Changpeng Cai

GLM-5 Team, Aohan Zeng, Xin Lv et al. · tsinghua

We present GLM-5, a next-generation foundation model designed to transition the paradigm of vibe coding to agentic engineering. Building upon the agentic, reasoning, and coding (ARC) capabilities of its predecessor, GLM-5 adopts DSA to significantly reduce training and inference costs while maintaining long-context fidelity. To advance model alignment and autonomy, we implement a new asynchronous reinforcement learning infrastructure that drastically improves post-training efficiency by decoupling generation from training. Furthermore, we propose novel asynchronous agent RL algorithms that further improve RL quality, enabling the model to learn from complex, long-horizon interactions more effectively. Through these innovations, GLM-5 achieves state-of-the-art performance on major open benchmarks. Most critically, GLM-5 demonstrates unprecedented capability in real-world coding tasks, surpassing previous baselines in handling end-to-end software engineering challenges. Code, models, and more information are available at https://github.com/zai-org/GLM-5.

Junhao Su, Changpeng Cai, Feiyu Zhu et al.

Deep neural networks conventionally employ end-to-end backpropagation for their training process, which lacks biological credibility and triggers a locking dilemma during network parameter updates, leading to significant GPU memory use. Supervised local learning, which segments the network into multiple local blocks updated by independent auxiliary networks. However, these methods cannot replace end-to-end training due to lower accuracy, as gradients only propagate within their local block, creating a lack of information exchange between blocks. To address this issue and establish information transfer across blocks, we propose a Momentum Auxiliary Network (MAN) that establishes a dynamic interaction mechanism. The MAN leverages an exponential moving average (EMA) of the parameters from adjacent local blocks to enhance information flow. This auxiliary network, updated through EMA, helps bridge the informational gap between blocks. Nevertheless, we observe that directly applying EMA parameters has certain limitations due to feature discrepancies among local blocks. To overcome this, we introduce learnable biases, further boosting performance. We have validated our method on four image classification datasets (CIFAR-10, STL-10, SVHN, ImageNet), attaining superior performance and substantial memory savings. Notably, our method can reduce GPU memory usage by more than 45\% on the ImageNet dataset compared to end-to-end training, while achieving higher performance. The Momentum Auxiliary Network thus offers a new perspective for supervised local learning. Our code is available at: https://github.com/JunhaoSu0/MAN.

GLM-4. 5 Team, Aohan Zeng, Xin Lv et al.

We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that supports both thinking and direct response modes. Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding (ARC) tasks, scoring 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified. With much fewer parameters than several competitors, GLM-4.5 ranks 3rd overall among all evaluated models and 2nd on agentic benchmarks. We release both GLM-4.5 (355B parameters) and a compact version, GLM-4.5-Air (106B parameters), to advance research in reasoning and agentic AI systems. Code, models, and more information are available at https://github.com/zai-org/GLM-4.5.

Xiuyuan Guo, Chengqi Xu, Guinan Guo et al.

Currently, training large-scale deep learning models is typically achieved through parallel training across multiple GPUs. However, due to the inherent communication overhead and synchronization delays in traditional model parallelism methods, seamless parallel training cannot be achieved, which, to some extent, affects overall training efficiency. To address this issue, we present PPLL (Pipeline Parallelism based on Local Learning), a novel framework that leverages local learning algorithms to enable effective parallel training across multiple GPUs. PPLL divides the model into several distinct blocks, each allocated to a separate GPU. By utilizing queues to manage data transfers between GPUs, PPLL ensures seamless cross-GPU communication, allowing multiple blocks to execute forward and backward passes in a pipelined manner. This design minimizes idle times and prevents bottlenecks typically caused by sequential gradient updates, thereby accelerating the overall training process. We validate PPLL through extensive experiments using ResNet and Vision Transformer (ViT) architectures on CIFAR-10, SVHN, and STL-10 datasets. Our results demonstrate that PPLL significantly enhances the training speed of the local learning method while achieving comparable or even superior training speed to traditional pipeline parallelism (PP) without sacrificing model performance. In a 4-GPU training setup, PPLL accelerated local learning training on ViT and ResNet by 162% and 33%, respectively, achieving 1.25x and 0.85x the speed of traditional pipeline parallelism.

Changpeng Cai, Guinan Guo, Jiao Li et al.

Most earlier researches on talking face generation have focused on the synchronization of lip motion and speech content. However, head pose and facial emotions are equally important characteristics of natural faces. While audio-driven talking face generation has seen notable advancements, existing methods either overlook facial emotions or are limited to specific individuals and cannot be applied to arbitrary subjects. In this paper, we propose a novel one-shot Talking Head Generation framework (SPEAK) that distinguishes itself from the general Talking Face Generation by enabling emotional and postural control. Specifically, we introduce Inter-Reconstructed Feature Disentanglement (IRFD) module to decouple facial features into three latent spaces. Then we design a face editing module that modifies speech content and facial latent codes into a single latent space. Subsequently, we present a novel generator that employs modified latent codes derived from the editing module to regulate emotional expression, head poses, and speech content in synthesizing facial animations. Extensive trials demonstrate that our method ensures lip synchronization with the audio while enabling decoupled control of facial features, it can generate realistic talking head with coordinated lip motions, authentic facial emotions, and smooth head movements. The demo video is available: https://anonymous.4open.science/r/SPEAK-8A22

Yuming Zhang, Shouxin Zhang, Peizhe Wang et al.

Deep neural networks (DNNs) typically employ an end-to-end (E2E) training paradigm which presents several challenges, including high GPU memory consumption, inefficiency, and difficulties in model parallelization during training. Recent research has sought to address these issues, with one promising approach being local learning. This method involves partitioning the backbone network into gradient-isolated modules and manually designing auxiliary networks to train these local modules. Existing methods often neglect the interaction of information between local modules, leading to myopic issues and a performance gap compared to E2E training. To address these limitations, we propose the Multilaminar Leap Augmented Auxiliary Network (MLAAN). Specifically, MLAAN comprises Multilaminar Local Modules (MLM) and Leap Augmented Modules (LAM). MLM captures both local and global features through independent and cascaded auxiliary networks, alleviating performance issues caused by insufficient global features. However, overly simplistic auxiliary networks can impede MLM's ability to capture global information. To address this, we further design LAM, an enhanced auxiliary network that uses the Exponential Moving Average (EMA) method to facilitate information exchange between local modules, thereby mitigating the shortsightedness resulting from inadequate interaction. The synergy between MLM and LAM has demonstrated excellent performance. Our experiments on the CIFAR-10, STL-10, SVHN, and ImageNet datasets show that MLAAN can be seamlessly integrated into existing local learning frameworks, significantly enhancing their performance and even surpassing end-to-end (E2E) training methods, while also reducing GPU memory consumption.