{"id":21338,"date":"2023-08-01T06:27:17","date_gmt":"2023-08-01T04:27:17","guid":{"rendered":"https:\/\/www.graviton.at\/letterswaplibrary\/llm-training-with-php-improved-using-txt-datasets\/"},"modified":"2023-08-01T06:27:17","modified_gmt":"2023-08-01T04:27:17","slug":"llm-training-with-php-improved-using-txt-datasets","status":"publish","type":"post","link":"https:\/\/www.graviton.at\/letterswaplibrary\/llm-training-with-php-improved-using-txt-datasets\/","title":{"rendered":"LLM Training With PHP Improved Using Txt Datasets!"},"content":{"rendered":"

<\/p>\n

\n

Hi guys how are you doing?
last week I share my first version of this simple Languaje model training with php.<\/p>\n

For thoose who missed, it use a simple Markov Chain for calculate the probabilities for the next word based on the previous words.<\/p>\n

Now I have improved the training dataset and the next word selector.<\/p>\n

Here’s is the link:<\/p>\n

https:\/\/github.com\/AcidBurn86\/LM-nGram-with-php\/<\/a><\/p>\n

is a good way to start understand how big LLM works. And of course I know this could never perform like GPT or Llama.<\/p>\n

Is just an educational code for php fans.<\/p>\n

\u200b<\/p>\n

Shares and github stars are welcome!<\/p>\n

\u200b<\/p>\n<\/div>\n

submitted by \/u\/OficialPimento <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>