Jin10 Data reported on September 29 that Cambricon announced on its official account that on September 29, 2025, Cambricon has simultaneously achieved adaptation to the latest model DeepSeek-V3.2-Exp of the Depth Exploration Company, and has open-sourced the large model inference engine vLLM-MLU Source Code. For this new model architecture DeepSeek-V3.2-Exp, Cambricon has achieved rapid adaptation through the Triton operator development, utilized the BangC fusion operator development for extreme performance optimization, and based on the parallel strategy of computation and communication, once again achieved industry-leading computational efficiency levels.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
Cambrian Day 0 is compatible with DeepSeek-V3.2-Exp, synchronizing Open Source inference engine vLLM-MLU
Jin10 Data reported on September 29 that Cambricon announced on its official account that on September 29, 2025, Cambricon has simultaneously achieved adaptation to the latest model DeepSeek-V3.2-Exp of the Depth Exploration Company, and has open-sourced the large model inference engine vLLM-MLU Source Code. For this new model architecture DeepSeek-V3.2-Exp, Cambricon has achieved rapid adaptation through the Triton operator development, utilized the BangC fusion operator development for extreme performance optimization, and based on the parallel strategy of computation and communication, once again achieved industry-leading computational efficiency levels.