Search Results for author:
Found 28 papers, 16 papers with code
ParetoQ: Scaling Laws in Extremely Low-bit LLM Quantization
The optimal bit-width for achieving the best trade-off between quantized model size and accuracy has been a subject of ongoing debate.
EdgeTAM: On-Device Track Anything Model
Given that video segmentation is a dense prediction task, we find preserving the spatial structure of the memories is essential so that the queries are split into global-level and patch-level groups.
Efficient Track Anything
The high computation complexity of multistage image encoder and memory module has limited its applications in real-world tasks, e. g., video object segmentation on mobile devices.
Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations
This paper presents Llama Guard 3-1B-INT4, a compact and efficient Llama Guard model, which has been open-sourced to the community during Meta Connect 2024.
LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding
Given a light-weight LLM, our LongVU also scales effectively into a smaller size with state-of-the-art video understanding performance.
Agent-as-a-Judge: Evaluate Agents with Agents
To address this, we introduce the Agent-as-a-Judge framework, wherein agentic systems are used to evaluate agentic systems.
SpinQuant: LLM quantization with learned rotations
With 4-bit quantization of weight, activation, and KV-cache, SpinQuant narrows the accuracy gap on zero-shot reasoning tasks with full precision to merely 2. 9 points on the LLaMA-2 7B model, surpassing LLM-QAT by 19. 1 points and SmoothQuant by 25. 0 points.
Communication Efficient Distributed Training with Distributed Lion
The Lion optimizer has been a promising competitor with the AdamW for training large AI models, with advantages on memory, computation, and sample efficiency.
MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases
The resultant models, denoted as MobileLLM-LS, demonstrate a further accuracy enhancement of 0. 7%/0. 8% than MobileLLM 125M/350M.
SqueezeSAM: User friendly mobile interactive segmentation
A common user expectation is that a click on a specific part of an object will result in the segmentation of the entire object.
Gen2Det: Generate to Detect
In the long-tailed detection setting on LVIS, Gen2Det improves the performance on rare categories by a large margin while also significantly improving the performance on other categories, e. g. we see an improvement of 2. 13 Box AP and 1. 84 Mask AP over just training on real data on LVIS with Mask R-CNN.
Diversify, Don't Fine-Tune: Scaling Up Visual Recognition Training with Synthetic Images
We present a new framework leveraging off-the-shelf generative models to generate synthetic training images, addressing multiple challenges: class name ambiguity, lack of diversity in naive prompts, and domain shifts.
EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anything
On segment anything task such as zero-shot instance segmentation, our EfficientSAMs with SAMI-pretrained lightweight image encoders perform favorably with a significant gain (e. g., ~4 AP on COCO/LVIS) over other fast SAM models.
Ranked #3 on
Zero-Shot Instance Segmentation
on LVIS v1.0 val
MiniGPT-v2: large language model as a unified interface for vision-language multi-task learning
Motivated by this, we target to build a unified interface for completing many vision-language tasks including image description, visual question answering, and visual grounding, among others.
TODM: Train Once Deploy Many Efficient Supernet-Based RNN-T Compression For On-device ASR Models
Results demonstrate that our TODM Supernet either matches or surpasses the performance of manually tuned models by up to a relative of 3% better in word error rate (WER), while efficiently keeping the cost of training many models at a small constant.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
Mixture-of-Supernets: Improving Weight-Sharing Supernet Training with Architecture-Routed Mixture-of-Experts
In NLP tasks like machine translation and pre-trained language modeling, there is a significant performance gap between supernet and training from scratch for the same model architecture, necessitating retraining post optimal architecture identification.
Binary and Ternary Natural Language Generation
For machine translation, we achieved BLEU scores of 21. 7 and 17. 6 on the WMT16 En-Ro benchmark, compared with a full precision mBART model score of 26. 8.
LLM-QAT: Data-Free Quantization Aware Training for Large Language Models
Several post-training quantization methods have been applied to large language models (LLMs), and have been shown to perform well down to 8-bits.
PathFusion: Path-consistent Lidar-Camera Deep Feature Fusion
Fusing 3D LiDAR features with 2D camera features is a promising technique for enhancing the accuracy of 3D detection, thanks to their complementary physical properties.
Fast Point Cloud Generation with Straight Flows
We perform evaluations on multiple 3D tasks and find that our PSF performs comparably to the standard diffusion model, outperforming other efficient 3D point cloud generation methods.
LiCo-Net: Linearized Convolution Network for Hardware-efficient Keyword Spotting
This paper proposes a hardware-efficient architecture, Linearized Convolution Network (LiCo-Net) for keyword spotting.
Learning a Dual-Mode Speech Recognition Model via Self-Pruning
There is growing interest in unifying the streaming and full-context automatic speech recognition (ASR) networks into a single end-to-end ASR model to simplify the model training and deployment for both use cases.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
DepthShrinker: A New Compression Paradigm Towards Boosting Real-Hardware Efficiency of Compact Neural Networks
Efficient deep neural network (DNN) models equipped with compact operators (e. g., depthwise convolutions) have shown great potential in reducing DNNs' theoretical complexity (e. g., the total number of weights/operations) while maintaining a decent model accuracy.
BiT: Robustly Binarized Multi-distilled Transformer
Modern pre-trained transformers have rapidly advanced the state-of-the-art in machine learning, but have also grown in parameters and computational complexity, making them increasingly difficult to deploy in resource-constrained environments.
Low-Precision Hardware Architectures Meet Recommendation Model Inference at Scale
We share in this paper our search strategies to adapt reference recommendation models to low-precision hardware, our optimization of low-precision compute kernels, and the design and development of tool chain so as to maintain our models' accuracy throughout their lifespan during which topic trends and users' interests inevitably evolve.
Check-N-Run: A Checkpointing System for Training Deep Learning Recommendation Models
While Check-N-Run is applicable to long running ML jobs, we focus on checkpointing recommendation models which are currently the largest ML models with Terabytes of model size.
Deep Learning Recommendation Model for Personalization and Recommendation Systems
With the advent of deep learning, neural network-based recommendation models have emerged as an important tool for tackling personalization and recommendation tasks.
Quantizing deep convolutional networks for efficient inference: A whitepaper
Per-channel quantization of weights and per-layer quantization of activations to 8-bits of precision post-training produces classification accuracies within 2% of floating point networks for a wide variety of CNN architectures.
