AMD Instinct MI300x
PulseAugur coverage of AMD Instinct MI300x — every cluster mentioning AMD Instinct MI300x across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
AMD Instinct MI300X GPU detailed with custom Python management tools
This article details the process of inventorying and measuring an AMD Instinct MI300X GPU on the AMD Developer Cloud. The author developed a suite of Python tools, collectively named 'MCP', to manage and interact with t…
-
vLLM adds speculative decoding for AMD GPUs, boosting inference speed
vLLM has implemented speculative decoding for AMD GPUs, a technique that allows for faster inference by having a smaller draft model propose tokens that a larger target model then verifies. This feature, optimized for A…
-
Instella-MoE: New open-source MoE language model released
A new technical report introduces Instella-MoE, an open-source Mixture-of-Experts (MoE) language model with 16 billion total parameters. Trained on AMD Instinct GPUs, the model incorporates innovations like Gated Multi-…
-
DeepSeek R1 reasoning LLM deployed via SGLang on AMD GPUs
A technical guide details how to deploy the DeepSeek R1 reasoning language model using SGLang on an AMD Instinct MI300X GPU server. The process involves setting up the environment with Docker, downloading the model, and…
-
Musk's SpaceX, xAI to exclusively use Nvidia GPUs, launching space-optimized AI hardware
Elon Musk announced that SpaceX and xAI will exclusively use Nvidia GPUs, citing the Vera Rubin NVL72 architecture as the best available. This decision sidelines competitors like AMD, Cerebras, and others in the merchan…
-
AMD releases open-source Instella-MoE language model
AMD has released Instella-MoE-16B-A3B-Think, a new open-source Mixture-of-Experts language model. This model features 16 billion total parameters with 2.8 billion active parameters per token and was trained from scratch…
-
Zyphra's ZAYA1-8B model shows strong reasoning on AMD hardware
Zyphra has released ZAYA1-8B, an Apache 2.0 licensed Mixture-of-Experts reasoning model with 8.4 billion total parameters and approximately 760 million active parameters. Notably, the model was trained entirely on AMD I…
-
Zyphra releases ZAYA1-8B MoE with sub-billion active parameters
Zyphra has released ZAYA1-8B, an 8.4 billion parameter Mixture-of-Experts model that only activates approximately 760 million parameters per token. This architecture allows it to achieve performance comparable to much l…
-
Clinical AI fine-tuned on AMD hardware, bypassing CUDA dependency
A project has successfully fine-tuned a clinical AI model, MedQA, using AMD hardware and ROCm, demonstrating that advanced AI development is possible without NVIDIA's CUDA. The fine-tuning process utilized the Qwen3-1.7…
-
Zyphra's ZAYA1-8B MoE model trained on AMD hardware outperforms larger rivals
Zyphra AI has released ZAYA1-8B, a Mixture of Experts (MoE) language model with 760 million active parameters and 8.4 billion total parameters. Trained on AMD hardware, this model demonstrates competitive performance ag…
-
Character.ai, DigitalOcean, AMD boost AI inference 2x
Character.ai, in collaboration with DigitalOcean and AMD, has achieved a twofold increase in production inference performance for its AI entertainment platform. This significant improvement was realized through deep tec…