Qizhen Weng

Zhexiang Zhang, Ye Wang, Yumiao Zhao et al.

Serving large Mixture-of-Experts (MoE) models is challenging because of their large memory footprints, heterogeneous resource demands, and highly dynamic inference workloads. Most existing MoE inference systems deploy the entire model as a monolithic unit, forcing attention and MoE layers to share the same resource configuration despite their different scaling behaviors and resource bottlenecks. Such coarse-grained provisioning leads to resource inefficiency and suboptimal performance. We present JANUS, a scalable and resource-efficient MoE inference system built around three key principles. First, JANUS disaggregates attention and MoE layers onto separate GPU worker pools, enabling independent resource provisioning for the two layer types, and uses an adaptive two-phase communication mechanism for low-latency data exchange. Second, because MoE-layer execution is often memory-bound and highly sensitive to activated-expert imbalance, JANUS introduces a lightweight, microsecond-scale activation scheduler that balances per-layer activated experts across MoE instances to reduce inference latency. Third, JANUS employs a fine-grained, SLO-aware resource scaling scheme that jointly selects attention resources, MoE resources, and expert placement to minimize GPU cost under token-level SLOs. Evaluation shows that JANUS improves per-GPU throughput by up to 4.7x over state-of-the-art MoE inference baselines while satisfying token-level latency SLOs.

Zheng Cai, Maosong Cao, Haojiong Chen et al. · pku

The evolution of Large Language Models (LLMs) like ChatGPT and GPT-4 has sparked discussions on the advent of Artificial General Intelligence (AGI). However, replicating such advancements in open-source models has been challenging. This paper introduces InternLM2, an open-source LLM that outperforms its predecessors in comprehensive evaluations across 6 dimensions and 30 benchmarks, long-context modeling, and open-ended subjective evaluations through innovative pre-training and optimization techniques. The pre-training process of InternLM2 is meticulously detailed, highlighting the preparation of diverse data types including text, code, and long-context data. InternLM2 efficiently captures long-term dependencies, initially trained on 4k tokens before advancing to 32k tokens in pre-training and fine-tuning stages, exhibiting remarkable performance on the 200k ``Needle-in-a-Haystack" test. InternLM2 is further aligned using Supervised Fine-Tuning (SFT) and a novel Conditional Online Reinforcement Learning from Human Feedback (COOL RLHF) strategy that addresses conflicting human preferences and reward hacking. By releasing InternLM2 models in different training stages and model sizes, we provide the community with insights into the model's evolution.

Jian Chen, Zhuoran Wang, Jiayu Qin et al.

Large language models rely on kv-caches to avoid redundant computation during autoregressive decoding, but as context length grows, reading and writing the cache can quickly saturate GPU memory bandwidth. Recent work has explored KV-cache compression, yet most approaches neglect the data-dependent nature of kv-caches and their variation across layers. We introduce KV-CoRE KV-cache Compressibility by Rank Evaluation), an SVD-based method for quantifying the data-dependent low-rank compressibility of kv-caches. KV-CoRE computes the optimal low-rank approximation under the Frobenius norm and, being gradient-free and incremental, enables efficient dataset-level, layer-wise evaluation. Using this method, we analyze multiple models and datasets spanning five English domains and sixteen languages, uncovering systematic patterns that link compressibility to model architecture, training data, and language coverage. As part of this analysis, we employ the Normalized Effective Rank as a metric of compressibility and show that it correlates strongly with performance degradation under compression. Our study establishes a principled evaluation framework and the first large-scale benchmark of kv-cache compressibility in LLMs, offering insights for dynamic, data-aware compression and data-centric model development.

Yabo Chen, Yuanzhi Liang, Jiepeng Wang et al.

World models aim to endow AI systems with the ability to represent, generate, and interact with dynamic environments in a coherent and temporally consistent manner. While recent video generation models have demonstrated impressive visual quality, they remain limited in real-time interaction, long-horizon consistency, and persistent memory of dynamic scenes, hindering their evolution into practical world models. In this report, we present TeleWorld, a real-time multimodal 4D world modeling framework that unifies video generation, dynamic scene reconstruction, and long-term world memory within a closed-loop system. TeleWorld introduces a novel generation-reconstruction-guidance paradigm, where generated video streams are continuously reconstructed into a dynamic 4D spatio-temporal representation, which in turn guides subsequent generation to maintain spatial, temporal, and physical consistency. To support long-horizon generation with low latency, we employ an autoregressive diffusion-based video model enhanced with Macro-from-Micro Planning (MMPL)--a hierarchical planning method that reduces error accumulation from frame-level to segment-level-alongside efficient Distribution Matching Distillation (DMD), enabling real-time synthesis under practical computational budgets. Our approach achieves seamless integration of dynamic object modeling and static scene representation within a unified 4D framework, advancing world models toward practical, interactive, and computationally accessible systems. Extensive experiments demonstrate that TeleWorld achieves strong performance in both static and dynamic world understanding, long-term consistency, and real-time generation efficiency, positioning it as a practical step toward interactive, memory-enabled world models for multimodal generation and embodied intelligence.

Longwen Wang, Xuan'er Wu, Xiaohui Hu et al.

Effective reward design is a central challenge in Reinforcement Learning (RL) for code generation. Mainstream pass/fail outcome rewards enforce functional correctness via executing unit tests, but the resulting sparsity limits potential performance gains. While recent work has explored external Reward Models (RM) to generate richer, continuous rewards, the learned RMs suffer from reward misalignment and prohibitive computational cost. In this paper, we introduce \textbf{VeRPO} (\textbf{V}erifiable D\textbf{e}nse \textbf{R}eward \textbf{P}olicy \textbf{O}ptimization), a novel RL framework for code generation that synthesizes \textit{robust and dense rewards fully grounded in verifiable execution feedback}. The core idea of VeRPO is constructing dense rewards from weighted partial success: by dynamically estimating the difficulty weight of each unit test based on the execution statistics during training, a dense reward is derived from the sum of weights of the passed unit tests. To solidify the consistency between partial success and end-to-end functional correctness, VeRPO further integrates the dense signal with global execution outcomes, establishing a robust and dense reward paradigm relying solely on verifiable execution feedback. Extensive experiments across diverse benchmarks and settings demonstrate that VeRPO consistently outperforms outcome-driven and RM-based baselines, achieving up to +8.83\% gain in pass@1 with negligible time cost (< 0.02\%) and zero GPU memory overhead.

Tianze Wang, Yifei Liu, Chen Chen et al.

Modern AI clusters, which host diverse workloads like data pre-processing, training and inference, often store the large-volume data in cloud storage and employ caching frameworks to facilitate remote data access. To avoid code-intrusion complexity and minimize cache space wastage, it is desirable to maintain a unified cache shared by all the workloads. However, existing cache management strategies, designed for specific workloads, struggle to handle the heterogeneous AI workloads in a cluster -- which usually exhibit heterogeneous access patterns and item storage granularities. In this paper, we propose IGTCache, a unified, high-efficacy cache for modern AI clusters. IGTCache leverages a hierarchical access abstraction, AccessStreamTree, to organize the recent data accesses in a tree structure, facilitating access pattern detection at various granularities. Using this abstraction, IGTCache applies hypothesis testing to categorize data access patterns as sequential, random, or skewed. Based on these detected access patterns and granularities, IGTCache tailors optimal cache management strategies including prefetching, eviction, and space allocation accordingly. Experimental results show that IGTCache increases the cache hit ratio by 55.6% over state-of-the-art caching frameworks, reducing the overall job completion time by 52.2%.

Chen Chen, Qizhen Weng, Wei Wang et al.

Machine learning (ML) models are increasingly trained in clusters with non-dedicated workers possessing heterogeneous resources. In such scenarios, model training efficiency can be negatively affected by stragglers -- workers that run much slower than others. Efficient model training requires eliminating such stragglers, yet for modern ML workloads, existing load balancing strategies are inefficient and even infeasible. In this paper, we propose a novel strategy called semi-dynamic load balancing to eliminate stragglers of distributed ML workloads. The key insight is that ML workers shall be load-balanced at iteration boundaries, being non-intrusive to intra-iteration execution. We develop LB-BSP based on such an insight, which is an integrated worker coordination mechanism that adapts workers' load to their instantaneous processing capabilities by right-sizing the sample batches at the synchronization barriers. We have custom-designed the batch sizing algorithm respectively for CPU and GPU clusters based on their own characteristics. LB-BSP has been implemented as a Python module for ML frameworks like TensorFlow and PyTorch. Our EC2 deployment confirms that LB-BSP is practical, effective and light-weight, and is able to accelerating distributed training by up to $54\%$.