OncoAgent Clinical: 266K Oncology Training Samples

266,854 clinical oncology training samples curated for fine-tuning large language models. The dataset was created by lablab-ai-amd-developer-hackathon and last updated on May 9, 2026. It is composed of multi-source data, including approximately 100,000 samples from PMC-Patients.

Use Cases

• Fine-tuning LLMs for cancer diagnosis based on clinical case presentations.
• Training models for oncology treatment recommendation using curated clinical samples.
• Developing clinical reasoning agents for oncology tasks.
• Benchmarking medical AI models on multi-source oncology data.

Strengths

• Contains 266,854 training samples, providing a substantial corpus for model training.
• Includes approximately 100,000 samples from the PMC-Patients source of real clinical case presentations.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

lablab-ai-amd-developer-hackathon

Collection Method

Curated from multiple sources, including PMC-Patients.

Freshness

Last updated 2026-05-09 06:58:39; freshness should be verified.