Dacheng Li

Lianmin Zheng, Wei-Lin Chiang, Ying Sheng et al. · berkeley

Evaluating large language model (LLM) based chat assistants is challenging due to their broad capabilities and the inadequacy of existing benchmarks in measuring human preferences. To address this, we explore using strong LLMs as judges to evaluate these models on more open-ended questions. We examine the usage and limitations of LLM-as-a-judge, including position, verbosity, and self-enhancement biases, as well as limited reasoning ability, and propose solutions to mitigate some of them. We then verify the agreement between LLM judges and human preferences by introducing two benchmarks: MT-bench, a multi-turn question set; and Chatbot Arena, a crowdsourced battle platform. Our results reveal that strong LLM judges like GPT-4 can match both controlled and crowdsourced human preferences well, achieving over 80% agreement, the same level of agreement between humans. Hence, LLM-as-a-judge is a scalable and explainable way to approximate human preferences, which are otherwise very expensive to obtain. Additionally, we show our benchmark and traditional benchmarks complement each other by evaluating several variants of LLaMA and Vicuna. The MT-bench questions, 3K expert votes, and 30K conversations with human preferences are publicly available at https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge.

Dacheng Li, Rulin Shao, Anze Xie et al. · cmu

FlashAttention (Dao, 2023) effectively reduces the quadratic peak memory usage to linear in training transformer-based large language models (LLMs) on a single GPU. In this paper, we introduce DISTFLASHATTN, a distributed memory-efficient attention mechanism optimized for long-context LLMs training. We propose three key techniques: token-level workload balancing, overlapping key-value communication, and a rematerialization-aware gradient checkpointing algorithm. We evaluate DISTFLASHATTN on Llama-7B and variants with sequence lengths from 32K to 512K. DISTFLASHATTN achieves 8x longer sequences, 4.45 - 5.64x speedup compared to Ring Self-Attention, 2 - 8x longer sequences, 1.24 - 2.01x speedup compared to Megatron-LM with FlashAttention. It achieves 1.67x and 1.26 - 1.88x speedup compared to recent Ring Attention and DeepSpeed-Ulysses. Code is available at https://github.com/RulinShao/LightSeq.

Dacheng Li, Rulin Shao, Hongyi Wang et al.

Enabling private inference is crucial for many cloud inference services that are based on Transformer models. However, existing private inference solutions can increase the inference latency by more than 60x or significantly compromise the inference quality. In this paper, we design the framework MPCFORMER as a practical solution, using Secure Multi-Party Computation (MPC) and Knowledge Distillation (KD). Through extensive evaluations, we show that MPCFORMER significantly speeds up Transformer inference in MPC settings while achieving similar ML performance to the input model. On the IMDb dataset, it achieves similar performance to BERTBASE, while being 5.3x faster. On the GLUE benchmark, it achieves 97% performance of BERTBASE with a 2.2x speedup. MPCFORMER remains effective with different trained Transformer weights such as ROBERTABASE and larger models including BERTLarge. Code is available at https://github.com/MccRee177/MPCFormer.

Ying Sheng, Shiyi Cao, Dacheng Li et al.

The "pretrain-then-finetune" paradigm is commonly adopted in the deployment of large language models. Low-Rank Adaptation (LoRA), a parameter-efficient fine-tuning method, is often employed to adapt a base model to a multitude of tasks, resulting in a substantial collection of LoRA adapters derived from one base model. We observe that this paradigm presents significant opportunities for batched inference during serving. To capitalize on these opportunities, we present S-LoRA, a system designed for the scalable serving of many LoRA adapters. S-LoRA stores all adapters in the main memory and fetches the adapters used by the currently running queries to the GPU memory. To efficiently use the GPU memory and reduce fragmentation, S-LoRA proposes Unified Paging. Unified Paging uses a unified memory pool to manage dynamic adapter weights with different ranks and KV cache tensors with varying sequence lengths. Additionally, S-LoRA employs a novel tensor parallelism strategy and highly optimized custom CUDA kernels for heterogeneous batching of LoRA computation. Collectively, these features enable S-LoRA to serve thousands of LoRA adapters on a single GPU or across multiple GPUs with a small overhead. Compared to state-of-the-art libraries such as HuggingFace PEFT and vLLM (with naive support of LoRA serving), S-LoRA can improve the throughput by up to 4 times and increase the number of served adapters by several orders of magnitude. As a result, S-LoRA enables scalable serving of many task-specific fine-tuned models and offers the potential for large-scale customized fine-tuning services. The code is available at https://github.com/S-LoRA/S-LoRA

Yukang Chen, Fuzhao Xue, Dacheng Li et al.

Long-context capability is critical for multi-modal foundation models, especially for long video understanding. We introduce LongVILA, a full-stack solution for long-context visual-language models by co-designing the algorithm and system. For model training, we upgrade existing VLMs to support long video understanding by incorporating two additional stages, i.e., long context extension and long video supervised fine-tuning. However, training on long video is computationally and memory intensive. We introduce the long-context Multi-Modal Sequence Parallelism (MM-SP) system that efficiently parallelizes long video training and inference, enabling 2M context length training on 256 GPUs without any gradient checkpointing. LongVILA efficiently extends the number of video frames of VILA from 8 to 2048, achieving 99.8% accuracy in 6,000-frame (more than 1 million tokens) video needle-in-a-haystack. LongVILA-7B demonstrates strong accuracy on 9 popular video benchmarks, e.g. 65.1% VideoMME with subtitle. Besides, MM-SP is 2.1x - 5.7x faster than ring style sequence parallelism and 1.1x - 1.4x faster than Megatron with a hybrid context and tensor parallelism. Moreover, it seamlessly integrates with Hugging Face Transformers.

Dacheng Li, Hongyi Wang, Eric Xing et al.

Scaling up model sizes can lead to fundamentally new capabilities in many machine learning (ML) tasks. However, training big models requires strong distributed system expertise to carefully design model-parallel execution strategies that suit the model architectures and cluster setups. In this paper, we develop AMP, a framework that automatically derives such strategies. AMP identifies a valid space of model parallelism strategies and efficiently searches the space for high-performed strategies, by leveraging a cost model designed to capture the heterogeneity of the model and cluster specifications. Unlike existing methods, AMP is specifically tailored to support complex models composed of uneven layers and cluster setups with more heterogeneous accelerators and bandwidth. We evaluate AMP on popular models and cluster setups from public clouds and show that AMP returns parallel strategies that match the expert-tuned strategies on typical cluster setups. On heterogeneous clusters or models with heterogeneous architectures, AMP finds strategies with 1.54x and 1.77x higher throughput than state-of-the-art model-parallel systems, respectively.

Song Bian, Dacheng Li, Hongyi Wang et al.

Large-scale Transformer models are known for their exceptional performance in a range of tasks, but training them can be difficult due to the requirement for communication-intensive model parallelism. One way to improve training speed is to compress the message size in communication. Previous approaches have primarily focused on compressing gradients in a data parallelism setting, but compression in a model-parallel setting is an understudied area. We have discovered that model parallelism has fundamentally different characteristics than data parallelism. In this work, we present the first empirical study on the effectiveness of compression methods for model parallelism. We implement and evaluate three common classes of compression algorithms - pruning-based, learning-based, and quantization-based - using a popular Transformer training framework. We evaluate these methods across more than 160 settings and 8 popular datasets, taking into account different hyperparameters, hardware, and both fine-tuning and pre-training stages. We also provide analysis when the model is scaled up. Finally, we provide insights for future development of model parallelism compression algorithms.

Zhijian Liu, Ligeng Zhu, Baifeng Shi et al.

Visual language models (VLMs) have made significant advances in accuracy in recent years. However, their efficiency has received much less attention. This paper introduces NVILA, a family of open VLMs designed to optimize both efficiency and accuracy. Building on top of VILA, we improve its model architecture by first scaling up the spatial and temporal resolutions, and then compressing visual tokens. This "scale-then-compress" approach enables NVILA to efficiently process high-resolution images and long videos. We also conduct a systematic investigation to enhance the efficiency of NVILA throughout its entire lifecycle, from training and fine-tuning to deployment. NVILA matches or surpasses the accuracy of many leading open and proprietary VLMs across a wide range of image and video benchmarks. At the same time, it reduces training costs by 4.5X, fine-tuning memory usage by 3.4X, pre-filling latency by 1.6-2.2X, and decoding latency by 1.2-2.8X. We will soon make our code and models available to facilitate reproducibility.

Ying Sheng, Shiyi Cao, Dacheng Li et al.

High-demand LLM inference services (e.g., ChatGPT and BARD) support a wide range of requests from short chat conversations to long document reading. To ensure that all client requests are processed fairly, most major LLM inference services have request rate limits, to ensure that no client can dominate the request queue. However, this rudimentary notion of fairness also results in under-utilization of the resources and poor client experience when there is spare capacity. While there is a rich literature on fair scheduling, serving LLMs presents new challenges due to their unpredictable request lengths and their unique batching characteristics on parallel accelerators. This paper introduces the definition of LLM serving fairness based on a cost function that accounts for the number of input and output tokens processed. To achieve fairness in serving, we propose a novel scheduling algorithm, the Virtual Token Counter (VTC), a fair scheduler based on the continuous batching mechanism. We prove a 2x tight upper bound on the service difference between two backlogged clients, adhering to the requirement of work-conserving. Through extensive experiments, we demonstrate the superior performance of VTC in ensuring fairness, especially in contrast to other baseline methods, which exhibit shortcomings under various conditions. The reproducible code is available at https://github.com/Ying1123/VTC-artifact

Shiyi Cao, Yichuan Wang, Ziming Mao et al.

Large language model (LLM) inference workload dominates a wide variety of modern AI applications, ranging from multi-turn conversation to document analysis. Balancing fairness and efficiency is critical for managing diverse client workloads with varying prefix patterns. Unfortunately, existing fair scheduling algorithms for LLM serving, such as Virtual Token Counter (VTC), fail to take prefix locality into consideration and thus suffer from poor performance. On the other hand, locality-aware scheduling algorithms in existing LLM serving frameworks tend to maximize the prefix cache hit rate without considering fair sharing among clients. This paper introduces the first locality-aware fair scheduling algorithm, Deficit Longest Prefix Match (DLPM), which can maintain a high degree of prefix locality with a fairness guarantee. We also introduce a novel algorithm, Double Deficit LPM (D$^2$LPM), extending DLPM for the distributed setup that can find a balance point among fairness, locality, and load-balancing. Our extensive evaluation demonstrates the superior performance of DLPM and D$^2$LPM in ensuring fairness while maintaining high throughput (up to 2.87$\times$ higher than VTC) and low per-client (up to 7.18$\times$ lower than state-of-the-art distributed LLM serving system) latency.

Tinghao Xie, Xiangyu Qi, Yi Zeng et al.

Evaluating aligned large language models' (LLMs) ability to recognize and reject unsafe user requests is crucial for safe, policy-compliant deployments. Existing evaluation efforts, however, face three limitations that we address with SORRY-Bench, our proposed benchmark. First, existing methods often use coarse-grained taxonomies of unsafe topics, and are over-representing some fine-grained topics. For example, among the ten existing datasets that we evaluated, tests for refusals of self-harm instructions are over 3x less represented than tests for fraudulent activities. SORRY-Bench improves on this by using a fine-grained taxonomy of 44 potentially unsafe topics, and 440 class-balanced unsafe instructions, compiled through human-in-the-loop methods. Second, linguistic characteristics and formatting of prompts are often overlooked, like different languages, dialects, and more -- which are only implicitly considered in many evaluations. We supplement SORRY-Bench with 20 diverse linguistic augmentations to systematically examine these effects. Third, existing evaluations rely on large LLMs (e.g., GPT-4) for evaluation, which can be computationally expensive. We investigate design choices for creating a fast, accurate automated safety evaluator. By collecting 7K+ human annotations and conducting a meta-evaluation of diverse LLM-as-a-judge designs, we show that fine-tuned 7B LLMs can achieve accuracy comparable to GPT-4 scale LLMs, with lower computational cost. Putting these together, we evaluate over 50 proprietary and open-weight LLMs on SORRY-Bench, analyzing their distinctive safety refusal behaviors. We hope our effort provides a building block for systematic evaluations of LLMs' safety refusal capabilities, in a balanced, granular, and efficient manner. Benchmark demo, data, code, and models are available through https://sorry-bench.github.io.

Gaurav Parmar, Dacheng Li, Kwonjoon Lee et al.

We present a new generative autoencoder model with dual contradistinctive losses to improve generative autoencoder that performs simultaneous inference (reconstruction) and synthesis (sampling). Our model, named dual contradistinctive generative autoencoder (DC-VAE), integrates an instance-level discriminative loss (maintaining the instance-level fidelity for the reconstruction/synthesis) with a set-level adversarial loss (encouraging the set-level fidelity for there construction/synthesis), both being contradistinctive. Extensive experimental results by DC-VAE across different resolutions including 32x32, 64x64, 128x128, and 512x512 are reported. The two contradistinctive losses in VAE work harmoniously in DC-VAE leading to a significant qualitative and quantitative performance enhancement over the baseline VAEs without architectural changes. State-of-the-art or competitive results among generative autoencoders for image reconstruction, image synthesis, image interpolation, and representation learning are observed. DC-VAE is a general-purpose VAE model, applicable to a wide variety of downstream tasks in computer vision and machine learning.