Search Results for "network data speed"

DeiT (Data-efficient Image Transformers) shows that Vision Transformers can be trained competitively on ImageNet-1k without external data by using strong training recipes and knowledge distillation. Its key idea is a specialized distillation strategy—including a learnable “distillation token”—that lets a transformer learn effectively from a CNN or transformer teacher on modest-scale datasets. The project provides compact ViT variants (Tiny/Small/Base) that achieve excellent...

...With a simple API and examples, it invites experimentation on tradeoffs between compression, fidelity, and speed.

...Pretrained checkpoints cover multiple model sizes so practitioners can trade accuracy for speed and memory depending on their deployment constraints.

FlashMLA is a high-performance decoding kernel library designed especially for Multi-Head Latent Attention (MLA) workloads, targeting NVIDIA Hopper GPU architectures. It provides optimized kernels for MLA decoding, including support for variable-length sequences, helping reduce latency and increase throughput in model inference systems using that attention style. The library supports both BF16 and FP16 data types, and includes a paged KV cache implementation with a block size of 64 to...

The Universal Sentence Encoder (USE) is a pre-trained deep learning model designed to encode sentences into fixed-length embeddings for use in various natural language processing (NLP) tasks. It leverages Transformer and Deep Averaging Network (DAN) architectures to generate embeddings that capture the semantic meaning of sentences. The model is designed for tasks like sentiment analysis, semantic textual similarity, and clustering, and provides high-quality sentence representations in a...

GigaChat 3 Ultra is a flagship instruct-model built on a custom Mixture-of-Experts architecture with 702B total and 36B active parameters. It leverages Multi-head Latent Attention to compress the KV cache into latent vectors, dramatically reducing memory demand and improving inference speed at scale. The model also employs Multi-Token Prediction, enabling multi-step token generation in a single pass for up to 40% faster output through speculative and parallel decoding techniques. Its training corpus incorporates ten languages, enriched with books, academic sources, code datasets, mathematical tasks, and more than 5.5 trillion tokens of high-quality synthetic data.

granite-timeseries-ttm-r2 is part of IBM’s TinyTimeMixers (TTM) series—compact, pre-trained models for multivariate time series forecasting. Unlike massive foundation models, TTM models are designed to be lightweight yet powerful, with only ~805K parameters, enabling high performance even on CPU or single-GPU machines. The r2 version is pre-trained on ~700M samples (r2.1 expands to ~1B), delivering up to 15% better accuracy than the r1 version. TTM supports both zero-shot and fine-tuned...