Feb 21, 2023 #9 Wasn't informed that a fine tuned product is harder to operate - will have to go through up on it as I have already been little by little collecting references regarding how to set a data set together with the intent of building something which may be used to fine tune.
TmpGuy mentioned: Why device translate to Chinese? Ultimately, the intention for most here is to translate to English. Now we have a lot of Chinese subs as they're those producing subtitles, not much Japanese, who have little will need for them (Most likely In addition to hearing impaired people).
This is certainly the first time I've tried using translating everything. I am wanting to study japanese and it absolutely was a fun method of getting some apply And that i acquired a whole lot. I made use of AI to tear a transcript and then went line by line comparing the hiragana to what was becoming mentioned to verify the same old MTL garble was not in it.
That is since there are no English subtitles for them. You will find hardly any people that translate JAV to English properly and just a few additional who equipment translate and add. I only deliver Chinese subtitles for anybody requesting translations.
condition_on_previous_text: Defaults to True, only other solution is False on the the time of the producing. This location encourages the design to know a particular topic, a selected model, or discern more dependable translations of suitable nouns. I strongly advise you utilize Phony when translating JAV. Considering the fact that JAV can have lots of parts which are tricky to detect and transcribe and translate (not adequate JAV of their schooling info), acquiring this established to Genuine leads to some locations the place a similar translation is applied line just after line, bringing about such a robust bias in direction of some line that the translation may possibly under no circumstances Get better.
three. To beat the final dilemma I parsed my here comprehensive map tree in two or three AI to talk to if it noticed any duplicates that have this problem. Deepseek, Chatgpt couldn't handle it, they just invented duplicate names. Claud.AI context window was much too compact (about 3000 actress names).
What about I generate a blogger were I'm able to collect my posts (EngSubs+ Raw) free without having ads or shit In lieu of losing a upcoming?
initial_prompt: You'll be able to enter a sentence or number of words and phrases being a string to try to bias the interpretation in some way. It's not at all distinct whether or not this is designed to applay to only the very first segment or your complete transcript, but in future variations it will most likely be somewhat a lot more distinct or adaptable.
Our Local community has been around for quite some time and pride ourselves on featuring unbiased, critical dialogue among people today of all different backgrounds. We're Operating everyday to make certain our Group is among the finest.
Or I need access to a different superior databases. It might also ensure it is easier to search for the actresses bring about I could just place a .txt named with all aliases inside the maps.
At some point there may be a whisper factor which can assess the video clip also to have proper context, but we are not there yet.
avatarthe stated: Discover how to fish, you’ll have additional fish than you’d get by just begging for fish.
I actually recognize these packs due to the fact OCRing subtitle has gotten more difficult in recent months. Thanks a great deal of for uploading these packs!
avatarthe explained: Figure out how to fish, you’ll have additional fish than you’d get by just begging for fish.