datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
testing_codealpaca_small
Dataset Card for "testing_codealpaca_small"
More Information needed
code-comments-small
Comment Dataset
Opening comments extracted from code datasets with CommentMiner and ML4SE-toolkit.
Files are grouped as <dataset>/<language>/part-*.parquet.
The Hugging Face dataset card declares one config per source dataset and one split-safe language name per language.
Each row contains dataset, record_id, opening_comment, language, path, repo, extracted_at, and metadata.
For Parquet exports, metadata is stored as a JSON string so every source dataset shares one stable… See the full description on the dataset page: https://huggingface.co/datasets/Jkatzy/code-comments-small.motherlode-code-small-en-v0.1-evidence
Evidence for motherlode-code-small-en-v0.1
Every figure on the card of RiverRider/motherlode-code-small-en-v0.1
comes from a file here. Each figure also has a row in Sunstone North Labs' claims ledger, named in the first column. The
table maps each row to its file and field. The per-instance and per-query files let you recompute a paired test without
our code or our hardware.
"The soup" is the model: the parameter-wise mean of two fine-tunes of bge-small-en-v1.5, a2 and d1, at… See the full description on the dataset page: https://huggingface.co/datasets/RiverRider/motherlode-code-small-en-v0.1-evidence.hardware_code_and_sec_smalltokenized-codeparrot-ds-smallagent-trace-privacy-scrubber-codex-tracesRTL-Coder_small
RTL-Coder_small
For implementation details, visit our GitHub repository: VeriReason
Check out our paper: VeriReason: Reinforcement Learning with Testbench Feedback for Reasoning-Enhanced Verilog Generation
Update Log
2025.05.17: Initial release of Nellyw888/Verireason-RTL-Coder_7b_reasoning_tb
Project Description
This study introduces VeriReason, a novel approach utilizing reinforcement learning with testbench feedback to enhance the performance of pre-trained… See the full description on the dataset page: https://huggingface.co/datasets/Nellyw888/RTL-Coder_small.small_repos_multi_file_chatgpt_5_qas_part5_code_qa-datasethackathon-advisor-codex-traces
Hackathon Advisor Codex Session Traces
Real Codex session logs for the Hackathon Advisor project, selected from local Codex
rollout JSONL files and redacted before publication. The event stream preserves user
requests, assistant messages, tool calls, tool outputs, browser/search events, and
minimal session provenance needed to audit how the project was built.
Privacy filtering
The publisher applied openai/privacy-filter
at revision… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/hackathon-advisor-codex-traces.code-ita-dpo-small
Dataset Card for "code-instructions-ita-dpo-small"
More Information needed
lost-found-desk-codex-traces
Lost & Found Desk Codex Trace Dataset
This dataset is an official-format Codex trace artifact for the Codex-assisted Build Small hackathon submission of Lost & Found Desk.
It follows the Hugging Face Agent Traces guidance: Codex sessions are published as JSONL files under traces/, preserving the Codex session event schema so the Hub trace viewer can open the session. For public release, the trace is redacted in-place: local absolute paths, token-shaped strings, and secret-label… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/lost-found-desk-codex-traces.kicky-ai-codex-trace
Kicky AI - Codex agent trace (sanitized)
A redacted OpenAI Codex CLI session trace from building
Kicky AI for the Build Small
Hackathon - shared for the Sharing is Caring badge so others can see how the build went.
Format: Codex CLI JSONL session log (each record = {payload, timestamp, type}).
All secrets removed (HF / Modal / Roboflow tokens, shared secrets, emails) - verified 0 leaks.
Blog write-up: https://dcrey7.substack.com/p/world-fut-coach
NeuroBait-Codex-Traces
Codex Session Traces
This folder contains Codex rollout JSONL traces related to the
NeuroBait Build Small Model project.
Included traces:
rollout-2026-06-08T17-03-23-019ea6af-db29-7801-ac01-46dfc88f90b0.jsonl
rollout-2026-06-09T07-10-21-019ea9b7-4610-7223-906e-2d0dba8bae7f.jsonl
rollout-2026-06-09T16-00-29-019eab9c-a18a-7de1-8967-ea63db425a4f.jsonl
The traces were selected because their session metadata contains the project
working directory:… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/NeuroBait-Codex-Traces.code-code-CodeRefinement-Java-Small
Dataset is imported from CodeXGLUE and pre-processed using their script.
Where to find in Semeru:
The dataset can be found at /nfs/semeru/semeru_datasets/code_xglue/code-to-code/code-refinement/data/small in Semeru
Task Definition
Code refinement aims to automatically fix bugs in the code, which can contribute to reducing the cost of bug-fixes for developers.
In CodeXGLUE, given a piece of Java code with bugs, the task is to remove the bugs to output the… See the full description on the dataset page: https://huggingface.co/datasets/semeru/code-code-CodeRefinement-Java-Small.High-Coder-SFT-Small
High-Coder-SFT-Small
A high-quality synthetic code dataset containing 54,950 long-form code samples across 8 programming languages. Generated using Hunter Alpha (1T+ parameter frontier model). Every single sample contains at least 200 lines of actual code — most contain 500+.
This is not a snippet dataset. Every file is a complete, production-quality source file with imports, error handling, design patterns, and modern language idioms. The average sample is 630 lines of code… See the full description on the dataset page: https://huggingface.co/datasets/Crownelius/High-Coder-SFT-Small.codeflow-agent-traces
CodeFlow — generation traces
Generation traces from CodeFlow, a code-to-flowchart generator built for the
Build Small Hackathon 2026. CodeFlow turns a code snippet into a readable
Mermaid.js control-flow diagram — generated by a 30B
coder model running entirely on CPU via llama.cpp, with every node wired back
to the source lines it came from.
Each trace is a complete witness of one end-to-end generation: the exact code the
user pasted, the model's hidden reasoning, the raw model… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/codeflow-agent-traces.fao-species-codes-handwritten-small
FAO Species Codes Handwritten Dataset
This dataset contains images of handwritten FAO species codes.
The collection includes 600 samples of 121 distinct species codes (roughly 5 samples per species)
These species codes are focused on those most common in the Western and Central Pacific Ocean.
Contact
For queries or collaborations related to this dataset, contact corey.cole+ocr@gmail.com
Datset Creation
Purpose
This dataset was created to enable the… See the full description on the dataset page: https://huggingface.co/datasets/coredog64/fao-species-codes-handwritten-small.best-man-speech-codex-traces
Best Man Speech Codex Traces
This dataset contains a sanitized Codex agent trace from development work on Best Man Speech
Coach, a Gradio app built for the Hugging Face Build Small Hackathon.
The trace captures a real software engineering session involving repository inspection,
implementation planning, code edits, validation checks, GitHub pull request context, and follow-up
project hygiene around sharing agent traces.
Contents… See the full description on the dataset page: https://huggingface.co/datasets/build-small-hackathon/best-man-speech-codex-traces.ds4sd-synth-code-net-smallmarc-code-mixed-small
marc-code-mixed-small
This dataset is based on The Multilingual Amazon Reviews Corpus.
It contains German (DE), English (EN), Spanish (ES), and French (FR) languages.
The labels are 0 (DE), 1 (EN), 2 (ES), and 3 (FR).
Each review contains all four languages.
Total number of tokens:
In training set: 10195342
In test set: 842760
In validation set: 842760
code-retrieval-stackoverflow-smalllca-codegen-small
LCA Project Level Code Completion
How to load the dataset
from datasets import load_dataset
ds = load_dataset('JetBrains-Research/lca-codegen-small', split='test')
Data Point Structure
repo – repository name in format {GitHub_user_name}__{repository_name}
commit_hash – commit hash
completion_file – dictionary with the completion file content in the following format:
filename – filepath to the completion file
content – content of the completion file… See the full description on the dataset page: https://huggingface.co/datasets/JetBrains-Research/lca-codegen-small.small_repos_multi_file_chatgpt_5_qas_part4_code_qa-datasetsmall_repos_multi_file_chatgpt_5_qas_code_qa_1k-datasetsmall_repos_multi_file_chatgpt_5_qas_part2_3_code_qa-datasetsmall-thoughts-code-try-run
Dataset card for small-thoughts-code-try-run
This dataset was made with Curator.
Dataset details
A sample from the dataset:
{
"question": "Generate an executable Python function generated from the given prompt. The function should take stdin as input and print the output. Simply call the function after the definition.Fox Ciel has some flowers: *r* red flowers, *g* green flowers and *b* blue flowers. She wants to use these flowers to make several bouquets. There… See the full description on the dataset page: https://huggingface.co/datasets/SmallDoge/small-thoughts-code-try-run.code-smallsmall_code_outputmath_dataset_smalldetails_microsoft__CodeGPT-small-py
Dataset Card for Evaluation run of microsoft/CodeGPT-small-py
Dataset Summary
Dataset automatically created during the evaluation run of model microsoft/CodeGPT-small-py on the Open LLM Leaderboard.
The dataset is composed of 3 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_microsoft__CodeGPT-small-py.