English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep learning for recommendation data is one of the most pervasive andchallenging AI workload in recent times. State-of-the-art recommendation modelsare one of the largest models matching the likes of GPT-3 and SwitchTransformer. Challenges in deep learning recommendation models (DLRM) stem fromlearning dense embeddings for each of the categorical tokens. These embeddingtables in industrial scale models can be as large as hundreds of terabytes.Such large models lead to a plethora of engineering challenges, not to mentionprohibitive communication overheads, and slower training and inference times.Of these, slower inference time directly impacts user experience. Modelcompression for DLRM is gaining traction and the community has recently shownimpressive compression results. In this paper, we present Random Offset BlockEmbedding Array (ROBE) as a low memory alternative to embedding tables whichprovide orders of magnitude reduction in memory usage while maintainingaccuracy and boosting execution speed. ROBE is a simple fundamental approach inimproving both cache performance and the variance of randomized hashing, whichcould be of independent interest in itself. We demonstrate that we cansuccessfully train DLRM models with same accuracy while using $1000 \times$less memory. A $1000\times$ compressed model directly results in fasterinference without any engineering effort. In particular, we show that we cantrain DLRM model using ROBE array of size 100MB on a single GPU to achieve AUCof 0.8025 or higher as required by official MLPerf CriteoTB benchmark DLRMmodel of 100GB while achieving about $3.1\times$ (209\%) improvement ininference throughput.

Random Offset Block Embedding Array (ROBE) for CriteoTB Benchmark MLPerf DLRM Model : 1000$\times$ Compression and 3.1$\times$ Faster Inference