Large language models deliver strong reasoning and tool-use skills, yet their computational demands make them impractical for edge or cost-sensitive deployments. We present \textbf{Xmodel-2.5}, a 1.3-billion-parameter small language model designed as a \emph{drop-in agent core}. Training with maximal-update parameterization ($μ$P) allows hyper-parameters tuned on a 20M-parameter proxy to transfer directly to the full model, even under the parameter-tied \emph{tie-word-embedding} architecture. A 1.4T-token Warmup--Stable--Decay curriculum is used, and we further show that \textbf{switching from AdamW to Muon during the decay phase} improves the 13-task reasoning average by 4.58\,\% while keeping every other hyper-parameter fixed, verifying that early AdamW stability can be paired with late Muon sharpening for better downstream performance. FP8-mixed-precision training balances accuracy and throughput. All checkpoints, recipes, and evaluation code are released under the Apache-2.0 license.\footnote{https://huggingface.co/XiaoduoAILab/Xmodel-2.5 and https://huggingface.co/XiaoduoAILab/Xmodel-2.5-history (training checkpoints).} Training code and evaluation harness: https://github.com/XiaoduoAILab/Xmodel-2.5.

Large Language Models (LLMs) have scaled rapidly in size and complexity, requiring increasingly intricate parallelism for distributed training, such as 3D parallelism. This sophistication motivates a shift toward simpler, more debuggable programming paradigm like Single Program Multiple Data (SPMD). However, SPMD in eager execution introduces two key challenges: ensuring consistency with single-device execution and achieving high performance at scale. In this paper, we introduce veScale, an eager-mode training system that fully embraces SPMD paradigm to democratize distributed tensor programming. veScale addresses the prevalent issue of inconsistent results in systems like PyTorch by introducing a novel algorithm of distributed Random Number Generation (RNG) compatible with arbitrary sharded operators. veScale also significantly boosts training performance by reducing PyTorch primitive's overhead and improving communication efficiency. Evaluations show that veScale delivers up to 2.2x speedup over the state-of-the-art training systems, like TorchTitan, and cuts code complexity by 78.4%, while preserving single-device-equivalent results.

Distributed LLM inference requires careful coordination of parallelization strategies across hundreds to thousands of NPUs to meet production SLOs. Current systems like Megatron-LM rely on static heuristics that separately configure parallelism degrees and per-operator sharding dimensions, leaving significant performance on the table as models scale and hardware topologies diversify. We introduce Learn to Shard, to our knowledge, the first RL-based approach to co-optimize both coarse-grained parallelism degrees and fine-grained per-operator sharding dimensions for distributed LLM inference. Our method employs an attention-based policy over an elite history that learns from high-performing strategies to efficiently navigate the vast combinatorial search space. Evaluated on H100 clusters with MoE models up to 1.6T parameters, Learn to Shard achieves up to 3.5x throughput improvement over metaheuristic baselines and 1.06x over Megatron heuristics.

The sparsely activated mixture-of-experts (MoE) transformer has become a common architecture for large language models (LLMs) due to its sparsity, which requires fewer computational demands while easily scaling the model size. In MoE models, each MoE layer requires to dynamically choose tokens to activate particular experts for computation while the activated experts may not be located in the same device or GPU as the token. However, this leads to substantial communication and load imbalances across all GPUs, which obstructs the scalability of distributed systems within a GPU cluster. To this end, we introduce HierMoE to accelerate the training of MoE models by two topology-aware techniques: 1) token deduplication to reduce the communication traffic, and 2) expert swap to balance the workloads among all GPUs. To enable the above two proposed approaches to be more general, we build theoretical models aimed at achieving the best token duplication and expert swap strategy under different model configurations and hardware environments. We implement our prototype HierMoE system atop Megatron-LM and conduct experiments on a 32-GPU cluster with DeepSeek-V3 and Qwen3-30B-A3B models. Experimental results show that our HierMoE achieves $1.55\times$ to $3.32\times$ faster communication and delivers $1.18\times$ to $1.27\times$ faster end-to-end training compared to state-of-the-art MoE training systems, Tutel-2DH, SmartMoE, and Megatron-LM.

We design and implement AXLearn, a production deep learning system that facilitates scalable and high-performance training of large deep learning models. Compared to other state-of-the-art deep learning systems, AXLearn has a unique focus on modularity and support for heterogeneous hardware infrastructure. AXLearn's internal interfaces between software components follow strict encapsulation, allowing different components to be assembled to facilitate rapid model development and experimentation on heterogeneous compute infrastructure. We introduce a novel method of quantifying modularity via Lines-of-Code (LoC)-complexity, which demonstrates how our system maintains constant complexity as we scale the components in the system, compared to linear or quadratic complexity in other systems. This allows integrating features such as Rotary Position Embeddings (RoPE) into AXLearn across hundred of modules with just 10 lines of code, compared to hundreds as required in other systems. At the same time, AXLearn maintains equivalent performance compared to state-of-the-art training systems. Finally, we share our experience in the development and operation of AXLearn.

Given the popularity of generative AI, Large Language Models (LLMs) often consume hundreds or thousands of GPUs for parallelizing and accelerating the training process. Communication overhead becomes more pronounced when training LLMs at scale. To eliminate communication overhead in distributed LLM training, we propose Domino, which provides a generic scheme to hide communication behind computation. By breaking data dependency of a single batch training into smaller independent pieces, Domino pipelines these independent pieces training and provides generic strategy of fine-grained communication and computation overlapping. Extensive results show that, comparing with Megatron-LM, Domino achieves up to 1.3x speedup for LLM training on Nvidia DGX-H100 GPUs.

Multimodal large language models (MLLMs) have extended the success of large language models (LLMs) to multiple data types, such as image, text and audio, achieving significant performance in various domains, including multimodal translation, visual question answering and content generation. Nonetheless, existing systems are inefficient to train MLLMs due to substantial GPU bubbles caused by the heterogeneous modality models and complex data dependencies in 3D parallelism. This paper proposes Optimus, a distributed MLLM training system that reduces end-to-end MLLM training time. Optimus is based on our principled analysis that scheduling the encoder computation within the LLM bubbles can reduce bubbles in MLLM training. To make scheduling encoder computation possible for all GPUs, Optimus searches the separate parallel plans for encoder and LLM, and adopts a bubble scheduling algorithm to enable exploiting LLM bubbles without breaking the original data dependencies in the MLLM model architecture. We further decompose encoder layer computation into a series of kernels, and analyze the common bubble pattern of 3D parallelism to carefully optimize the sub-millisecond bubble scheduling, minimizing the overall training time. Our experiments in a production cluster show that Optimus accelerates MLLM training by 20.5%-21.3% with ViT-22B and GPT-175B model over 3072 GPUs compared to baselines.

We present the design, implementation and engineering experience in building and deploying MegaScale, a production system for training large language models (LLMs) at the scale of more than 10,000 GPUs. Training LLMs at this scale brings unprecedented challenges to training efficiency and stability. We take a full-stack approach that co-designs the algorithmic and system components across model block and optimizer design, computation and communication overlapping, operator optimization, data pipeline, and network performance tuning. Maintaining high efficiency throughout the training process (i.e., stability) is an important consideration in production given the long extent of LLM training jobs. Many hard stability issues only emerge at large scale, and in-depth observability is the key to address them. We develop a set of diagnosis tools to monitor system components and events deep in the stack, identify root causes, and derive effective techniques to achieve fault tolerance and mitigate stragglers. MegaScale achieves 55.2% Model FLOPs Utilization (MFU) when training a 175B LLM model on 12,288 GPUs, improving the MFU by 1.34x compared to Megatron-LM. We share our operational experience in identifying and fixing failures and stragglers. We hope by articulating the problems and sharing our experience from a systems perspective, this work can inspire future LLM systems research.

Multi-task model training has been adopted to enable a single deep neural network model (often a large language model) to handle multiple tasks (e.g., question answering and text summarization). Multi-task training commonly receives input sequences of highly different lengths due to the diverse contexts of different tasks. Padding (to the same sequence length) or packing (short examples into long sequences of the same length) is usually adopted to prepare input samples for model training, which is nonetheless not space or computation efficient. This paper proposes a dynamic micro-batching approach to tackle sequence length variation and enable efficient multi-task model training. We advocate pipeline-parallel training of the large model with variable-length micro-batches, each of which potentially comprises a different number of samples. We optimize micro-batch construction using a dynamic programming-based approach, and handle micro-batch execution time variation through dynamic pipeline and communication scheduling, enabling highly efficient pipeline training. Extensive evaluation on the FLANv2 dataset demonstrates up to 4.39x higher training throughput when training T5, and 3.25x when training GPT, as compared with packing-based baselines. DynaPipe's source code is publicly available at https://github.com/awslabs/optimizing-multitask-training-through-dynamic-pipelines.

We introduce Breadth-First Pipeline Parallelism, a novel training schedule which optimizes the combination of pipeline and data parallelism. Breadth-First Pipeline Parallelism lowers training time, cost and memory usage by combining a high GPU utilization with a small batch size per GPU, and by making use of fully sharded data parallelism. Experimentally, we observed an increase of up to 43% in training throughput for a 52 billion-parameter model using a small batch size per GPU compared to Megatron-LM, which would reduce the training time and cost by the same amount on a large GPU cluster.