Overview | NeMo Curator

Score and remove low-quality content using heuristics and ML classifiers to prepare your data for model training using NeMo Curator’s tools and utilities.

Large datasets often contain many documents considered “low quality.” In this context, “low quality” means data we do not want downstream models to learn from, and “high quality” is data we do want them to learn from. The metrics that define quality can vary widely.

How It Works

NeMo Curator’s filtering framework is built around several key components that work within the data processing architecture :

ScoreFilter

Filter and Score Modules

The ScoreFilter is at the center of filtering in NeMo Curator. It applies a filter to a document and optionally saves the score as metadata:

1 from nemo_curator.pipeline import Pipeline
2 from nemo_curator.stages.text.io.reader import JsonlReader
3 from nemo_curator.stages.text.io.writer import JsonlWriter
4 from nemo_curator.stages.text.modules import ScoreFilter
5 from nemo_curator.stages.text.filters import WordCountFilter
6 
7 # Create pipeline
8 pipeline = Pipeline(name="quality_filtering")
9 
10 # Load dataset
11 reader = JsonlReader(
12     file_paths="books_dataset/*.jsonl",
13     fields=["text", "id"]
14 )
15 pipeline.add_stage(reader)
16 
17 # Create and apply filter
18 filter_stage = ScoreFilter(
19     filter_obj=WordCountFilter(min_words=80),
20     text_field="text",
21     score_field="word_count",
22 )
23 pipeline.add_stage(filter_stage)
24 
25 # Save filtered dataset
26 writer = JsonlWriter(path="long_books/")
27 pipeline.add_stage(writer)
28 
29 # Execute pipeline (uses XennaExecutor by default)
30 results = pipeline.run()

Default Executor: When you call pipeline.run() without specifying an executor, NeMo Curator automatically uses XennaExecutor() as the default. You can optionally specify a different executor by passing it as a parameter: pipeline.run(executor=my_executor).

The filter object implements two key methods:

score_document: Computes a quality score for a document
keep_document: Determines if a document should be kept based on its score

Filtering Approaches

Heuristic Filtering

Filter text using configurable rules and metrics rules metrics fast

Classifier Filtering

Filter text using trained quality classifiers ml-models quality scoring

Distributed Classification

GPU-accelerated classification with pre-trained models gpu distributed scalable

Usage

NeMo Curator provides programmatic interfaces for document filtering through the Pipeline framework:

1 from nemo_curator.pipeline import Pipeline
2 from nemo_curator.stages.text.io.reader import JsonlReader
3 from nemo_curator.stages.text.io.writer import JsonlWriter
4 from nemo_curator.stages.text.modules import ScoreFilter
5 from nemo_curator.stages.text.filters import WordCountFilter
6 
7 # Create and configure pipeline
8 pipeline = Pipeline(name="document_filtering")
9 
10 # Add data loading
11 reader = JsonlReader(
12     file_paths="/path/to/input/data/*.jsonl",
13     fields=["text", "id"]
14 )
15 pipeline.add_stage(reader)
16 
17 # Add filtering stage
18 filter_stage = ScoreFilter(
19     filter_obj=WordCountFilter(min_words=80),
20     text_field="text",
21     score_field="word_count"
22 )
23 pipeline.add_stage(filter_stage)
24 
25 # Add output stage
26 writer = JsonlWriter(path="/path/to/output/filtered/")
27 pipeline.add_stage(writer)
28 
29 # Execute pipeline (uses XennaExecutor by default)
30 results = pipeline.run()

Best Practices

When filtering large datasets, consider these performance tips:

Order matters: Place computationally inexpensive filters early in your pipeline
Batch size tuning: Adjust batch sizes based on your hardware capabilities
Use vectorization: Implement batched methods for compute-intensive filters
Disk I/O: Consider compression and chunking strategies for large datasets
Distributed processing: For TB-scale datasets, use distributed filtering with the XennaExecutor