Adaptive Input Representations for Neural Language Modeling
read the original abstract
We introduce adaptive input representations for neural language modeling which extend the adaptive softmax of Grave et al. (2017) to input representations of variable capacity. There are several choices on how to factorize the input and output layers, and whether to model words, characters or sub-word units. We perform a systematic comparison of popular choices for a self-attentional architecture. Our experiments show that models equipped with adaptive embeddings are more than twice as fast to train than the popular character input CNN while having a lower number of parameters. On the WikiText-103 benchmark we achieve 18.7 perplexity, an improvement of 10.5 perplexity compared to the previously best published result and on the Billion Word benchmark, we achieve 23.02 perplexity.
This paper has not been read by Pith yet.
Forward citations
Cited by 11 Pith papers
-
Efficiently Modeling Long Sequences with Structured State Spaces
S4 is an efficient state space sequence model that captures long-range dependencies via structured parameterization of the SSM, achieving state-of-the-art results on the Long Range Arena and other benchmarks while bei...
-
Sundial: A Family of Highly Capable Time Series Foundation Models
Sundial uses TimeFlow Loss for native pre-training of Transformers on continuous time series from TimeBench, achieving SOTA point and probabilistic forecasting with millisecond inference.
-
LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale
LLM.int8() performs 8-bit inference for transformers up to 175B parameters with no accuracy loss by combining vector-wise quantization for most features with 16-bit mixed-precision handling of systematic outlier dimensions.
-
ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
ALBERT reduces BERT parameters via embedding factorization and layer sharing, adds inter-sentence coherence pretraining, and reaches SOTA on GLUE, RACE, and SQuAD with fewer parameters than BERT-large.
-
XLNet: Generalized Autoregressive Pretraining for Language Understanding
XLNet is a generalized autoregressive pretraining method that learns bidirectional contexts via permutation-based factorization and outperforms BERT on 20 NLP tasks.
-
Beyond N-gram: Data-Aware X-GRAM Extraction for Efficient Embedding Parameter Scaling
X-GRAM applies data-aware dynamic token injection with hybrid hashing and local feature extraction to achieve up to 4.4 accuracy point gains over vanilla backbones and 3.2 over retrieval baselines at 0.73B-1.15B scale...
-
MoE-LLaVA: Mixture of Experts for Large Vision-Language Models
MoE-LLaVA applies mixture-of-experts sparsity to LVLMs via MoE-Tuning, delivering LLaVA-1.5-7B level visual understanding and better hallucination resistance with only ~3B active parameters.
-
Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation
ALiBi enables transformers trained on length-1024 sequences to extrapolate to length-2048 with the same perplexity as a sinusoidal model trained on 2048, while training 11% faster and using 11% less memory.
-
Compressive Transformers for Long-Range Sequence Modelling
Compressive Transformer sets new records on WikiText-103 (17.1 ppl) and Enwik8 (0.97 bpc) via memory compression and introduces the PG-19 long-range language benchmark.
-
DA-Cramming: Enhancing Cost-Effective Language Model Pretraining with Dependency Agreement Integration
DA-Cramming inserts chunk-level dependency agreement embeddings into a dual-stage pretraining pipeline and reports better downstream performance than prior Cramming baselines.
-
A Comprehensive Overview of Large Language Models
A survey paper providing an overview of Large Language Models, their background, and recent advances in the field.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.