<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Building DBRX-class Custom LLMs with Mosaic AI Training in Announcements</title>
    <link>https://community.databricks.com/t5/announcements/building-dbrx-class-custom-llms-with-mosaic-ai-training/m-p/69265#M96</link>
    <description>&lt;P&gt;We recently introduced&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm" target="_blank" rel="nofollow noopener noreferrer"&gt;DBRX&lt;/A&gt;: an open, state-of-the-art, general-purpose LLM. DBRX was trained, fine-tuned, and evaluated using Mosaic AI Training, scaling training to 3072 NVIDIA H100s and processing more than 12 trillion tokens in the process.&lt;/P&gt;
&lt;P&gt;Training LLMs, and in particular MoE models such as DBRX, is hard. It requires overcoming many infrastructure, performance, and scientific challenges. Mosaic AI Training was purposely built to address these challenges and was battle-tested through the training of&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm" target="_blank" rel="nofollow noopener noreferrer"&gt;DBRX&lt;/A&gt;, the&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://www.databricks.com/blog/mpt-30b" target="_blank" rel="nofollow noopener noreferrer"&gt;MPT series of models&lt;/A&gt;, and many other LLMs such as&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://blog.olakrutrim.com/public-rollout-krutrim/" target="_blank" rel="nofollow noopener noreferrer"&gt;Ola’s Krutrim&lt;/A&gt;,&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://blog.allenai.org/hello-olmo-a-truly-open-llm-43f7e7359222" target="_blank" rel="nofollow noopener noreferrer"&gt;AI2’s OLMo&lt;/A&gt;,&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://dynamo.ai/blog/dynamo-8b-a-multilingual-foundation-for-global-enterprises" target="_blank" rel="nofollow noopener noreferrer"&gt;Dynamo AI’s Dynamo 8B&lt;/A&gt;,&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://www.refuel.ai/blog-posts/announcing-refuel-llm-2" target="_blank" rel="nofollow noopener noreferrer"&gt;Refuel’s LLM-2&lt;/A&gt;, and others.&amp;nbsp;&lt;BR /&gt;&lt;BR /&gt;&lt;STRONG&gt;&lt;A href="https://www.databricks.com/blog/mosaic-ai-training-capabilities?utm_source=bambu&amp;amp;utm_medium=social&amp;amp;utm_campaign=advocacy" target="_blank" rel="noopener"&gt;Continue to read more here.&lt;/A&gt;&lt;/STRONG&gt;&lt;/P&gt;</description>
    <pubDate>Fri, 17 May 2024 13:03:24 GMT</pubDate>
    <dc:creator>Sujitha</dc:creator>
    <dc:date>2024-05-17T13:03:24Z</dc:date>
    <item>
      <title>Building DBRX-class Custom LLMs with Mosaic AI Training</title>
      <link>https://community.databricks.com/t5/announcements/building-dbrx-class-custom-llms-with-mosaic-ai-training/m-p/69265#M96</link>
      <description>&lt;P&gt;We recently introduced&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm" target="_blank" rel="nofollow noopener noreferrer"&gt;DBRX&lt;/A&gt;: an open, state-of-the-art, general-purpose LLM. DBRX was trained, fine-tuned, and evaluated using Mosaic AI Training, scaling training to 3072 NVIDIA H100s and processing more than 12 trillion tokens in the process.&lt;/P&gt;
&lt;P&gt;Training LLMs, and in particular MoE models such as DBRX, is hard. It requires overcoming many infrastructure, performance, and scientific challenges. Mosaic AI Training was purposely built to address these challenges and was battle-tested through the training of&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm" target="_blank" rel="nofollow noopener noreferrer"&gt;DBRX&lt;/A&gt;, the&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://www.databricks.com/blog/mpt-30b" target="_blank" rel="nofollow noopener noreferrer"&gt;MPT series of models&lt;/A&gt;, and many other LLMs such as&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://blog.olakrutrim.com/public-rollout-krutrim/" target="_blank" rel="nofollow noopener noreferrer"&gt;Ola’s Krutrim&lt;/A&gt;,&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://blog.allenai.org/hello-olmo-a-truly-open-llm-43f7e7359222" target="_blank" rel="nofollow noopener noreferrer"&gt;AI2’s OLMo&lt;/A&gt;,&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://dynamo.ai/blog/dynamo-8b-a-multilingual-foundation-for-global-enterprises" target="_blank" rel="nofollow noopener noreferrer"&gt;Dynamo AI’s Dynamo 8B&lt;/A&gt;,&lt;SPAN&gt;&amp;nbsp;&lt;/SPAN&gt;&lt;A href="https://www.refuel.ai/blog-posts/announcing-refuel-llm-2" target="_blank" rel="nofollow noopener noreferrer"&gt;Refuel’s LLM-2&lt;/A&gt;, and others.&amp;nbsp;&lt;BR /&gt;&lt;BR /&gt;&lt;STRONG&gt;&lt;A href="https://www.databricks.com/blog/mosaic-ai-training-capabilities?utm_source=bambu&amp;amp;utm_medium=social&amp;amp;utm_campaign=advocacy" target="_blank" rel="noopener"&gt;Continue to read more here.&lt;/A&gt;&lt;/STRONG&gt;&lt;/P&gt;</description>
      <pubDate>Fri, 17 May 2024 13:03:24 GMT</pubDate>
      <guid>https://community.databricks.com/t5/announcements/building-dbrx-class-custom-llms-with-mosaic-ai-training/m-p/69265#M96</guid>
      <dc:creator>Sujitha</dc:creator>
      <dc:date>2024-05-17T13:03:24Z</dc:date>
    </item>
    <item>
      <title>Re: Building DBRX-class Custom LLMs with Mosaic AI Training</title>
      <link>https://community.databricks.com/t5/announcements/building-dbrx-class-custom-llms-with-mosaic-ai-training/m-p/73826#M128</link>
      <description>&lt;P&gt;&lt;STRONG&gt;exciting!&lt;/STRONG&gt;&lt;/P&gt;</description>
      <pubDate>Thu, 13 Jun 2024 17:51:26 GMT</pubDate>
      <guid>https://community.databricks.com/t5/announcements/building-dbrx-class-custom-llms-with-mosaic-ai-training/m-p/73826#M128</guid>
      <dc:creator>Insurance</dc:creator>
      <dc:date>2024-06-13T17:51:26Z</dc:date>
    </item>
  </channel>
</rss>

