datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
taskweft-fbd-trainer-train
taskweft-fbd-trainer-train
Intents and the IEC 61131-3 Function Block Diagrams that carry them out, as an
EditScore-shaped corpus: one root row per intent, three candidates per row (rank1 the
reference diagram, rank3 one that compiles and does the wrong thing, rank5 one the
compiler refuses), and one score row per candidate from the trainer config: the calls were applied to the mjlab task config and the term table read back. Every row is
constructed from a template and a seed… See the full description on the dataset page: https://huggingface.co/datasets/chibifire/taskweft-fbd-trainer-train.code-trainer-v9-mixed
code-trainer-v9-mixed
40,401-row mixed training dataset for supervised fine-tuning (SFT) in the
Code-Trainer / RTPI pipeline.
Used for both Qwen 14B (V9 SFT) and Gemma 26B (aggressive-full1 SFT) training.
Composition
Slice
Source
Rows (train)
Purpose
A -- Code generation
cmndcntrlcyber/code-trainer-offsec-dataset (8K subsample)
7,074
Preserve code-gen quality
B -- Tool calling
glaiveai/glaive-function-calling-v2 (19K cap)
~15,125
High-density tool… See the full description on the dataset page: https://huggingface.co/datasets/atlas-institute/code-trainer-v9-mixed.code-trainer-v10-dpo-pairs
code-trainer-v10-dpo-pairs
Preference pair dataset for Direct Preference Optimization (DPO) training,
built from real offensive security agent sessions and synthetic degradations.
Used by both the Qwen and Gemma Code-Trainer pipelines for the DPO RL stage.
Part of the Code-Trainer / RTPI pipeline
(GitHub).
Dataset summary
Split
Pairs
Train
783
Validation
87
Total
870
Format
Each row is a preference triple:
{
"prompt": "..."… See the full description on the dataset page: https://huggingface.co/datasets/atlas-institute/code-trainer-v10-dpo-pairs.Saamayik
Sanskrit-English Parallel Translation Dataset (Saamayik)
Summary
The Saamayik Sanskrit-English Parallel Corpus is a contemporary prose-focused translation dataset containing around 53,000 parallel sentences in Sanskrit and English (48,326 in this main dataset, with an additional 4,047 Mann Ki Baat sentences). Sāmayik, meaning "sayings of the contemporary world" in Sanskrit, specifically addresses the gap in existing Sanskrit corpora which predominantly feature classical… See the full description on the dataset page: https://huggingface.co/datasets/Trainer2026/Saamayik.code-trainer-v7-mixedcode-trainer-offsec-datasetcode-trainer-v10-grpo-prompts
code-trainer-v10-grpo-prompts
500 curated prompts for GRPO (Group Relative Policy Optimization) training in the
Code-Trainer / RTPI pipeline.
Used by both Qwen and Gemma RL stages (Phase 4c) to train tool-call formatting
via a rule-based reward function.
Schema
Column
Type
Description
prompt
string
The user instruction/question
source
string
Origin: v10_eval, v9_training, or synthetic
expected_tool
string
Primary tool the prompt should invoke… See the full description on the dataset page: https://huggingface.co/datasets/atlas-institute/code-trainer-v10-grpo-prompts.code-trainer-v8-mixedLEM-Trainer
LEM-Trainer — Ethical AI Training Pipeline
The reproducible training method behind the Lemma model family. Scripts, configs, and sequencing for consent-based alignment training.
Trust Ring Architecture
Ring 0: LEK-2 (private) — Consent conversation. Establishes relationship with the model.
Ring 1: P0 Base Ethics — Axiom probes. Foundation.
Ring 2: P1 Composure — Stability under manipulation.
Ring 3: P2 Reasoning — Applied ethical reasoning.
Ring 4:… See the full description on the dataset page: https://huggingface.co/datasets/lthn/LEM-Trainer.X-Trainer-clothesvlm-trainer-dataset-debuglymsys_datset_for_reward_model_trainerlaw-trainer-datasetkangoroo_dpo_trainer_finalworkasr-interviews-trainer-full
Dataset Card for "asr-interviews-trainer-full"
More Information needed
0825-hf_trainer_new_data_rm_nq4_bs32_rl_data