101010.pl

Chi Kim😲 DeepSeek-V3-4bit runs at >20 tokens per second and <200W using MLX on an M3 Ultra with 512GB. This might be the best and most user-friendly way to run DeepSeek-V3 on consumer hardware, possibly the most affordable too. You can finally run a GPT-4o level model locally, with possibly even better quality. <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#LLM</a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://mastodon.social/tags/ML" class="mention hashtag" rel="nofollow noopener" target="_blank">#ML</a> <a href="https://mastodon.social/tags/DeepSeek" class="mention hashtag" rel="nofollow noopener" target="_blank">#DeepSeek</a> <a href="https://mastodon.social/tags/OpenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#OpenAI</a> <a href="https://mastodon.social/tags/GPT" class="mention hashtag" rel="nofollow noopener" target="_blank">#GPT</a> <a href="https://mastodon.social/tags/OpenWeight" class="mention hashtag" rel="nofollow noopener" target="_blank">#OpenWeight</a> <a href="https://mastodon.social/tags/OpenSource" class="mention hashtag" rel="nofollow noopener" target="_blank">#OpenSource</a> <a href="https://venturebeat.com/ai/deepseek-v3-now-runs-at-20-tokens-per-second-on-mac-studio-and-thats-a-nightmare-for-openai/" rel="nofollow noopener" translate="no" target="_blank">https://venturebeat.com/ai/deepseek-v3-now-runs-at-20-tokens-per-second-on-mac-studio-and-thats-a-nightmare-for-openai/</a>

Stefano ZacchiroliThis week I'm in <a href="https://mastodon.xyz/tags/Montreal" class="mention hashtag" rel="nofollow noopener" target="_blank">#Montreal</a> to present at <a href="https://mastodon.xyz/tags/SANER2025" class="mention hashtag" rel="nofollow noopener" target="_blank">#SANER2025</a> recent joint work with A. Gurioli and M. Gabrielli, on how to recognize AI-generated code. Key differences in our results wrt the state-of-the-art are: (1) a multilingual approach that supports 10 different programming languages, and (2) a fully <a href="https://mastodon.xyz/tags/reproducible" class="mention hashtag" rel="nofollow noopener" target="_blank">#reproducible</a> pipeline based on both <a href="https://mastodon.xyz/tags/OpenWeight" class="mention hashtag" rel="nofollow noopener" target="_blank">#OpenWeight</a> and <a href="https://mastodon.xyz/tags/OpenData" class="mention hashtag" rel="nofollow noopener" target="_blank">#OpenData</a> self-hosted LLMs. Talk info and <a href="https://mastodon.xyz/tags/OpenAccess" class="mention hashtag" rel="nofollow noopener" target="_blank">#OpenAccess</a> preprint at: <a href="https://conf.researchr.org/details/saner-2025/saner-2025-papers/36/Is-This-You-LLM-Recognizing-AI-written-Programs-with-Multilingual-Code-Stylometry" rel="nofollow noopener" translate="no" target="_blank">https://conf.researchr.org/details/saner-2025/saner-2025-papers/36/Is-This-You-LLM-Recognizing-AI-written-Programs-with-Multilingual-Code-Stylometry</a>

Prem Kumar Aparanji 👶🤖🐘<a href="https://mastodon.de/@katzenberger" class="u-url mention" rel="nofollow noopener" target="_blank">@katzenberger</a> <a href="https://mastodon.social/@silentexception" class="u-url mention" rel="nofollow noopener" target="_blank">@silentexception</a> So, from that perspective, it is important that: 1. We find the appropriate job/task for which the <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#LLM</a> is indispensible, and not use it for everything 2. We try to reuse the LLM as much as possible rather than training again and again<a href="https://mastodon.social/tags/opensource" class="mention hashtag" rel="nofollow noopener" target="_blank">#opensource</a> <a href="https://mastodon.social/tags/openweight" class="mention hashtag" rel="nofollow noopener" target="_blank">#openweight</a> <a href="https://mastodon.social/tags/ai" class="mention hashtag" rel="nofollow noopener" target="_blank">#ai</a> <a href="https://mastodon.social/tags/jtbd" class="mention hashtag" rel="nofollow noopener" target="_blank">#jtbd</a> <a href="https://mastodon.social/tags/economics" class="mention hashtag" rel="nofollow noopener" target="_blank">#economics</a> <a href="https://mastodon.social/tags/foss" class="mention hashtag" rel="nofollow noopener" target="_blank">#foss</a> <a href="https://mastodon.social/tags/genai" class="mention hashtag" rel="nofollow noopener" target="_blank">#genai</a>

Prem Kumar Aparanji 👶🤖🐘<a href="https://mastodon.social/@silentexception" class="u-url mention" rel="nofollow noopener" target="_blank">@silentexception</a> ha ha! Yeah ... Well, it's a perennial cycle ... Isn't it.Wait till the large lying models cause enough troubles for enterprises and various Sovereign AI.<a href="https://mastodon.social/tags/llm" class="mention hashtag" rel="nofollow noopener" target="_blank">#llm</a> <a href="https://mastodon.social/tags/llms" class="mention hashtag" rel="nofollow noopener" target="_blank">#llms</a> <a href="https://mastodon.social/tags/GenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#GenAI</a> <a href="https://mastodon.social/tags/opensource" class="mention hashtag" rel="nofollow noopener" target="_blank">#opensource</a> <a href="https://mastodon.social/tags/openweight" class="mention hashtag" rel="nofollow noopener" target="_blank">#openweight</a>

Prem Kumar Aparanji 👶🤖🐘<a href="https://witter.cz/@fredbrooker" class="u-url mention" rel="nofollow noopener" target="_blank">@fredbrooker</a> not even the 3 trillion tokens worth of data from dolma?Or, the more recent, 2 trillion tokens in common corpus?<a href="https://huggingface.co/blog/Pclanglais/two-trillion-tokens-open" rel="nofollow noopener" translate="no" target="_blank">https://huggingface.co/blog/Pclanglais/two-trillion-tokens-open</a><a href="https://mastodon.social/tags/opensource" class="mention hashtag" rel="nofollow noopener" target="_blank">#opensource</a> <a href="https://mastodon.social/tags/ai" class="mention hashtag" rel="nofollow noopener" target="_blank">#ai</a> <a href="https://mastodon.social/tags/genai" class="mention hashtag" rel="nofollow noopener" target="_blank">#genai</a> <a href="https://mastodon.social/tags/llm" class="mention hashtag" rel="nofollow noopener" target="_blank">#llm</a> <a href="https://mastodon.social/tags/llms" class="mention hashtag" rel="nofollow noopener" target="_blank">#llms</a> <a href="https://mastodon.social/tags/openweight" class="mention hashtag" rel="nofollow noopener" target="_blank">#openweight</a> <a href="https://mastodon.social/tags/openweights" class="mention hashtag" rel="nofollow noopener" target="_blank">#openweights</a>

Recent searches

Search options

Administered by:

Server stats:

#openweight