Large Language Models (LLMs) have significantly advanced natural language processing with exceptional task generalization capabilities. Low-Rank Adaption (LoRA) offers a cost-effective fine-tuning solution, freezing the original model parameters and training only lightweight, low-rank adapter matrices. However, the memory footprint of LoRA is largely dominated by the original model parameters. To mitigate this, we propose LoRAM, a memory-efficient LoRA training scheme founded on the intuition that many neurons in over-parameterized LLMs have low training utility but are essential for inference. LoRAM presents a unique twist: it trains on a pruned (small) model to obtain pruned low-rank matrices, which are then recovered and utilized with the original (large) model for inference. Additionally, minimal-cost continual pre-training, performed by the model publishers in advance, aligns the knowledge discrepancy between pruned and original models. Our extensive experiments demonstrate the efficacy of LoRAM across various pruning strategies and downstream tasks. For a model with 70 billion parameters, LoRAM enables training on a GPU with only 20G HBM, replacing an A100-80G GPU for LoRA training and 15 GPUs for full fine-tuning. Specifically, QLoRAM implemented by structured pruning combined with 4-bit quantization, for LLaMA-3.1-70B (LLaMA-2-70B), reduces the parameter storage cost that dominates the memory usage in low-rank matrix training by 15.81$\times$ (16.95$\times$), while achieving dominant performance gains over both the original LLaMA-3.1-70B (LLaMA-2-70B) and LoRA-trained LLaMA-3.1-8B (LLaMA-2-13B). Code is available at https://github.com/junzhang-zj/LoRAM.

Video generation models have achieved remarkable progress in the past year. The quality of AI video continues to improve, but at the cost of larger model size, increased data quantity, and greater demand for training compute. In this report, we present Open-Sora 2.0, a commercial-level video generation model trained for only $200k. With this model, we demonstrate that the cost of training a top-performing video generation model is highly controllable. We detail all techniques that contribute to this efficiency breakthrough, including data curation, model architecture, training strategy, and system optimization. According to human evaluation results and VBench scores, Open-Sora 2.0 is comparable to global leading video generation models including the open-source HunyuanVideo and the closed-source Runway Gen-3 Alpha. By making Open-Sora 2.0 fully open-source, we aim to democratize access to advanced video generation technology, fostering broader innovation and creativity in content creation. All resources are publicly available at: https://github.com/hpcaitech/Open-Sora.

Language models are strong few-shot learners and achieve good overall accuracy in text classification tasks, masking the fact that their results suffer from great class accuracy imbalance. We believe that the pursuit of overall accuracy should not come from enriching the strong classes, but from raising up the weak ones. To address the imbalance, we propose a post-hoc nonlinear integer programming based debiasing method that ensembles weight correction and membership correction to enable flexible rectifications of class probabilities at both class and sample levels, enhancing the performance of LLMs directly from their outputs. Evaluations with Llama-2-13B on seven text classification benchmarks show that our approach achieves state-of-the-art overall accuracy gains with balanced class accuracies. The resulted probability correction scheme demonstrates that sample-level corrections are necessary to elevate weak classes. In addition, due to effectively correcting weak classes, our method also brings significant performance gains to Llama-2-70B, especially on a biomedical domain task, demonstrating its effectiveness across both small and large model variants.

Diffusion models (DMs) have become the leading choice for generative tasks across diverse domains. However, their reliance on multiple sequential forward passes significantly limits real-time performance. Previous acceleration methods have primarily focused on reducing the number of sampling steps or reusing intermediate results, failing to leverage variations across spatial regions within the image due to the constraints of convolutional U-Net structures. By harnessing the flexibility of Diffusion Transformers (DiTs) in handling variable number of tokens, we introduce RAS, a novel, training-free sampling strategy that dynamically assigns different sampling ratios to regions within an image based on the focus of the DiT model. Our key observation is that during each sampling step, the model concentrates on semantically meaningful regions, and these areas of focus exhibit strong continuity across consecutive steps. Leveraging this insight, RAS updates only the regions currently in focus, while other regions are updated using cached noise from the previous step. The model's focus is determined based on the output from the preceding step, capitalizing on the temporal consistency we observed. We evaluate RAS on Stable Diffusion 3 and Lumina-Next-T2I, achieving speedups up to 2.36x and 2.51x, respectively, with minimal degradation in generation quality. Additionally, a user study reveals that RAS delivers comparable qualities under human evaluation while achieving a 1.6x speedup. Our approach makes a significant step towards more efficient diffusion transformers, enhancing their potential for real-time applications.

Parameter generation has long struggled to match the scale of today large vision and language models, curbing its broader utility. In this paper, we introduce Recurrent Diffusion for Large Scale Parameter Generation (RPG), a novel framework that generates full neural network parameters up to hundreds of millions on a single GPU. Our approach first partitions a networks parameters into non-overlapping tokens, each corresponding to a distinct portion of the model. A recurrent mechanism then learns the inter token relationships, producing prototypes which serve as conditions for a diffusion process that ultimately synthesizes the full parameters. Across a spectrum of architectures and tasks including ResNets, ConvNeXts and ViTs on ImageNet 1K and COCO, and even LoRA based LLMs RPG achieves performance on par with fully trained networks while avoiding excessive memory overhead. Notably, it generalizes beyond its training set to generate valid parameters for previously unseen tasks, highlighting its flexibility in dynamic and open ended scenarios. By overcoming the longstanding memory and scalability barriers, RPG serves as a critical advance in AI generating AI, potentially enabling efficient weight generation at scales previously deemed infeasible.

Despite significant advancements in general-purpose AI agents, several challenges still hinder their practical application in real-world scenarios. First, the limited planning capabilities of Large Language Models (LLM) restrict AI agents from effectively solving complex tasks that require long-horizon planning. Second, general-purpose AI agents struggle to efficiently utilize domain-specific knowledge and human expertise. In this paper, we introduce the Standard Operational Procedure-guided Agent (SOP-agent), a novel framework for constructing domain-specific agents through pseudocode-style Standard Operational Procedures (SOPs) written in natural language. Formally, we represent a SOP as a decision graph, which is traversed to guide the agent in completing tasks specified by the SOP. We conduct extensive experiments across tasks in multiple domains, including decision-making, search and reasoning, code generation, data cleaning, and grounded customer service. The SOP-agent demonstrates excellent versatility, achieving performance superior to general-purpose agent frameworks and comparable to domain-specific agent systems. Additionally, we introduce the Grounded Customer Service Benchmark, the first benchmark designed to evaluate the grounded decision-making capabilities of AI agents in customer service scenarios based on SOPs.

In-context learning, which allows large language models to perform diverse tasks with a few demonstrations, is found to have imbalanced per-class prediction accuracy on multi-class text classification. Although notable output correction methods have been developed to tackle the issue and simultaneously improve downstream prediction accuracy, they may fail to answer the core interpretability challenges: why and which certain classes need corrections, and more importantly, a tailored correction for per-sample, per-class's probability. To address such interpretability gaps, we first find that the imbalance arises from certain classes consistently receiving high ICL output probabilities, whereas others receiving lower or mixed ranges, so the former is more frequently chosen, resulting in higher accuracy; more crucially, we find that these ranges have significantly varying degrees of influence on the accuracy bias, highlighting the need for precise, interpretable probability corrections by range. Motivated by this, we propose FuRud, a Fuzzy Rule Optimization based Debiasing method, that (1) detects which classes need corrections, and (2) for each correction-needed class, detects its probability ranges and applies asymmetric amplifications or reductions to correct them interpretably. Notably, across seven benchmark datasets, FuRud reduces the pairwise class accuracy bias (COBias) by more than half (56%), while achieving a relative increase of 21% in accuracy, outperforming state-of-the-art debiasing methods. Moreover, FuRud can optimize downstream tasks with as few as 10 optimization examples. Furthermore, FuRud can work for prompt formats that lead to highly skewed predictions. For example, FuRud greatly improves ICL outputs which use letter options, with 44% relative accuracy increase and 54% relative COBias reduction.

Datasets nowadays are generally constructed from multiple sources and using different synthetic techniques, making data de-noising and de-duplication crucial before being used for post-training. In this work, we propose to perform instruction tuning by iterative data selection (\ApproachName{}). We measure the quality of a sample from complexity and diversity simultaneously. Instead of calculating the complexity score once for all before fine-tuning, we highlight the importance of updating this model-specific score during fine-tuning to accurately accommodate the dynamic changes of the model. On the other hand, the diversity score is defined on top of the samples' responses under the consideration of their informativeness. IterIT integrates the strengths of both worlds by iteratively updating the complexity score for the top-ranked samples and greedily selecting the ones with the highest complexity-diversity score. Experiments on multiple instruction-tuning data demonstrate consistent improvements of IterIT over strong baselines. Moreover, our approach also generalizes well to domain-specific scenarios and different backbone models. All resources will be available at https://github.com/JiaQiSJTU/IterIT.

Vision Mamba (e.g., Vim) has successfully been integrated into computer vision, and token reduction has yielded promising outcomes in Vision Transformers (ViTs). However, token reduction performs less effectively on Vision Mamba compared to ViTs. Pruning informative tokens in Mamba leads to a high loss of key knowledge and bad performance. This makes it not a good solution for enhancing efficiency in Mamba. Token merging, which preserves more token information than pruning, has demonstrated commendable performance in ViTs. Nevertheless, vanilla merging performance decreases as the reduction ratio increases either, failing to maintain the key knowledge in Mamba. Re-training the token-reduced model enhances the performance of Mamba, by effectively rebuilding the key knowledge. Empirically, pruned Vims only drop up to 0.9% accuracy on ImageNet-1K, recovered by our proposed framework R-MeeTo in our main evaluation. We show how simple and effective the fast recovery can be achieved at minute-level, in particular, a 35.9% accuracy spike over 3 epochs of training on Vim-Ti. Moreover, Vim-Ti/S/B are re-trained within 5/7/17 minutes, and Vim-S only drop 1.3% with 1.2x (up to 1.5x) speed up in inference.

Recent advances in Generative Artificial Intelligence have fueled numerous applications, particularly those involving Generative Adversarial Networks (GANs), which are essential for synthesizing realistic photos and videos. However, efficiently training GANs remains a critical challenge due to their computationally intensive and numerically unstable nature. Existing methods often require days or even weeks for training, posing significant resource and time constraints. In this work, we introduce ParaGAN, a scalable distributed GAN training framework that leverages asynchronous training and an asymmetric optimization policy to accelerate GAN training. ParaGAN employs a congestion-aware data pipeline and hardware-aware layout transformation to enhance accelerator utilization, resulting in over 30% improvements in throughput. With ParaGAN, we reduce the training time of BigGAN from 15 days to 14 hours while achieving 91% scaling efficiency. Additionally, ParaGAN enables unprecedented high-resolution image generation using BigGAN.