Why It Matters
Significant Training Efficiency Gains : Through multi-level co-optimization — including fine-grained MoE communication optimization, selective recomputation, DVM automatic graph-operator fusion, and Ascend C mHC fused operators — overall training throughput was improved by approximately 96% over out-of-the-box performance.
Overview
I find another new model at HF: https://huggingface.co/XingChen-AGI/Xing4.0-29B-A4B "Xing4.0-29B-A4B is a next-generation large language model in the Xing series (formerly TeleChat ), developed by China Telecom Artificial Intelligence Technology Co., Ltd. With 29B total parameters and only 4B activated per token, it natively supports a 256K context length, extensible to 512K. It is the first model of this scale trained entirely on the Ascend NPU platform with the MindSpore framework, and deeply optimized for complex engineering tasks.
Why It Matters
Significant Training Efficiency Gains : Through multi-level co-optimization — including fine-grained MoE communication optimization, selective recomputation, DVM automatic graph-operator fusion, and Ascend C mHC fused operators — overall training throughput was improved by approximately 96% over out-of-the-box performance.
Key Details
For more information, please refer to our GitHub repository .
Highlights Agent-Oriented Architecture : Built on the mHC + MLA + MTP architecture, supporting multi-step planning, tool calling, and complex reasoning chain execution, ensuring task coherence and execution stability under long contexts.
Deep Co-optimization with Ascend NPU : Adapted for Ascend 910C clusters using MindSpore/MindFormers, including feature adaptation for mHC and fused operator development, enabling stable and efficient training on the Ascend platform.
Significant Training Efficiency Gains : Through multi-level co-optimization — including fine-grained MoE communication optimization, selective recomputation, DVM automatic graph-operator fusion, and Ascend C mHC fused operators — overall training throughput was improved by approximately 96% over out-of-the-box performance.
Full Open-Source Ecosystem Compatibility : Supports LLaMA-Factory and MindFormers for fine-tuning; SGLang, vLLM, and KTransformers for inference and deployment; with targeted adaptation and format alignment for agent frameworks such as OpenCode, Claude Code, OpenClaw, and Hermes, enabling seamless integration into existing workflows.
--- *Original report published by [Reddit LocalLLaMA (Model Leaks & Local AI)](https://www.reddit.com/r/LocalLLaMA/comments/1wimf5p/xingchenagixing4029ba4b_moe/).*
