{ "version": "full-ultra-volume-1", "generated_from": "output/full_ultra/dataset_final.parquet", "totals": { "datasets": 51, "rows": 31180492, "available_rows": 29004908, "sampled_rows": 432722, "source_groups": { "v3": { "datasets": 28, "rows": 25525218, "sampled_rows": 350015 }, "ultra": { "datasets": 23, "rows": 5655274, "sampled_rows": 82707 } } }, "datasets": [ { "id": "nemotron-math-v2", "dataset": "Math-v2", "dataset_full_name": "Nemotron-Math-v2", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 7085839, "available_rows": 7085839, "sampled_rows": 96558, "url": "https://huggingface.co/datasets/nvidia/Nemotron-Math-v2" }, { "id": "nemotron-competitive-programming-v1", "dataset": "CompProg-v1", "dataset_full_name": "Nemotron-Competitive-Programming-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 3927984, "available_rows": 1792697, "sampled_rows": 53521, "url": "https://huggingface.co/datasets/nvidia/Nemotron-Competitive-Programming-v1" }, { "id": "nemotron-sft-math-v3", "dataset": "SFT-Math-v3", "dataset_full_name": "Nemotron-SFT-Math-v3", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 3638783, "available_rows": 3638783, "sampled_rows": 49581, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Math-v3" }, { "id": "nemotron-sft-multilingual-v1", "dataset": "Multilingual", "dataset_full_name": "Nemotron-SFT-Multilingual-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 3065255, "available_rows": 3065006, "sampled_rows": 41773, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Multilingual-v1" }, { "id": "nemotron-sft-science-v2", "dataset": "SFT-Science-v2", "dataset_full_name": "Nemotron-SFT-Science-v2", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 2837680, "available_rows": 2837680, "sampled_rows": 38910, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Science-v2" }, { "id": "nemotron-sft-instruction-following-chat-v2", "dataset": "SFT-Chat-IF-v2", "dataset_full_name": "Nemotron-SFT-Instruction-Following-Chat-v2", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 1997510, "available_rows": 1992667, "sampled_rows": 27219, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Instruction-Following-Chat-v2" }, { "id": "nemotron-math-proofs-v1", "dataset": "Math-Proofs", "dataset_full_name": "Nemotron-Math-Proofs-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 1376663, "available_rows": 1376663, "sampled_rows": 18759, "url": "https://huggingface.co/datasets/nvidia/Nemotron-Math-Proofs-v1" }, { "id": "nemotron-sft-competitive-programming-v2", "dataset": "SFT-CompProg-v2", "dataset_full_name": "Nemotron-SFT-Competitive-Programming-v2", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 841555, "available_rows": 841555, "sampled_rows": 11468, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Competitive-Programming-v2" }, { "id": "nemotron-sft-math-v4", "dataset": "SFT-Math-v4", "dataset_full_name": "Nemotron-SFT-Math-v4", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 545425, "available_rows": 545425, "sampled_rows": 7479, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Math-v4" }, { "id": "nemotron-sft-instruction-following-chat-v3", "dataset": "SFT-Chat-IF-v3", "dataset_full_name": "Nemotron-SFT-Instruction-Following-Chat-v3", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 530575, "available_rows": 530575, "sampled_rows": 7275, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Instruction-Following-Chat-v3" }, { "id": "nemotron-rl-super-training-blends", "dataset": "RL-Super-Blend", "dataset_full_name": "Nemotron-RL-Super-Training-Blends", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 477859, "available_rows": 464860, "sampled_rows": 6512, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Super-Training-Blends" }, { "id": "nemotron-sft-opencode-v1", "dataset": "SFT-OpenCode", "dataset_full_name": "Nemotron-SFT-OpenCode-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 460254, "available_rows": 460254, "sampled_rows": 6272, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-OpenCode-v1" }, { "id": "nemotron-instruction-following-chat-v1", "dataset": "Chat-IF-v1", "dataset_full_name": "Nemotron-Instruction-Following-Chat-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 430978, "available_rows": 430977, "sampled_rows": 5872, "url": "https://huggingface.co/datasets/nvidia/Nemotron-Instruction-Following-Chat-v1" }, { "id": "nemotron-rl-ultra-training-blends", "dataset": "RL-Ultra-Blend", "dataset_full_name": "Nemotron-RL-Ultra-Training-Blends", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 394142, "available_rows": 394142, "sampled_rows": 5404, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Ultra-Training-Blends" }, { "id": "nemotron-sft-multilingual-v2", "dataset": "SFT-Multilingual-v2", "dataset_full_name": "Nemotron-SFT-Multilingual-v2", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 370081, "available_rows": 370081, "sampled_rows": 5074, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Multilingual-v2" }, { "id": "nemotron-agentic-v1", "dataset": "Agentic", "dataset_full_name": "Nemotron-Agentic-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 335122, "available_rows": 335121, "sampled_rows": 4567, "url": "https://huggingface.co/datasets/nvidia/Nemotron-Agentic-v1" }, { "id": "nemotron-specializeddomains-finance-v1", "dataset": "Finance", "dataset_full_name": "Nemotron-SpecializedDomains-Finance-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 326698, "available_rows": 326698, "sampled_rows": 4452, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SpecializedDomains-Finance-v1" }, { "id": "nemotron-rlhf-genrm-v1", "dataset": "RLHF-GenRM", "dataset_full_name": "Nemotron-RLHF-GenRM-v1", "source_group": "v3", "pipeline_stage": "Reward Modeling", "stage_short": "RM", "rows": 299517, "available_rows": 299509, "sampled_rows": 4082, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RLHF-GenRM-v1" }, { "id": "nemotron-sft-agentic-v2", "dataset": "SFT-Agentic-v2", "dataset_full_name": "Nemotron-SFT-Agentic-v2", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 293292, "available_rows": 293291, "sampled_rows": 3997, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Agentic-v2" }, { "id": "nemotron-sft-swe-v2", "dataset": "SFT-SWE-v2", "dataset_full_name": "Nemotron-SFT-SWE-v2", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 256254, "available_rows": 256248, "sampled_rows": 3492, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-SWE-v2" }, { "id": "nemotron-sft-arc-agi-v1", "dataset": "SFT-ARC-AGI-v1", "dataset_full_name": "Nemotron-SFT-ARC-AGI-v1", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 252069, "available_rows": 252069, "sampled_rows": 3456, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-ARC-AGI-v1" }, { "id": "nemotron-sft-swe-v3", "dataset": "SFT-SWE-v3", "dataset_full_name": "Nemotron-SFT-SWE-v3", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 237970, "available_rows": 237970, "sampled_rows": 3263, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-SWE-v3" }, { "id": "nemotron-science-v1", "dataset": "Science", "dataset_full_name": "Nemotron-Science-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 226334, "available_rows": 226334, "sampled_rows": 3084, "url": "https://huggingface.co/datasets/nvidia/Nemotron-Science-v1" }, { "id": "nemotron-rl-science-v1", "dataset": "RL-Science-v1", "dataset_full_name": "Nemotron-RL-Science-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 150644, "available_rows": 150644, "sampled_rows": 2066, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Science-v1" }, { "id": "nemotron-sft-safety-v2", "dataset": "SFT-Safety-v2", "dataset_full_name": "Nemotron-SFT-Safety-v2", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 130244, "available_rows": 130244, "sampled_rows": 1786, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Safety-v2" }, { "id": "nemotron-rl-agentic-conversational-tool-use-pivot-v1", "dataset": "RL-Conv-Tools", "dataset_full_name": "Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 96968, "available_rows": 96968, "sampled_rows": 1322, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1" }, { "id": "nemotron-3-nano-rl-training-blend", "dataset": "Nano-RL-Blend", "dataset_full_name": "Nemotron-3-Nano-RL-Training-Blend", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 93244, "available_rows": 71188, "sampled_rows": 1271, "url": "https://huggingface.co/datasets/nvidia/Nemotron-3-Nano-RL-Training-Blend" }, { "id": "nemotron-rl-safety-v1", "dataset": "RL-Safety", "dataset_full_name": "Nemotron-RL-Safety-v1", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 89068, "available_rows": 88984, "sampled_rows": 1213, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Safety-v1" }, { "id": "nemotron-sft-math-proofs-v4", "dataset": "SFT-Math-Proofs-v4", "dataset_full_name": "Nemotron-SFT-Math-Proofs-v4", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 82737, "available_rows": 82737, "sampled_rows": 1134, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Math-Proofs-v4" }, { "id": "nemotron-rl-instruction-following-structured-outputs-v2", "dataset": "RL-IF-Structured-v2", "dataset_full_name": "Nemotron-RL-Instruction-Following-Structured-Outputs-v2", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 62694, "available_rows": 62694, "sampled_rows": 860, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Structured-Outputs-v2" }, { "id": "nemotron-swe-v1", "dataset": "SWE-v1", "dataset_full_name": "Nemotron-SWE-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 51029, "available_rows": 51029, "sampled_rows": 695, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SWE-v1" }, { "id": "nemotron-rl-agentic-swe-pivot-v1", "dataset": "RL-SWE-Pivot", "dataset_full_name": "Nemotron-RL-Agentic-SWE-Pivot-v1", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 50661, "available_rows": 50661, "sampled_rows": 690, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-SWE-Pivot-v1" }, { "id": "nemotron-sft-safety-v1", "dataset": "Safety", "dataset_full_name": "Nemotron-SFT-Safety-v1", "source_group": "v3", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 45145, "available_rows": 45102, "sampled_rows": 615, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-Safety-v1" }, { "id": "nemotron-rl-identity-following-v1", "dataset": "RL-Identity", "dataset_full_name": "Nemotron-RL-Identity-Following-v1", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 21660, "available_rows": 21660, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Identity-Following-v1" }, { "id": "nemotron-rl-arc-agi-v1", "dataset": "RL-ARC-AGI-v1", "dataset_full_name": "Nemotron-RL-ARC-AGI-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 21028, "available_rows": 21028, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-ARC-AGI-v1" }, { "id": "nemotron-rl-reasoninggym-v1", "dataset": "RL-ReasoningGym", "dataset_full_name": "Nemotron-RL-ReasoningGym-v1", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 15000, "available_rows": 14994, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-ReasoningGym-v1" }, { "id": "nemotron-rl-instruction-following-calendar-v2", "dataset": "RL-IF-Calendar", "dataset_full_name": "Nemotron-RL-Instruction-Following-Calendar-v2", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 9915, "available_rows": 9915, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Calendar-v2" }, { "id": "nemotron-rl-agentic-function-calling-pivot-v1", "dataset": "RL-FuncCall", "dataset_full_name": "Nemotron-RL-Agentic-Function-Calling-Pivot-v1", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 9620, "available_rows": 9620, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1" }, { "id": "nemotron-rl-instruction-following-citation-formatting-v1", "dataset": "RL-IF-Citation-v1", "dataset_full_name": "Nemotron-RL-Instruction-Following-Citation-Formatting-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 9540, "available_rows": 9540, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Citation-Formatting-v1" }, { "id": "nemotron-rl-instruction-following-free-form-formatting-v1", "dataset": "RL-IF-FreeForm-v1", "dataset_full_name": "Nemotron-RL-Instruction-Following-Free-Form-Formatting-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 9036, "available_rows": 9036, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Free-Form-Formatting-v1" }, { "id": "nemotron-rl-litmus-bench-v0-1", "dataset": "RL-Litmus-v0.1", "dataset_full_name": "Nemotron-RL-Litmus-Bench-v0.1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 5714, "available_rows": 5714, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Litmus-Bench-v0.1" }, { "id": "nemotron-rl-math-v4", "dataset": "RL-Math-v4", "dataset_full_name": "Nemotron-RL-Math-v4", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 3748, "available_rows": 3748, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Math-v4" }, { "id": "nemotron-rl-qa-abstention-v1", "dataset": "RL-QA-Abstention-v1", "dataset_full_name": "Nemotron-RL-QA-Abstention-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 3150, "available_rows": 3150, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-QA-Abstention-v1" }, { "id": "nemotron-sft-cuda-v1", "dataset": "SFT-CUDA-v1", "dataset_full_name": "Nemotron-SFT-CUDA-v1", "source_group": "ultra", "pipeline_stage": "Supervised Fine-Tuning", "stage_short": "SFT", "rows": 2276, "available_rows": 2276, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-SFT-CUDA-v1" }, { "id": "nemotron-rl-multichallenge-v1", "dataset": "RL-Multichallenge-v1", "dataset_full_name": "Nemotron-RL-Multichallenge-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 2118, "available_rows": 2118, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Multichallenge-v1" }, { "id": "nemotron-rl-instruction-following-multiturnchat-v1", "dataset": "RL-IF-MultiTurn", "dataset_full_name": "Nemotron-RL-Instruction-Following-MultiTurnChat-v1", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 2011, "available_rows": 2011, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-MultiTurnChat-v1" }, { "id": "nemotron-rl-agentic-indirect-prompt-injection-v1", "dataset": "RL-Agentic-IPI-v1", "dataset_full_name": "Nemotron-RL-Agentic-Indirect-Prompt-Injection-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 1272, "available_rows": 1272, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Indirect-Prompt-Injection-v1" }, { "id": "nemotron-rl-cfbench-v1", "dataset": "RL-CFBench-v1", "dataset_full_name": "Nemotron-RL-CFBench-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 1121, "available_rows": 1121, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-CFBench-v1" }, { "id": "nemotron-rl-sysbench-v1", "dataset": "RL-SysBench-v1", "dataset_full_name": "Nemotron-RL-SysBench-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 1010, "available_rows": 1010, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-SysBench-v1" }, { "id": "nemotron-rl-instruction-following-adversarial-v1", "dataset": "RL-IF-Adversarial", "dataset_full_name": "Nemotron-RL-Instruction-Following-Adversarial-v1", "source_group": "v3", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 1000, "available_rows": 1000, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Adversarial-v1" }, { "id": "nemotron-rl-inverseifeval-v1", "dataset": "RL-InverseIFEval-v1", "dataset_full_name": "Nemotron-RL-InverseIFEval-v1", "source_group": "ultra", "pipeline_stage": "Reinforcement Learning", "stage_short": "RL", "rows": 1000, "available_rows": 1000, "sampled_rows": 500, "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-InverseIFEval-v1" } ] }