The fusion of speech and language in the era of large language models has garnered significant attention. Discrete speech token is often utilized in text-to-speech tasks for speech compression and portability, which is convenient for joint training with text and have good compression efficiency. However, we found that the discrete speech tokenizer still suffers from information loss. Therefore, we propose a simple yet effective continuous speech tokenizer named Cont-SPT, and a text-to-speech model based on continuous speech tokens. Our results show that the speech language model based on the continuous speech tokenizer has better continuity and higher estimated Mean Opinion Scores (MoS). This enhancement is attributed to better information preservation rate of the continuous speech tokenizer across both low and high frequencies in the frequency domain. The code and resources for Cont-SPT can be found in https://github.com/Yixing-Li/Continuous-Speech-Tokenizer.

In typical multimodal tasks, such as Visual Question Answering (VQA), adversarial attacks targeting a specific image and question can lead large vision-language models (LVLMs) to provide incorrect answers. However, it is common for a single image to be associated with multiple questions, and LVLMs may still answer other questions correctly even for an adversarial image attacked by a specific question. To address this, we introduce the query-agnostic visual attack (QAVA), which aims to create robust adversarial examples that generate incorrect responses to unspecified and unknown questions. Compared to traditional adversarial attacks focused on specific images and questions, QAVA significantly enhances the effectiveness and efficiency of attacks on images when the question is unknown, achieving performance comparable to attacks on known target questions. Our research broadens the scope of visual adversarial attacks on LVLMs in practical settings, uncovering previously overlooked vulnerabilities, particularly in the context of visual adversarial threats. The code is available at https://github.com/btzyd/qava.

We examine the pre-training dynamics of language models, focusing on their ability to copy text from preceding context—a fundamental skill for various LLM applications, including in-context learning (ICL) and retrieval-augmented generation (RAG). We propose a novel perspective that Transformer-based language models develop copying abilities similarly to grokking, which refers to sudden generalization on test set long after the model fit to the training set. Our experiments yield three arguments: (1) The pre-training loss decreases rapidly, while the context copying ability of models initially lags and then abruptly saturates. (2) The speed of developing copying ability is independent of the number of tokens trained, similarly to how grokking speed is unaffected by dataset size as long as the data distribution is preserved. (3) Induction heads, the attention heads responsible for copying, form from shallow to deep layers during training, mirroring the development of circuits in deeper layers during grokking. We contend that the connection between grokking and context copying can provide valuable insights for more effective language model training, ultimately improving in-context performance. For example, we demonstrated that techniques that enhance grokking, such as regularization, either accelerate or enhance the development of context copying.

This paper studies vision-language (V&L) pre-training for deep cross-modal representations. Recently, pre-trained V&L models have shown great success in V&L tasks. However, most existing models apply multi-modal encoders to encode the image and text, at the cost of high training complexity because of the input sequence length. In addition, they suffer from noisy training corpora caused by V&L mismatching. In this work, we propose a lightweight vision-language pre-training (LightVLP) for efficient and effective V&L pre-training. First, we design a new V&L framework with two autoencoders. Each autoencoder involves an encoder, which only takes in unmasked tokens (removes masked ones), as well as a lightweight decoder that reconstructs the masked tokens. Besides, we mask and remove large portions of input tokens to accelerate the training. Moreover, we propose a gated interaction mechanism to cope with noise in aligned image-text pairs. As for a matched image-text pair, the model tends to apply cross-modal representations for reconstructions. By contrast, for an unmatched pair, the model conducts reconstructions mainly using uni-modal representations. Benefiting from the above-mentioned designs, our base model shows competitive results compared to ALBEF while saving 44% FLOPs. Further, we compare our large model with ALBEF under the setting of similar FLOPs on six datasets and show the superiority of LightVLP. In particular, our model achieves 2.2% R@1 gains on COCO Text Retrieval and 1.1% on refCOCO+.

Recently, the success of pre-training in text domain has been fully extended to vision, audio, and cross-modal scenarios. The proposed pre-training models of different modalities are showing a rising trend of homogeneity in their model structures, which brings the opportunity to implement different pre-training models within a uniform framework. In this paper, we present TencentPretrain, a toolkit supporting pre-training models of different modalities. The core feature of TencentPretrain is the modular design. The toolkit uniformly divides pre-training models into 5 components: embedding, encoder, target embedding, decoder, and target. As almost all of common modules are provided in each component, users can choose the desired modules from different components to build a complete pre-training model. The modular design enables users to efficiently reproduce existing pre-training models or build brand-new one. We test the toolkit on text, vision, and audio benchmarks and show that it can match the performance of the original implementations.

Large language models (LLMs) have proven to be very superior to conventional methods in various tasks. However, their expensive computations and high memory requirements are prohibitive for deployment. Model quantization is an effective method for reducing this overhead. The problem is that in most previous works, the quantized model was calibrated using few samples from the training data, which might affect the generalization of the quantized LLMs to unknown cases and tasks. Hence in this work, we explore an important question: Can we design a data-independent quantization method for LLMs to guarantee its generalization performance? In this work, we propose EasyQuant, a training-free and data-independent weight-only quantization algorithm for LLMs. Our observation indicates that two factors: outliers in the weight and quantization ranges, are essential for reducing the quantization error. Therefore, in EasyQuant, we leave the outliers (less than 1%) unchanged and optimize the quantization range to reduce the reconstruction error. With these methods, we surprisingly find that EasyQuant achieves comparable performance to the original model. Since EasyQuant does not depend on any training data, the generalization performance of quantized LLMs is safely guaranteed. Moreover, EasyQuant can be implemented in parallel so that the quantized model could be attained in a few minutes even for LLMs over 100B. To our best knowledge, we are the first work that achieves almost lossless quantization performance for LLMs under a data-independent setting and our algorithm runs over 10 times faster than the data-dependent methods.

Position Embedding (PE) is essential for transformer to capture the sequence ordering of input tokens. Despite its general effectiveness verified in Natural Language Processing (NLP) and Computer Vision (CV), its application in cross-modal tasks remains unexplored and suffers from two challenges: 1) the input text tokens and image patches are not aligned, 2) the encoding space of each modality is different, making it unavailable for feature comparison. In this paper, we propose a unified position embedding method for these problems, called AnChor-basEd Relative Position Embedding (ACE-RPE), in which we first introduce an anchor locating mechanism to bridge the semantic gap and locate anchors from different modalities. Then we conduct the distance calculation of each text token and image patch by computing their shortest paths from the located anchors. Last, we embed the anchor-based distance to guide the computation of cross-attention. In this way, it calculates cross-modal relative position embedding for cross-modal transformer. Benefiting from ACE-RPE, our method obtains new SOTA results on a wide range of benchmarks, such as Image-Text Retrieval on MS-COCO and Flickr30K, Visual Entailment on SNLI-VE, Visual Reasoning on NLVR2 and Weakly-supervised Visual Grounding on RefCOCO+.

This paper introduces TexSmart, a text understanding system that supports fine-grained named entity recognition (NER) and enhanced semantic analysis functionalities. Compared to most previous publicly available text understanding systems and tools, TexSmart holds some unique features. First, the NER function of TexSmart supports over 1,000 entity types, while most other public tools typically support several to (at most) dozens of entity types. Second, TexSmart introduces new semantic analysis functions like semantic expansion and deep semantic representation, that are absent in most previous systems. Third, a spectrum of algorithms (from very fast algorithms to those that are relatively slow but more accurate) are implemented for one function in TexSmart, to fulfill the requirements of different academic and industrial applications. The adoption of unsupervised or weakly-supervised algorithms is especially emphasized, with the goal of easily updating our models to include fresh data with less human annotation efforts.