Kudasai

TXT file with Japanese Text

Replacements JSON file

Knowledge Base Single File

Knowledge Base Directory

Indexed text

Indexing Results

Debug Log

TXT file with Japanese Text

Replacements JSON file

Japanese Text

Preprocessed text

Preprocessing Results

Debug Log

TXT file with Japanese Text

Japanese Text

Translation Settings File

translation_settings .json

1.1 KB ⇣

Genders.json File

genders .json

41.0 B ⇣

Translation Method

API Key

Translated Text

JE Check Text

Debug Log

Base Translation Settings

These settings are used for OpenAI, Gemini, DeepL, and Google Translate

Please ensure to thoroughly read and understand these settings before making any modifications. Each setting has a specific impact on the translation methods. Some settings may affect one or two translation methods, but not the others. Incorrect adjustments could lead to unexpected results or errors in the translation process.

Prompt Assembly Mode

prompt_assembly_mode : 1 or 2. 1 means the system message will actually be treated as a system message. 2 means it'll be treated as a user message. 1 is recommend for gpt-4 otherwise either works. For Gemini & DeepL, this setting is ignored.

Number of Lines Per Batch

number_of_lines_per_batch : The number of lines to be built into a prompt at once. Theoretically, more lines would be more cost effective, but other complications may occur with higher lines. So far been tested up to 48 by me.

Sentence Fragmenter Mode

sentence_fragmenter_mode : 1 or 2 (1 - via regex and other nonsense) 2 - None (Takes formatting and text directly from API return)) the API can sometimes return a result on a single line, so this determines the way Kudasai fragments the sentences if at all. Use 2 for newer models and Deepl.

JE Check Mode

je_check_mode : 1 or 2, 1 will print out the jap then the english below separated by ---, 2 will attempt to pair the english and jap sentences, placing the jap above the eng. If it cannot, it will default to 1. Use 2 for newer models and DeepL.

Number Of Malformed Batch Retries

number_of_malformed_batch_retries : (Malformed batch is when je-fixing fails) How many times Kudasai will attempt to mend a malformed batch (mending is resending the request). Be careful with increasing as cost increases at (cost * length * n) at worst case. This setting is ignored if je_check_mode is set to 1.

Batch Retry Timeout

batch_retry_timeout : How long Kudasai will try to translate a batch in seconds, if a requests exceeds this duration, Kudasai will leave it untranslated.

Number Of Concurrent Batches

number_of_concurrent_batches : How many translations batches Kudasai will send to the translation API at a time. For OpenAI, be conservative as rate-limiting is aggressive, I'd suggest 3-5. For Gemini, do not exceed 15 for 1.0 or 2 for 1.5. This setting more or less doesn't matter for DeepL.

gender_context_insertion : true or false - Whether to insert gender info into system prompts when translating text. Kudasai will look for a "gender.json" (exactly) in the root directory and scan each text to translate for the names in the file. If a name is found, Kudasai will insert the gender info into the system prompt. For more info look at the README.md.

Gender Context Insertion

is_cote: true or false - Whether you are translating COTE (Classroom of the Elite), Kudasai has specialized functions for COTE that will be enabled if this is set to true.

Is Cote

OpenAI API Settings

See https://platform.openai.com/docs/api-reference/chat/create for further details

openai_stream, openai_logit_bias, openai_stop and openai_n are included for completion's sake, current versions of Kudasai will hardcode their values when validating the translation_settings.json to their default values. As different values for these settings do not have a use case in Kudasai's current implementation.

OpenAI Model

openai_model : ID of the model to use. Kudasai only works with 'chat' models.

OpenAI System Message

openai_system_message : Instructions to the model. Basically tells the model how to translate.

OpenAI Temperature

openai_temperature : What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. Lower values are typically better for translation.

0 2

OpenAI Top P

openai_top_p : An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. I generally recommend altering this or temperature but not both.

0 1

OpenAI N

openai_n : How many chat completion choices to generate for each input message. Do not change this, as Kudasai will always use 1.

OpenAI Stream

openai_stream : If set, partial message deltas will be sent, like in ChatGPT. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE] message. See the OpenAI python library on GitHub for example code. Do not change this as Kudasai does not support this feature.

OpenAI Stop

openai_stop : Up to 4 sequences where the API will stop generating further tokens. Do not change this as Kudasai does not support this feature.

OpenAI Logit Bias

openai_logit_bias : Modifies the likelihood of specified tokens appearing in the completion. Do not change this as Kudasai does not support this feature.

OpenAI Max Tokens

openai_max_tokens : The maximum number of tokens to generate in the chat completion. The total length of input tokens and generated tokens is limited by the model's context length. I wouldn't recommend changing this. Is none by default. If you change to an integer, make sure it doesn't exceed that model's context length or your request will fail and repeat till timeout.

OpenAI Presence Penalty

openai_presence_penalty : Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. While negative values encourage repetition. Should leave this at 0.0.

-2 2

OpenAI Frequency Penalty

openai_frequency_penalty : Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. Negative values encourage repetition. Should leave this at 0.0.

-2 2

Gemini API Settings

See https://ai.google.dev/docs/concepts#model-parameters for further details

gemini_stream, gemini_stop_sequences and gemini_candidate_count are included for completion's sake, current versions of Kudasai will hardcode their values when validating the translation_settings.json to their default values. As different values for these settings do not have a use case in Kudasai's current implementation.

Gemini Model

gemini_model : The model to use. Currently only supports gemini-pro and gemini-pro-vision, the 1.0 model and 1.5 models and their aliases.

Gemini Prompt

gemini_prompt : Instructions to the model. Basically tells the model how to translate.

Gemini Temperature

gemini_temperature : What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. Lower values are typically better for translation.

0 2

Gemini Top P

gemini_top_p : An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. I generally recommend altering this or temperature but not both.

Gemini Top K

gemini_top_k : Determines the number of most probable tokens to consider for each selection step. A higher value increases diversity, a lower value makes the output more deterministic.

Gemini Candidate Count

gemini_candidate_count : The number of candidates to generate for each input message. Do not change this as Kudasai will always use 1.

Gemini Stream

gemini_stream : If set, partial message deltas will be sent, like in Gemini Chat. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE] message. Do not change this as Kudasai does not support this feature.

Gemini Stop Sequences

gemini_stop_sequences : Up to 4 sequences where the API will stop generating further tokens. Do not change this as Kudasai does not support this feature.

Gemini Max Output Tokens

gemini_max_output_tokens : The maximum number of tokens to generate in the chat completion. The total length of input tokens and generated tokens is limited by the model's context length. I wouldn't recommend changing this. Is none by default. If you change to an integer, make sure it doesn't exceed that model's context length or your request will fail and repeat till timeout.

DeepL API Settings

See https://developers.deepl.com/docs/api-reference/translate for further details

DeepL API settings are not as extensive as OpenAI and Gemini, a lot of the settings are simply not included for Kudasai as they do not have a decent use case to warrant their inclusion. Settings may be added in the future if a use case is found or is suggested.

DeepL Context

deepl_context : The context in which the text should be translated. This is used to improve the translation. If you don't have any context, you can leave this empty. This is a DeepL Alpha feature and could be subject to change.

DeepL Split Sentences

deepl_split_sentences : How the text should be split into sentences. Possible values are 'OFF', 'ALL', 'NO_NEWLINES'.

deepl_preserve_formatting : Whether the formatting of the text should be preserved. If you don't want to preserve the formatting, you can set this to False. Otherwise, set it to True.

DeepL Preserve Formatting

DeepL Formality

deepl_formality : The formality of the text. Possible values are 'default', 'more', 'less', 'prefer_more', 'prefer_less'.

Debug Log

Error Log