{"id":30273,"date":"2024-09-03T07:27:13","date_gmt":"2024-09-03T05:27:13","guid":{"rendered":"https:\/\/www.graviton.at\/letterswaplibrary\/q-fine-tuning-coding-llms-on-github-histories-rather-than-just-final-code\/"},"modified":"2024-09-03T07:27:13","modified_gmt":"2024-09-03T05:27:13","slug":"q-fine-tuning-coding-llms-on-github-histories-rather-than-just-final-code","status":"publish","type":"post","link":"https:\/\/www.graviton.at\/letterswaplibrary\/q-fine-tuning-coding-llms-on-github-histories-rather-than-just-final-code\/","title":{"rendered":"Q: Fine-tuning Coding LLMs On Git[hub] Histories Rather Than Just Final Code?"},"content":{"rendered":"<p><!-- SC_OFF --><\/p>\n<div class=\"md\">\n<p>I run a small software company creating traditional C++ desktop apps for font &amp; graphic design work. We have 10+ years of Git histories of our apps.<\/p>\n<p>What open &#8220;coding&#8221; LLMs are there out there that weren&#8217;t just trained on final code but on Git histories (commits &amp; pull requests), and Github stuff (PR discussions, issues etc.)? <\/p>\n<p>What dataset formats for such data would be advisable to use? <\/p>\n<p>I&#8217;d like to fine-tune a coding LLM to privately assist in our software development, ideally not just on the current state of the code but on its evolution. <\/p>\n<p>I have a &#8220;feeling&#8221; that this would be much better. \ud83d\ude42 <\/p>\n<\/div>\n<p><!-- SC_ON -->   submitted by   <a href=\"https:\/\/www.reddit.com\/user\/Minimum_Art_2263\"> \/u\/Minimum_Art_2263 <\/a> <br \/> <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1f7e8eq\/q_finetuning_coding_llms_on_github_histories\/\">[link]<\/a><\/span>   <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1f7e8eq\/q_finetuning_coding_llms_on_github_histories\/\">[comments]<\/a><\/span><\/p><div class='watch-action'><div class='watch-position align-right'><div class='action-like'><a class='lbg-style1 like-30273 jlk' href='javascript:void(0)' data-task='like' data-post_id='30273' data-nonce='65e0e39b87' rel='nofollow'><img class='wti-pixel' src='https:\/\/www.graviton.at\/letterswaplibrary\/wp-content\/plugins\/wti-like-post\/images\/pixel.gif' title='Like' \/><span class='lc-30273 lc'>0<\/span><\/a><\/div><\/div> <div class='status-30273 status align-right'><\/div><\/div><div class='wti-clear'><\/div>","protected":false},"excerpt":{"rendered":"<p>I run a small software company creating traditional C++ desktop apps for font &amp; graphic design work&#8230;.<\/p>\n","protected":false},"author":0,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[85],"tags":[],"class_list":["post-30273","post","type-post","status-publish","format-standard","hentry","category-datatards","wpcat-85-id"],"_links":{"self":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/30273","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/comments?post=30273"}],"version-history":[{"count":0,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/30273\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/media?parent=30273"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/categories?post=30273"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/tags?post=30273"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}