Arxiv Papers

The paper proposes a speed-up alternative method for the segment anything model (SAM) in computer vision tasks, using a regular CNN detector with an instance segmentation branch, achieving comparable performance with 50x higher run-time speed. https://arxiv.org/abs//2306.12156 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers...

The paper presents a new class of conditional denoising diffusion probabilistic models that can sample from distributions of signals that are never observed directly, but only through a known differentiable forward model. The approach is demonstrated on three challenging computer vision tasks. https://arxiv.org/abs//2306.11719 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts:...

Phi-1 is a new, smaller language model for code with 1.3B parameters, trained on a selection of web data and synthetically generated textbooks and exercises, achieving high accuracy on HumanEval and MBPP. https://arxiv.org/abs//2306.11644 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016 Spotify: https://podcaste...

The paper proposes glimmer, a memory-retrieval hybrid that improves on the lumen approach by incorporating multi-task training and a shallow reranker to achieve strong gains in performance on knowledge-intensive tasks. https://arxiv.org/abs//2306.10231 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016 Spotify: ht...

AssistGPT is a multi-modal AI assistant that integrates LLMs with various tools using a Plan, Execute, Inspect, and Learn (PEIL) approach to address complex visual-based tasks. It achieved state-of-the-art results on A-OKVQA and NExT-QA benchmarks. https://arxiv.org/abs//2306.08640 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-p...

The paper analyzes the effectiveness of self-repair in large language models (LLMs) for code generation, specifically GPT-3.5 and GPT-4, on a challenging dataset called APPS. The study finds that self-repair is only effective in GPT-4 and is bottlenecked by the feedback stage. Using expert human programmers to give feedback on the programs generated by GPT-4 unlocks significant performance gains....

Researchers introduce MagicBrush, a large-scale, manually annotated dataset for instruction-guided real image editing, to address the limitations of existing methods and improve text-guided image editing models. https://arxiv.org/abs//2306.10012 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016 Spotify: https://p...

Researchers propose a new optimizer, LOMO, that reduces memory usage to enable full parameter fine-tuning of large language models on a single machine. https://arxiv.org/abs//2306.09782 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016 Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers

The paper proposes a diffusion-based architecture called TryOnDiffusion that generates photorealistic visualizations of garments on different body shapes and poses, achieving state-of-the-art performance. https://arxiv.org/abs//2306.0827600:00 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016 Spotify: https://pod...

This paper proposes a zero-shot text-guided video-to-video translation framework to adapt image models to videos, achieving global style and local texture temporal consistency at a low cost. https://arxiv.org/abs//2306.07954 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016 Spotify: https://podcasters.spotify.com...

The paper presents GLoRA, an advanced approach for universal parameter-efficient fine-tuning tasks. GLoRA employs a generalized prompt module to optimize pre-trained model weights and adjust intermediate activations, providing more flexibility and capability across diverse tasks and datasets. It outperforms previous methods in natural, specialized, and structured benchmarks, achieving superior acc...

The LongMem framework enables large language models to utilize extended historical data by introducing a decoupled network architecture with memory encoding and adaptive memory retrieval components. This method significantly improves performance on long-context modeling and in-context learning tasks. https://arxiv.org/abs//2306.07174 YouTube: https://www.youtube.com/@ArxivPapers PODCASTS: Apple Po...

Orthogonal Finetuning (OFT) is introduced to guide text-to-image diffusion models for downstream tasks, preserving hyper-spherical energy essential for semantic generation. Constrained Orthogonal Finetuning (COFT) improves stability. It outperforms existing methods in quality and speed. https://arxiv.org/abs//2306.07280 YouTube: https://youtu.be/STUPtDIyDGg PODCASTS: Apple Podcasts: https://podcas...

FasterViT is a hybrid CNN-ViT neural network that combines the strengths of CNNs and ViTs, achieving high image throughput for computer vision tasks. It introduces a Hierarchical Attention (HAT) approach for efficient self-attention, resulting in state-of-the-art accuracy and faster performance. https://arxiv.org/abs//2306.06189