- NLP
- SUMMARIZATION
- TRANSFORMERS
Text Summarizer
An abstractive text summariser with chunking for long documents: package, CLI and live demo. Code at github.com/delcenjo/text-summarizer.
Abstractive summarisation that does not fall apart on long text: paste a whole article and get back a readable summary instead of a truncated one. It runs as a Python package, a CLI and a hosted web demo.
- Chunking700 words
- ModeldistilBART
- DemoHF Spaces
The real problem is long text
Summarisation models have a context limit, so the work is in the chunking: split on sentences, group into chunks of up to 700 words so each reads on its own, summarise each chunk with distilBART, and if the joined result is still long, re-summarise hierarchically.
Keeping logic away from the model
All the text wrangling — sentence splitting, chunk grouping, joining summaries — is pure code without the model, so it tests without a GPU and in milliseconds. The model only enters at the last step. That separation kept the project maintainable.
Try it live
There is a HuggingFace Spaces demo to try it by pasting text, and a Colab quickstart notebook to use it as a library. Three ways to consume the same package: web, CLI and import.
Text Summarizer
Interactive demo hosted on HuggingFace Spaces
Next projects:
Credit Risk Platform
End-to-end MLOps platform for credit-default scoring: reproducible training with a versioned model registry, a FastAPI inference API with input validation, PSI drift monitoring and a Prometheus and Grafana observability stack, all containerised with CI. Code at github.com/delcenjo/credit-risk-platform.
Transformer from scratch
A GPT-style language model implemented from scratch in PyTorch: multi-head attention, causal masking and residual blocks written by hand, plus a byte-pair tokenizer and an ablation study. Code at github.com/delcenjo/transformer-from-scratch.
Built with
