doing them manually would take a very long time because of how incredibly inefficient VX is.
The vast majority of them are just this though:
Do Not Skip
Skip
Adjust
Don't Adjust
Add
Do not Add
Give up body
Do not
Stuff
More Advanced Stuff
Even More Advanced Stuff
Super Advanced Stuff
Cancel
Solve Puzzle
Do not, even though you have to anyway.
Last edited:
Requesting another translation boss Shisaye
https://f95zone.to/threads/sword-of-meridyth-rj195652-hakkaku.151456/
Someone kindly provided a mega download link on page 2.
https://f95zone.to/threads/sword-of-meridyth-rj195652-hakkaku.151456/
Someone kindly provided a mega download link on page 2.
- Dec 29, 2017
- 2,709
- 4,899
According to the readme the mega link is v1.1.3 (2017.4.17).
The latest version would be 1.1.4 (2017.4.20)
Any clue where I can find the latest?
- Nov 3, 2021
- 320
- 765
It's slightly off-topic, but a while ago we said:
I finished it! It took way longer than it should have, though I likely would've had an easier time if I had found/followed some sort of guide. It's a small, straight-forward game so it likely isn't the best representative of the kind of stuff WolfRPG can throw your way, but all-in-all it was easier than some of the more complex VX Ace games I've seen. Even though I did my editing in-engine, T++ was really useful since I could search "きてしま" and it would tell me the exact line and CommonEvent I would need to edit. I think someone with as much RPGM experience as you wouldn't have any real trouble with a WolfRPG game (aside from the inability to automate it).
Ah I thought the mega link provided was the latest. I've asked the guy about it and waiting for the answer, hopefully it is the latest and the readme is just unupdated lol. Because yeah I'm having trouble finding this game for free online.
- Dec 29, 2017
- 2,709
- 4,899
Well, good job.
I'm glad you didn't jump off a bridge.
But Manually MTLing something is almost the same work as a manual translation, taking away the point of a MTL, and for me it's a massive waste of time to do a manual translation because people neither like my English nor my general translation style.
Every time I tried that the reaction was basically "barely above MTL".
As a result unless someone finds a way to make Wolf stuff readable and correctly convert it back, there is practically no chance that I will ever work with Wolf RPG again.
Currently I even struggle with VX Ace... because my "user friendly" SLR revamp is now forced to use yaml conversions instead of my Json system and it's awful.
My options are currently just,
Option 2 I'm basically stuck on because Dreamsavior's code is somehow even worse than mine.
Option 3 would take several months.
So that's fun.
Edit: Dreamsavior did in fact not fix something for once.
Last edited:
- Dec 29, 2017
- 2,709
- 4,899
Both the executable and the main package (rgss3a) are dated to be last modified on the 17.
1.1.4 did not exist yet.
Well since that's a bust for now what about this one?
https://f95zone.to/threads/shining-...rl-lily-rj01073233-studio-little-fish.180366/
Latest free version available on h sharing / mexashare.
I admit someone is supposed to be translating this but their last post / progress report was on Dec 14 and was last seen online Dec 31 with no news whatsoever until now. So yay or nay?
https://f95zone.to/threads/shining-...rl-lily-rj01073233-studio-little-fish.180366/
Latest free version available on h sharing / mexashare.
I admit someone is supposed to be translating this but their last post / progress report was on Dec 14 and was last seen online Dec 31 with no news whatsoever until now. So yay or nay?
- Dec 29, 2017
- 2,709
- 4,899
As a general guideline it would probably be best to wait at least 3 months since his last post about it,
then ask openly in the thread (so people waiting can see the post) whether or not he's still working on it, and if there's no reply within 2 weeks consider it dead.
That would at least reduce drama.
Alright I'll wait for til end of March / early April and hope he's back by then. Thanks for the response Shisaye.
Sorry for double posting but should have a doable request hopefully.
https://f95zone.to/threads/translation-request-iolite-lutnak-quest-v-1-7-rj01066434.168165/
Latest free version available on h sharing.
https://f95zone.to/threads/translation-request-iolite-lutnak-quest-v-1-7-rj01066434.168165/
Latest free version available on h sharing.
- Dec 29, 2017
- 2,709
- 4,899
I started working on RPGM because there was a certain game I really cared about, and I couldn't understand why nobody else did and why it was never localized to the west.
It turned out the reason was, that I just have odd taste and nobody actually really liked the game, but I still wanted to translate it.
I just made a pretty bad Google MTL, but the process of debugging it taught me the basics, and that case is also where the "working on hopeless games" idea came from.
Since I already knew the basics I started to make other Google MTLs for stuff that somewhat interested me, but it kept coming up that my Google stuff was "pointless because it's just like using MTooL", which created a "fun" rivalry in the sense that I was determined to create something better.
First with the work intensive DMTL format and then transitioning into my current advanced SLR system.
At this point MTooL is a shadow of it's former self and instead of improving, the free version actually got worse to the point that my Google MTLs would actually be better now, so that drive disappeared.
On top of that I don't actually play any of these games anymore. I'm completely burned out on the RPGM format.
But I've now built this elaborate setup over the last 2 years and It just seems like such a waste to just scrap it.
Just then for someone to have to remake it if they find themselves in the same situation I was in back then.
(Really liking a game, but nobody else does.)
Essentially I have accumulated so much otherwise useless knowledge and written these giant scripts, that without wanting to sound too arrogant, I'm kinda hard to replace in this space.
Sure there's dazedanon for example, but since he's using ChatGPT he has to charge a bunch of money for each translation, which just doesn't work for most games.
So I'm stuck. lol
- Dec 29, 2017
- 2,709
- 4,899
- Dec 29, 2017
- 2,709
- 4,899
This thread is about RPGM games.
I don't have any automation for unity.
As far as I am aware the only currently available somewhat working way of MTLing unity games is the XUnity.Autotranslator plugins by Bepis. I cannot write you a tutorial on how to use them because I haven't worked with unity, yet.
- Jan 12, 2020
- 68
- 215
The level of affinity I have for your cause cannot be understated. Unfortunately, I am broke.
Have you considered open sourcing your work so that other people can extend it? If the problem is this or that parser is just bad, then having other people able to look at it certainly couldn't hurt. In addition, closed source software just dies when the developer stops supporting it, and I would hate to see all the work you have put into your system for the last few years go unused when I am extremely interested in it.
For backend translation engines, in addition to dazed's paid GPT model, there are also freely available models that can be used locally through koboldcpp and other interfacing software. These can consider context as well so their results on the
You must be registered to see the links
should not be surprising. They can be better than GPT 3.5 and close to 4.0.And of course there is also DeepL. DeepL in particular consumes minimal local resources and has a beta feature where they are allowing arbitrary context entered by the user to influence their translations. DeepL has only been an NMT, like Sugoi, for a long time but they seem to be integrating AI technology in order to stay relevant as AI translation takes over.
It would not be unreasonable to ask people who make game requests to go sign up for a free DeepL account and lend you their free API key if they want the game translated. Is there a reason not to do this?
- Dec 29, 2017
- 2,709
- 4,899
I am actually trying to get it into a usable state for other people, but it's just one of those cases in which things that seem obvious to me aren't for someone else (It's a giant, inefficient mess).
That's why I am working on a modified Translator++ using my system in order to make it make sense. I'm also not planning to encrypt any of it so you can just look at the raw code and edit it how you like.
But the issue is that the standard parsers Dreamsavior uses are pretty bad (At least the ones I have and nobody was willing to provide me the newer ones, yet.), but they are required to be able to use the context system. (To be able to tag cells properly, unless you remake the entire system.)
So for now I can either try to include my properly working parser, with no guidance whatsoever, which basically guarantees that stuff will break because people don't have the experience what, where, can be translated how.
Or I use dreamsavior's parsers, have everything automatically tagged and formatted using a modified version of Reddo9999's tag system, but then some stuff breaks because the parser fucks up the VX Ace hash values.
It already actually works almost flawlessly for MV though so if you want to play around with the current prototype, I'll send it to you.
You'd have to ask dazed that.
My Sugoi/Dictionary system is fully offline and produces more consistent results than DeepL at least.
- Jan 12, 2020
- 68
- 215
What do you mean by context system?
For parsing, I had the same exact issue with the Translator++ parsers being sub-optimal for translation.
Sugoi in particular would often produce a result comparable to Google when parsed properly, but the Translator++ parsers would often break lines into very small chunks that made it impossible for Sugoi to tell what the context is. That meant going back to the raw source and manually rebuilding the untranslated string to get Sugoi to produce usable output because the parser broke it into unusable little chunks.
Instead of getting more into Translator++, like you have, I ran the other way and chose to only use it as an external tool instead of getting involved in its internals. The GPL license allows to be forked and distributed (minus the branding), but unless someone else is going to make it their full time project, then it is not really worth doing so because the project is under active development.
Yes, please! If I can use it on an MV/MZ game to see how it works, then I can better understand how it should be working for VX Ace and can dive into the internals to see if it is a fixable issue or if the problem warrants a completely different approach.
What I meant was that I am not sure what your local hardware looks like.
Unless you train your own NMT models or otherwise update the sentencepieces model, Sugoi will never really improve on its own. The use of dictionaries to fix input/output is more of a band-aid fix than anything helping Sugoi improve its underlying quality. That drawback of being stale, not improving on its own over time, makes it worth looking into different approaches to translate text from Japanese or at least train your own models based upon the the model Sugoi uses or JPCrawl.
The other approaches available right now are to use Large Language Models, overt AI, which are a linear extension of NMT technology but can consider arbitrary context information when translating and are familiar with a wide range of topics, as opposed to being focused on translation. Basically, this is the GPT, and Meta's Large Language Model AI (LLaMA) however, it is now possible to run these AI models locally, fully offline just like Sugoi. The leaderboard I linked to lists several models able to work this way like Minstral8x7b, Tulu-2-DPO, and Yi-34B.
If your hardware is high-end enough to run Sugoi with GPU acceleration, then getting it to work with KoboldCPP and an LLM would likely also be doable so it might be worth doing a quality comparison check. There are also LoRAs for text-to-text models which are like extensions to a base AI model that allows that models to specialize in a particular task, like translation. I have not found any LoRAs suitable for JPN->ENG translation but I also have not looked very hard and, due to recent advances in the field, it is possible to train LoRAs on consumer grade hardware. But again, I am not sure what your local hardware situation looks like.
If it is low-end enough that running KoboldCPP is not really viable, then DeepL would be the better option to look into because DeepL is also allowing context and other AI-like features to be used with their API for free, although they might start charging for it later.
For your dictionaries, I am not sure what format they are in, but all of your current dictionaries should be forward-compatible with both KoboldCPP and DeepL meaning that you can benefit from creating them and the work that was put into them even if switching translation engines. If they are not forward compatible right now, then it would be trivial to write a parser to convert them into a format usable by DeepL or in a more generic spreadsheet format to allow them to be used with KoboldCPP or another LLM interface.