In this technical report, we tackle the challenges of training large-scale Mixture of Experts (MoE) models, focusing on overcoming cost inefficiency and resource limitations prevalent in such systems. To address these issues, we present two differently sized MoE large language models (LLMs), namely Ling-Lite and Ling-Plus (referred to as "Bailing" in Chinese, spelled B\v{a}il\'ing in Pinyin). Ling-Lite contains 16.8 billion parameters with 2.75 billion activated parameters, while Ling-Plus boasts 290 billion parameters with 28.8 billion activated parameters. Both models exhibit comparable performance to leading industry benchmarks. This report offers actionable insights to improve the efficiency and accessibility of AI development in resource-constrained settings, promoting more scalable and sustainable technologies. Specifically, to reduce training costs for large-scale MoE models, we propose innovative methods for (1) optimization of model architecture and training processes, (2) refinement of training anomaly handling, and (3) enhancement of model evaluation efficiency. Additionally, leveraging high-quality data generated from knowledge graphs, our models demonstrate superior capabilities in tool use compared to other models. Ultimately, our experimental findings demonstrate that a 300B MoE LLM can be effectively trained on lower-performance devices while achieving comparable performance to models of a similar scale, including dense and MoE models. Compared to high-performance devices, utilizing a lower-specification hardware system during the pre-training phase demonstrates significant cost savings, reducing computing costs by approximately 20%. The models can be accessed at https://huggingface.co/inclusionAI.

Large models are a hot research topic in the field of artificial intelligence. Leveraging their generative capabilities has the potential to enhance the level and quality of medical services. In response to the limitations of current large language models, which often struggle with accuracy and have narrow capabilities in medical applications, this paper presents a Chinese medical large language model, MedGo. MedGo was trained using a combination of high quality unsupervised medical data, supervised data, and preference alignment data, aimed at enhancing both its versatility and precision in medical tasks. The model was evaluated through the public CBLUE benchmark and a manually constructed dataset ClinicalQA. The results demonstrate that MedGo achieved promising performance across various Chinese medical information processing tasks, achieved the first place in the CBLUE evaluation. Additionally, on our constructed dataset ClinicalQA, MedGo outperformed its base model Qwen2, highlighting its potential to improve both automated medical question answering and clinical decision support. These experimental results demonstrate that MedGo possesses strong information processing capabilities in the medical field. At present, we have successfully deployed MedGo at Shanghai East Hospital.

The cloud is still a popular platform for distributed deep learning (DL) training jobs since resource sharing in the cloud can improve resource utilization and reduce overall costs. However, such sharing also brings multiple challenges for DL training jobs, e.g., high-priority jobs could impact, even interrupt, low-priority jobs. Meanwhile, most existing distributed DL training systems require users to configure the resources (i.e., the number of nodes and resources like CPU and memory allocated to each node) of jobs manually before job submission and can not adjust the job's resources during the runtime. The resource configuration of a job deeply affect this job's performance (e.g., training throughput, resource utilization, and completion rate). However, this usually leads to poor performance of jobs since users fail to provide optimal resource configuration in most cases. \system~is a distributed DL framework can auto-configure a DL job's initial resources and dynamically tune the job's resources to win the better performance. With elastic capability, \system~can effectively adjusts the resources of a job when there are performance issues detected or a job fails because of faults or eviction. Evaluations results show \system~can outperform manual well-tuned resource configurations. Furthermore, in the production Kubernetes cluster of \company, \system~reduces the medium of job completion time by 31\%, and improves the job completion rate by 6\%, CPU utilization by 15\%, and memory utilization by 20\% compared with manual configuration.