mstdn.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A general-purpose Mastodon server with a 500 character limit. All languages are welcome.

Administered by:

Server stats:

13K
active users

#santacoder

0 posts0 participants0 posts today
Chuck Darwin<p><a href="https://c.im/tags/BigCode" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>BigCode</span></a> is an open scientific collaboration working on responsible training of large language models for coding applications. </p><p>In this organization you can find the artefacts of this collaboration:<br>👉 <a href="https://c.im/tags/StarCoder" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>StarCoder</span></a>, a state-of-the-art language model for code, <br>👉 The <a href="https://c.im/tags/Stack" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Stack</span></a>, the largest available pretraining dataset with perimssive code, and 👉 <a href="https://c.im/tags/SantaCoder" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>SantaCoder</span></a>, a 1.1B parameter model for code.</p><p><a href="https://c.im/tags/StarCoder" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>StarCoder</span></a> is a 15.5B parameters language model for code trained for 1T tokens on 80+ programming languages. <br>It uses MQA for efficient generation, has 8,192 tokens context window and can do fill-in-the-middle.</p><p>Chat with StarCoder here: <a href="https://huggingface.co/chat/?model=bigcode/starcoder" rel="nofollow noopener noreferrer" target="_blank"><span class="invisible">https://</span><span class="ellipsis">huggingface.co/chat/?model=big</span><span class="invisible">code/starcoder</span></a></p><p><a href="https://huggingface.co/bigcode" rel="nofollow noopener noreferrer" target="_blank"><span class="invisible">https://</span><span class="">huggingface.co/bigcode</span><span class="invisible"></span></a></p>
Alex Chaffee<p>There are quite a few code-generating “AI” systems now — GitHub <a href="https://ruby.social/tags/CoPilot" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CoPilot</span></a>, Amazon <a href="https://ruby.social/tags/CodeWhisperer" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>CodeWhisperer</span></a>, BigCode <a href="https://ruby.social/tags/SantaCoder" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>SantaCoder</span></a>, Facebook <a href="https://ruby.social/tags/Incoder" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Incoder</span></a>, maybe even more.</p><p>I wonder how hard it would be to get these BS bots to play <a href="https://ruby.social/tags/TDD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>TDD</span></a> ping-pong… I write a test, then they generate code until all tests pass, then i refactor, then we repeat.</p>
Vincent HETRU<p><a href="https://sigmoid.social/tags/HuggingFace" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>HuggingFace</span></a> just released the <a href="https://sigmoid.social/tags/SantaCoder" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>SantaCoder</span></a> models for the holiday season. Part of the <a href="https://sigmoid.social/tags/BigCode" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>BigCode</span></a> project, these 1.1B parameter models are trained on <a href="https://sigmoid.social/tags/Python" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Python</span></a>, <a href="https://sigmoid.social/tags/Java" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>Java</span></a>, and <a href="https://sigmoid.social/tags/JavaScript" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>JavaScript</span></a> and use advanced techniques like near-deduplication and comment-to-code ratio. </p><p><a href="https://huggingface.co/bigcode/santacoder" rel="nofollow noopener noreferrer" target="_blank"><span class="invisible">https://</span><span class="ellipsis">huggingface.co/bigcode/santaco</span><span class="invisible">der</span></a></p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/DeepLearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#<span>DeepLearning</span></a> 🤗</p>