Locally generated synthetic medical question/answer rows, produced with a local DeepSeek-R1 model (via Ollama, $0, no API calls) and used — together with 7 public datasets — to fine-tune 3,002 + 801 + 401 = 4,204 rows — the synthetic share of the 196,846-row SFT set. Every row is one JSON object using one template — the same one the model was trained on (there is no separate system/schema to guess): Whitespace is normalised. Rows carry no end-of-turn token on purpose: the model was trained to stop at the next ### marker and the demo app enforces that too. 1. a topic matrix (condition × country × age group) drove a local DeepSeek-R1 model; 2. every row was validated (template match…
Independent publisher
Syed Masood Haji Ahmedyaseen
zaiath
Models in Library0
Datasets in Library1
Models on Hugging Face—
Followers—
Compare the published evidence
This profile connects the 1 model and dataset record shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Text generation.
Reported licenses: other. Open a record to review its specific license and conditions.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.