datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SFT-UZ-9kThis is SFT version of MLDataScientist/DPO-uz-9k (Uzbek translated dataset with two answers for each prompt for DPO fine-tuning).
I selected ['answer'][0] from each example and saved them in this dataset for easy fine-tuning of LLMs.
DPO-uz-9kThis is DPO Uzbek translated dataset with 9k chat pairs.
Original English dataset comes from DPO-En-Zh-20k (commit 9ad5f7428419d3cf78493cf3f4be832cf5346ba8. File: dpo_en.json).
I translated 10k pairs of chat examples into Uzbek using NLLB 3.3B model.
After translation was completed, I used local lilac instance to remove records with coding examples since NLLB is not good at translating text with coding examples.
Note that each prompt has two answers. The first answer should be the 'selected'… See the full description on the dataset page: https://huggingface.co/datasets/MLDataScientist/DPO-uz-9k.MLDSUM_NEWml-design-doc-reviewer-data
ml-system-design/ml-design-doc-reviewer-data (v1.0.0)
Evaluation artifacts for the ML Design Doc Reviewer project.
Layout
Path
Description
manifest/sample_manifest.csv
Stratified 100-case sample manifest
manifest/error_topology.csv
Controlled error taxonomy for flawed docs
raw/
Raw markdown exports, metadata sidecars, OCR image blocks
raw/images/
Downloaded article images
normalized/
Canonical 14-section ML design documents
flawed/
Normalized… See the full description on the dataset page: https://huggingface.co/datasets/ml-system-design/ml-design-doc-reviewer-data.