2.3 Python language packages in practice

Key points

1. Subword tokenization splits words into smaller, reusable pieces, for example “anyplace” into “any” and “place”. NVIDIA's RAPIDS natural language processing (NLP) post describes a graphics processing unit (GPU) Bidirectional Encoder Representations from Transformers (BERT) subword tokenizer, so tokenization runs on the GPU too.

   What NVIDIA says (3)

   “We first introduced the GPU BERT subword tokenizer in a previous”

   — Run State of the Art NLP Workloads at Scale with RAPIDS, HuggingFace, and Dask

   “To solve these problems, we use subword tokenization.”

   — Run State of the Art NLP Workloads at Scale with RAPIDS, HuggingFace, and Dask

   “For example, the word “anyplace” can be broken down into “any” and “place,””

   — Run State of the Art NLP Workloads at Scale with RAPIDS, HuggingFace, and Dask

2. The post explains that spaCy needed its inputs on the central processing unit (CPU), so the graphics processing unit (GPU) pipeline had to copy data to CPU memory and back, which slowed it down.

   What NVIDIA says (1)

   “spaCy currently needs your inputs on CPU and thus was slow as it required a copy to CPU memory and back to GPU memory.”

   — Run State of the Art NLP Workloads at Scale with RAPIDS, HuggingFace, and Dask

3. NVIDIA's pandas glossary describes the DataFrame as a two-dimensional, array-like table where each column represents a variable and each row a set of values.

   What NVIDIA says (1)

   “A pandas DataFrame is a two-dimensional, array-like table where each column represents values of a specific variable”

   — What Is Pandas and Why Does it Matter?

Key terms

• cuDF: RAPIDS GPU DataFrame library with a pandas-like API.
• pandas / DataFrame: Python library for tabular data; a DataFrame is a table of rows and columns.
• spaCy: Python NLP library; used here for named-entity recognition on the CPU.

Sample question

Which NVIDIA-documented capability does cuDF provide for text work before feeding a Bidirectional Encoder Representations from Transformers (BERT)-style model?