57 citations · 194 across the 19 of their papers we have counts for
20 papers
Improving Video Sparse Attention with Fine-grained Router and Sparse Rebasing
Peiyuan Zhang, Guoqiang Wei, Yilong Zhao +7
We present VSA2, a frontier trainable sparse attention for video DiTs. VSA2 includes a variety of new architectural features and training procedures that we apply across all stages…
Unleashing Scalable Context Parallelism for Foundation Models Pre-Training via FCP
Yilong Zhao, Xiaonan Nie, Kan Zhu +6
Context parallelism (CP) has been widely adopted to support the growing context length in foundation model pretraining. However, existing designs fail to handle the large variation…
ByteScale: Efficient Scaling of LLM Training with a 2048K Context Length on More Than 12,000 GPUs
Hao Ge, Junda Feng, Qi Huang +6
Scaling long-context ability is essential for Large Language Models (LLMs). To amortize the memory consumption across multiple devices in long-context training, inter-data partitio…
LSH-MoE: Communication-efficient MoE Training via Locality-Sensitive Hashing
Xiaonan Nie, Qibin Liu, Fangcheng Fu +6
Larger transformer models always perform better on various tasks but require more costs to scale up the model size. To efficiently enlarge models, the mixture-of-experts (MoE) arch…
DataSculpt: Crafting Data Landscapes for Long-Context LLMs through Multi-Objective Partitioning
Keer Lu, Xiaonan Nie, Zheng Liang +8
In recent years, Large Language Models (LLMs) have demonstrated significant improvements across a variety of tasks, one of which is the long-context capability. The key to improvin…
Malleus: Straggler-Resilient Hybrid Parallel Training of Large-scale Models via Malleable Data and Model Parallelization
Haoyang Li, Fangcheng Fu, Hao Ge +7
As the scale of models and training data continues to grow, there is an expanding reliance on more GPUs to train large-scale models, which inevitably increases the likelihood of en…
HexiScale: Facilitating Large Language Model Training over Heterogeneous Hardware
Ran Yan, Youhe Jiang, Xiaonan Nie +3
Training large language models (LLMs) is a computationally intensive task, which is typically conducted in data centers with homogeneous high-performance GPUs. In this paper, we ex…
BaichuanSEED: Sharing the Potential of ExtensivE Data Collection and Deduplication by Introducing a Competitive Large Language Model Baseline
Guosheng Dong, Da Pan, Yiding Sun +17
The general capabilities of Large Language Models (LLM) highly rely on the composition and selection on extensive pretraining datasets, treated as commercial secrets by several ins…
PQCache: Product Quantization-based KVCache for Long Context LLM Inference
Hailin Zhang, Xiaodong Ji, Yilin Chen +5
As the field of Large Language Models (LLMs) continues to evolve, the context length in inference is steadily growing. Key-Value Cache (KVCache), the intermediate representations o…
MEMO: Fine-grained Tensor Management For Ultra-long Context LLM Training
Pinxue Zhao, Hailin Zhang, Fangcheng Fu +9
Nowadays, Large Language Models (LLMs) have been trained using extended context lengths to foster more creative applications. However, long context training poses great challenges…
Clover: Regressive Lightweight Speculative Decoding with Sequential Knowledge
Bin Xiao, Chunan Shi, Xiaonan Nie +5
Large language models (LLMs) suffer from low efficiency as the mismatch between the requirement of auto-regressive decoding and the design of most contemporary GPUs. Specifically,…
Improving Automatic Parallel Training via Balanced Memory Workload Optimization
Yujie Wang, Youhe Jiang, Xupeng Miao +5
Transformer models have emerged as the leading approach for achieving state-of-the-art performance across various application domains, serving as the foundation for advanced large-…
OSDP: Optimal Sharded Data Parallel for Distributed Deep Learning
Youhe Jiang, Fangcheng Fu, Xupeng Miao +2
Large-scale deep learning models contribute to significant performance improvements on varieties of downstream tasks. Current data and model parallelism approaches utilize model re…
FlexMoE: Scaling Large-scale Sparse Pre-trained Model Training via Dynamic Device Placement
Xiaonan Nie, Xupeng Miao, Zilong Wang +5
With the increasing data volume, there is a trend of using large-scale pre-trained models to store the knowledge into an enormous number of model parameters. The training of these…
Angel-PTM: A Scalable and Economical Large-scale Pre-training System in Tencent
Xiaonan Nie, Yi Liu, Fangcheng Fu +5
Recent years have witnessed the unprecedented achievements of large-scale pre-trained models, especially the Transformer models. Many products and services in Tencent Inc., such as…
Galvatron: Efficient Transformer Training over Multiple GPUs Using Automatic Parallelism
Xupeng Miao, Yujie Wang, Youhe Jiang +4
Transformer models have achieved state-of-the-art performance on various domains of applications and gradually becomes the foundations of the advanced large deep learning (DL) mode…
OSDP: Optimal Sharded Data Parallel for Distributed Deep Learning
Youhe Jiang, Fangcheng Fu, Xupeng Miao +2
Large-scale deep learning models contribute to significant performance improvements on varieties of downstream tasks. Current data and model parallelism approaches utilize model re…
HetuMoE: An Efficient Trillion-scale Mixture-of-Expert Distributed Training System
Xiaonan Nie, Pinxue Zhao, Xupeng Miao +2
As giant dense models advance quality but require large amounts of GPU budgets for training, the sparsely gated Mixture-of-Experts (MoE), a kind of conditional computation architec…
HET: Scaling out Huge Embedding Model Training via Cache-enabled Distributed Framework
Xupeng Miao, Hailin Zhang, Yining Shi +4
Embedding models have been an effective learning paradigm for high-dimensional data. However, one open issue of embedding models is that their representations (latent factors) ofte…
EvoMoE: An Evolutional Mixture-of-Experts Training Framework via Dense-To-Sparse Gate
Xiaonan Nie, Xupeng Miao, Shijie Cao +7
Mixture-of-experts (MoE) is becoming popular due to its success in improving the model quality, especially in Transformers. By routing tokens with a sparse gate to a few experts (i…