4 ms·

Oh I found this https://raku.land/zef:antononcube/DSL::Examples https://raku.land/zef:antononcube/DSL::Examples So Raku has a module for “few shot” LLM trainin

Oh I found this https://raku.land/zef:antononcube/DSL::Examples https://raku.land/zef:antononcube/DSL::Examples So Raku has a module for “few shot” LLM training for DSLs … would be cool to see an example DSL interpreter in Selkie (eg a window for DSL code and a window for output…

apogee 6mo ago
Yeah I have thought about doing this with Roaring::Tags (https://raku.land/zef:apogee/Roaring::Tags https://raku.land/zef:apogee/Roaring::Tags) so you type a Roaring::Tags query on one side and it shows the built Raku on the other. Wouldn't be too difficult: - FileBrowser widget to get the serialised bitmap. - MultiLineInput for the query text input. - ListView to show tags and fields. - RichText for the AST pretty printer.
antononcube 6mo ago
There is a set of Raku modules that leverage LLMs for different tasks (mostly code generation) using different techniques: - https://raku.land/zef:antononcube/LLM::Resources https://raku.land/zef:antononcube/LLM::Resources : Uses agentic LLM-graphs with asynchronous execution - https://raku.land/zef:antononcube/ML::FindTextualAnswer https://raku.land/zef:antononcube/ML::FindTextualAnswer : Finds answers to questions over provided texts (e.g. natural language code generation commands) - https://raku.land/zef:antononcube/ML::NLPTemplateEngine https://raku.land/zef:antononcube/ML::NLPTemplateEngine : Fills-in predefined code templates based on natural language code descriptions/commands - https://raku.land/zef:antononcube/DSL::Examples https://raku.land/zef:antononcube/DSL::Examples : Example translations of natural language commands to executable code
- apogee 6mo ago
  I've got a few LLM modules too, mostly for handling context management: - https://raku.land/zef:apogee/LLM::Character https://raku.land/zef:apogee/LLM::Character implements CCv3 which is a standard for managing characters (system prompts) and lorebooks (injected snippets) - https://raku.land/zef:apogee/LLM::Chat https://raku.land/zef:apogee/LLM::Chat handles context shifting for long contexts, sampler settings, templating for text completion & inferencing with or without streaming using supply/tap - https://raku.land/zef:apogee/LLM::Data::Inference https://raku.land/zef:apogee/LLM::Data::Inference adds retries, JSON parsing & multi-model route handling to LLM::Chat - https://raku.land/zef:apogee/LLM::Data::Pipeline https://raku.land/zef:apogee/LLM::Data::Pipeline allows you to declaratively build multi-step pipelines (simple agentic LLM use) - https://raku.land/zef:apogee/HuggingFace::API https://raku.land/zef:apogee/HuggingFace::API is a partial wrapper around HF API for grabbing tokenizers.json & tokenizer_config.json - https://raku.land/zef:apogee/Template::Jinja2 https://raku.land/zef:apogee/Template::Jinja2 is a near-complete impl of Jinja2 for parsing LLM text completion templates (can be used for anything you'd use Jinja2 for) - https://raku.land/zef:apogee/Tokenizers https://raku.land/zef:apogee/Tokenizers is a thin wrapper around HF tokenizers, for token counting mostly

apogee 6mo ago

Yeah I have thought about doing this with Roaring::Tags (https://raku.land/zef:apogee/Roaring::Tags https://raku.land/zef:apogee/Roaring::Tags) so you type a Roaring::Tags query on one side and it shows the built Raku on the other. Wouldn't be too difficult: - FileBrowser widget to get the serialised bitmap. - MultiLineInput for the query text input. - ListView to show tags and fields. - RichText for the AST pretty printer.

antononcube 6mo ago

There is a set of Raku modules that leverage LLMs for different tasks (mostly code generation) using different techniques: - https://raku.land/zef:antononcube/LLM::Resources https://raku.land/zef:antononcube/LLM::Resources : Uses agentic LLM-graphs with asynchronous execution - https://raku.land/zef:antononcube/ML::FindTextualAnswer https://raku.land/zef:antononcube/ML::FindTextualAnswer : Finds answers to questions over provided texts (e.g. natural language code generation commands) - https://raku.land/zef:antononcube/ML::NLPTemplateEngine https://raku.land/zef:antononcube/ML::NLPTemplateEngine : Fills-in predefined code templates based on natural language code descriptions/commands - https://raku.land/zef:antononcube/DSL::Examples https://raku.land/zef:antononcube/DSL::Examples : Example translations of natural language commands to executable code

apogee 6mo ago
I've got a few LLM modules too, mostly for handling context management: - https://raku.land/zef:apogee/LLM::Character https://raku.land/zef:apogee/LLM::Character implements CCv3 which is a standard for managing characters (system prompts) and lorebooks (injected snippets) - https://raku.land/zef:apogee/LLM::Chat https://raku.land/zef:apogee/LLM::Chat handles context shifting for long contexts, sampler settings, templating for text completion & inferencing with or without streaming using supply/tap - https://raku.land/zef:apogee/LLM::Data::Inference https://raku.land/zef:apogee/LLM::Data::Inference adds retries, JSON parsing & multi-model route handling to LLM::Chat - https://raku.land/zef:apogee/LLM::Data::Pipeline https://raku.land/zef:apogee/LLM::Data::Pipeline allows you to declaratively build multi-step pipelines (simple agentic LLM use) - https://raku.land/zef:apogee/HuggingFace::API https://raku.land/zef:apogee/HuggingFace::API is a partial wrapper around HF API for grabbing tokenizers.json & tokenizer_config.json - https://raku.land/zef:apogee/Template::Jinja2 https://raku.land/zef:apogee/Template::Jinja2 is a near-complete impl of Jinja2 for parsing LLM text completion templates (can be used for anything you'd use Jinja2 for) - https://raku.land/zef:apogee/Tokenizers https://raku.land/zef:apogee/Tokenizers is a thin wrapper around HF tokenizers, for token counting mostly