Introducing RWKV - An RNN with the advantages of a transformer
Hugging Face introduces RWKV, a recurrent neural network architecture that claims to combine the parallelizable training of transformers with the efficient linear-time inference of RNNs. The model avoids the quadratic attention bottleneck of standard transformers while maintaining competitive performance. RWKV represents an alternative architectural direction to the dominant transformer paradigm for language modeling.
Related guides (3)
Related events (8)
The Reformer - Pushing the limits of language modeling
This Hugging Face blog post covers the Reformer, a memory-efficient transformer architecture that uses locality-sensitive hashing (LSH) attention and reversible residual layers to handle very long sequences. The post explains the technical mechanisms that allow Reformer to process sequences up to 1 million tokens with significantly reduced memory footprint compared to standard transformers. It serves as an educational deep-dive into the architectural innovations introduced in the original Reformer paper by Kitaev et al.
Dynamic short convolutions yield 1.33–1.60× compute advantage over standard Transformers
A new arXiv preprint introduces dynamic short convolutions as an architectural primitive for Transformers, using input-dependent filters to combine locality bias with increased expressivity. Experiments across 150M–2B parameter language models show consistent perplexity improvements over standard Transformers and static convolution variants, with scaling-law fits indicating a 1.33× compute advantage when applied to key/query/value vectors and 1.60× when added after every linear layer. The technique also improves linear RNNs (Mamba-2, Gated DeltaNet) and mixture-of-experts architectures, with custom Triton kernels making training practical.
Introducing Decision Transformers on Hugging Face
Hugging Face introduces support for Decision Transformers, a framework that casts offline reinforcement learning as a sequence modeling problem using transformer architectures. The blog post covers the conceptual basis of Decision Transformers and their integration into the Hugging Face ecosystem. This represents an early step in bringing RL-based model paradigms into the standard ML tooling stack.
Triplet-Block Diffusion RWKV: Unifying Linear-Time Causal Models with Bidirectional Discrete Diffusion
The paper introduces B³D-RWKV, a 7.2B-parameter language model that combines RWKV's O(L) linear-time inference with parallel bidirectional discrete diffusion via a triplet-block layout. This architecture resolves the fundamental tension between causal (unidirectional) and diffusion (bidirectional) attention requirements. On an 8-task evaluation suite, B³D-RWKV-7.2B achieves comparable accuracy to existing models while delivering an average 1.6× decoding throughput speedup over baselines.
Probabilistic Time Series Forecasting with Transformers
This Hugging Face blog post introduces probabilistic time series forecasting using Transformer-based models available in the Hugging Face ecosystem. It covers the application of attention-based architectures to sequential prediction tasks with uncertainty quantification. The post serves as a tutorial and capability demonstration for time series modeling within the Transformers library.
Supervised Memory Training enables parallel RNN pretraining without backpropagation through time
A new arXiv preprint proposes Supervised Memory Training (SMT), a method that trains recurrent neural networks by reducing the problem to supervised learning on one-step memory transitions, bypassing backpropagation through time entirely. A Transformer-based encoder generates memory labels via a predictive state objective, enabling time-parallel training with O(1) gradient path length between any two tokens. SMT outperforms BPTT on language modeling and pixel sequence modeling tasks across multiple RNN architectures. The approach could enable RNNs to scale more effectively by decoupling memory content from update mechanics.
RayDer: Scalable Self-Supervised Novel View Synthesis via Unified Feed-Forward Transformer
RayDer is a unified feed-forward transformer that consolidates camera estimation, scene reconstruction, and rendering into a single backbone for self-supervised novel view synthesis (NVS). By treating dynamic content as a nuisance factor absorbed by a minimal dynamic state, it enables stable training on unconstrained real-world video without requiring dynamic-scene reconstruction. The model exhibits clean power-law scaling with both data and compute across multiple model sizes, and achieves zero-shot open-set performance competitive with supervised state-of-the-art methods on multiple benchmarks.
Bamba: Inference-Efficient Hybrid Mamba2 Model
Hugging Face published a blog post introducing Bamba, a hybrid architecture combining Mamba2 state-space layers with attention layers, designed for inference efficiency. The model targets reduced KV-cache memory and improved throughput compared to pure transformer architectures. The post covers architecture details, training approach, and benchmarking results positioning Bamba as a practical alternative for deployment-constrained settings.