Introducing Instella: New State-of-the-art Fully Open 3B Language Models — ROCm Blogs
rocm.blogs.amd.com
AMD is excited to announce Instella, a family of fully open state-of-the-art 3-billion-parameter language models (LMs). , In this blog we explain how the Instella models were trained, and how to access them.

This is again a big win on the red team at least for me. They developed a “fully open” 3B parameters model family trained from scratch on AMD Instinct™ MI300X GPUs.

AMD is excited to announce Instella, a family of fully open state-of-the-art 3-billion-parameter language models (LMs) […]. Instella models outperform existing fully open models of similar sizes and achieve competitive performance compared to state-of-the-art open-weight models such as Llama-3.2-3B, Gemma-2-2B, and Qwen-2.5-3B […].

As shown in this image (https://rocm.blogs.amd.com/_images/scaling_perf_instruct.png) this model outperforms current other “fully open” models, coming next to open weight only models.

A step further, thank you AMD.

PS : not doing AMD propaganda but thanks them to help and contribute to the Open Source World.

  • sugar_in_your_tea@sh.itjust.worksEnglish
    2·
    5 hours ago

    The issues are primarily social, not technical.

    Right, and having a FOSS alternative is certainly a good thing.

    I think it’s important to separate opposition to AI policy from a specific implementation. If your concerns are related to the social impact of a given technology, that is where the opposition should go, not toward the technology itself.

    That said, this is largely similar to opposition to other types of technological change. Every time a significant change in technology comes about, there is a significant impact to jobs. The printing press destroyed the livelihood of scribes, but it made books dramatically cheaper, which created new jobs for typesetters, booksellers, etc. The automobile dramatically cut back jobs like farriers, stable hands, etc, but created new jobs for drivers, mechanics, etc. I’m sure each of those large shifts in technology also had an overreaction by business owners as they adjusted to the new normal. It certainly sucks for those impacted, but it tends to benefit those who can quickly adapt and make use of the new technology.

    So I totally understand the hesitation around AI, especially given the overreaction by C-suites in gutting their workforce based on the promises made by AI marketing teams. However, that has nothing to do with the technology, but the social issues around the technology. Instead of hating AI in general, redirect that anger onto the actual problems:

    • poor social safety net
    • expensive education
    • lack of consequences for false marketing
    • lack of consequences for C-suite mistakes

    Hating on a FOSS model just because it’s related to an industry that is seeing abuse is the wrong approach.

    • frezik@midwest.socialEnglish
      0·
      5 hours ago

      Was there anything in the posts above mine that suggest this was a technical issue, or did you read that in as an assumption?

      Every time a significant change in technology comes about, there is a significant impact to jobs. The printing press destroyed the livelihood of scribes, but it made books dramatically cheaper, which created new jobs for typesetters, booksellers, etc.

      Take a look at the history of the first people called “Luddites”. They were early socialists focusing on the dismal working conditions that new automation would bring to the workers. And they were correct.

      Not every technological change is good. Our society has defaulted to saying yes to every change, and it’s caused a whole lot of problems.

      • sugar_in_your_tea@sh.itjust.worksEnglish
        1·
        3 hours ago

        Was there anything in the posts above mine that suggest this was a technical issue, or did you read that in as an assumption?

        I was responding both to you and to the parent to your comment and making it clear that it’s not a technical issue. I’m agreeing with you.

        And they were correct.

        I disagree.

        Yes, not every technological change is good, we can look at Social Media as a shining example of that. However, technological change is usually inevitable, especially if you value freedom in your society, so it’s a lot better to solve the issues that surround it than ban it.

        • frezik@midwest.socialEnglish
          1·
          3 hours ago

          There is absolutely nothing inevitable about technological change. We think that way because of the specific place we are in history. A specific place that is an aberration in how fast those changes have come. For the most part, humans throughout history have used much the same techniques and tools that their parents did.

          You also can’t separate AI technology from the social change. They’re not dumping billions into data centers and talking about using entire nuclear reactors to power them just because they think AI is a fun toy.

          • sugar_in_your_tea@sh.itjust.worksEnglish
            1·
            54 minutes ago

            A specific place that is an aberration in how fast those changes have come

            That’s really hard to quantify, but yeah, innovation is probably happening faster today than it has in the past, which is likely due to:

            • increased connectivity - more people have access to advanced technology
            • lower barriers to trade - despite Trump’s best efforts, trade/competition between countries still happens
            • better access to education

            People generally fear change, and change comes with work. Just because you were screwing on toothpaste caps in a factory yesterday doesn’t mean that job will make sense forever. Nor should it. Jobs that don’t need to be done by humans shouldn’t, and people should instead take more useful and fulfilling jobs.

            But sometimes people get caught in the crossfire, such as creative people having to compete with machines that can churn out decent, derivative works far more quickly. But that just means that the nature of work will change. If we use the printing press eliminating scribe jobs as an example, people have largely moved from reproducing text to designing new typefaces for branding purposes (or being commissioned for a calligraphy piece).

            I think the same is happening w/ art right now. Traditional, 9-5 artists producing largely derivative work is going away, because most people don’t need something truly original. So the number of artists will go down, but the truly great artists will still have a place in creating original works and innovating new types of art. We will still need people with an artistic eye to tune what the AI produces, so instead of manually creating the art, they’ll guide the art w/ tools, much like how farmers don’t hoe fields manually and instead use tractors (which will become increasing autonomous as time passes).

            I’ve gotten into chess recently, and chess is a game that is largely “solved” by AI, meaning the best bot will beat or tie the best human player every time. There’s still some competition between the best bots, but bot v human is pretty firmly in the bot camp and has been for years. However, chess is still a vibrant sport, and people still earn a living playing it (and perhaps more than ever!). It turns out we value the human aspect of chess, and I don’t see that changing anytime soon. I think the same applies to art and other fields AI can “replace,” because that human touch still very much has value.

            If you fight technology, you will lose. So instead of that, fight for fairness and opportunity.

            They’re not dumping billions into data centers and talking about using entire nuclear reactors to power them just because they think AI is a fun toy.

            Well yeah, they’re doing it because they think it’ll make us more productive. For a business owner/exec, that means higher profits. For the rest of us, that usually means higher inflation-adjusted incomes (either through increased wages or reduces costs).