diff --git a/LICENSE.md b/LICENSE.md index 73ad3714d3139f6edadcd0cf5b3f9c0c3add0cec..fdc375ff2a4fa8fcfbe435bee38e0d7d6b797fdc 100644 --- a/LICENSE.md +++ b/LICENSE.md @@ -1,6 +1,6 @@ # Licenses and attribution -FastBrain-Instruct is a collection of upstream material and adaptations. Each component retains its applicable terms. This file does not replace those licenses or apply a blanket MIT license to the collection. The `source`, `source_repo` and `provenance_json` fields identify components in the mixed choice view. [Pinned revisions and acquired-file hashes](audit/source_manifest.json) accompany this notice. +FastBrain-Instruct is a collection of upstream material and adaptations. Each component retains its applicable terms. This file does not replace those licenses or apply a blanket MIT license to the collection. The `source`, `source_repo` and `provenance_json` fields identify components in the mixed choice view. [Pinned revisions and acquired-file hashes](audit/v3/input_manifest.json) accompany this notice. | Component | Attribution | Applicable terms and source | |---|---|---| @@ -10,12 +10,28 @@ FastBrain-Instruct is a collection of upstream material and adaptations. Each co | Nemotron interactive-agent trajectories, through AgentToolDecisions | NVIDIA Corporation; decision-format transformation by Maziyar Panahi | [CC BY 4.0 legal text](licenses/CC-BY-4.0.txt); [NVIDIA's pinned terms](https://huggingface.co/datasets/nvidia/Nemotron-SFT-Agentic-v2/blob/7c804833427f633ccd53b582dbf02525fd680f78/README.md) and [derived source](https://huggingface.co/datasets/MaziyarPanahi/AgentToolDecisions-180K/blob/22b91136648789a2881183dfc2f77d756659444c/README.md). NVIDIA additionally lists Apache 2.0 and MIT for components; those notices are not removed or relicensed here. | | NLU++ English hotel utterances, intent modules and slots | PolyAI / NLU++ authors | [CC BY 4.0](licenses/CC-BY-4.0.txt), copied from the [pinned original LICENSE](https://github.com/PolyAI-LDN/task-specific-datasets/blob/57ec275d8078af65b7731c2a98be812d844a6d6b/LICENSE); [source data](https://github.com/PolyAI-LDN/task-specific-datasets/tree/57ec275d8078af65b7731c2a98be812d844a6d6b/nlupp) | -Changes in v2 include filtering, grouping and split isolation; assistant-reviewed intent descriptions and alias mapping; candidate sampling and paired target removal; restored available policy/context; and schema/metadata reformatting. Source utterances are retained. The MTOP/TOPv2 derived semantic data, including the adapted catalog, is distributed under CC BY-SA 4.0. Attribution does not imply upstream endorsement of this adaptation. +The v3 adaptation includes filtering, grouping and split isolation; assistant-reviewed intent descriptions and alias mapping; candidate sampling and paired target removal; restored available policy/context; and schema/metadata reformatting. Source utterances are retained. The MTOP/TOPv2 derived semantic data, including the adapted catalog, is distributed under CC BY-SA 4.0. Attribution does not imply upstream endorsement of this adaptation. -The frozen `default` and retained legacy replay include Google Research SGD (CC BY-SA 4.0), Amazon MASSIVE (CC BY 4.0), Glaive v2 and NousResearch Hermes (Apache 2.0 as recorded in the original card), attributed Nemotron records, NL2Bash (consult its source terms), Git natural-language commands (consult its source terms), and maintainer DevOps templates (Apache 2.0). Their original links and notices remain in the [archived legacy data card](docs/legacy-data-card.md). The v2 release does not claim to resolve unspecified legacy-source terms. The four original data files are unchanged from `a84ea0337a424e287dc6f36f6b9fa320fb920fb0`. +## Retained replay and recovered-source attribution -## v3 additions and modifications +The following sources contribute to `legacy_replay_v3` or the native Hermes view, and therefore also to the corresponding integrated v3 choice records where eligible. These notices preserve the source links and terms recorded in the original attribution. Removing superseded FastBrain release files does not remove upstream rights or attribution. + +| Component | Attribution | Applicable terms and source | +|---|---|---| +| Schema-Guided Dialogue (SGD) | Google Research; Rastogi et al., AAAI 2020 | CC BY-SA 4.0; [original source and notices](https://github.com/google-research-datasets/dstc8-schema-guided-dialogue) | +| MASSIVE English intents | Amazon Science; FitzGerald et al., 2022 | CC BY 4.0; [original source](https://huggingface.co/datasets/AmazonScience/massive), via [the intent mirror](https://huggingface.co/datasets/mteb/amazon_massive_intent) | +| Glaive Function Calling v2 | Glaive AI | Apache 2.0 as recorded in the original attribution; [source and terms](https://huggingface.co/datasets/glaiveai/glaive-function-calling-v2) | +| Hermes Function Calling v1, including recovered requests, schemas and call sets | NousResearch | Apache 2.0 as recorded in the original attribution; [source and terms](https://huggingface.co/datasets/NousResearch/hermes-function-calling-v1); [pinned row mapping](curation/v3/hermes_source_mapping.json) | +| NL2Bash | Lin et al., LREC 2018; Hugging Face mirror by AnishJoshi | Consult the [source terms](https://huggingface.co/datasets/AnishJoshi/nl2bash-custom); no new license is assigned here | +| Git natural-language commands | burak29 | Consult the [source terms](https://huggingface.co/datasets/burak29/git-natural-language-commands); no new license is assigned here | +| DevOps / terminal templates | FastBrain-Instruct maintainer | Apache 2.0, as recorded in the original attribution | + +This release does not resolve unspecified legacy-source terms. NVIDIA / Maziyar Panahi attribution and terms are retained in the component table above. Glaive and Nemotron source material includes synthetic generation. All applicable upstream terms continue to apply. + +## Modifications and HINT3 attribution The repaired semantic descriptions, conflict catalog and MTOP/TOPv2 examples retain CC BY-SA 4.0. Recovered Hermes requests, call sets and schemas retain the original NousResearch source terms; native Nemotron tool schemas retain the NVIDIA component terms listed above. Repairs include filtering, descriptor clarification, candidate regeneration, source recovery, additional group reservations and explicit supervision masks. These changes do not certify upstream outputs or imply author endorsement. HINT3 v2 (Curekart and SOFMattress), by Haptik and Gaurav Arora, Chirag Jain, Manas Chaturvedi and Krupal Modi, is provided as a separate native auxiliary database. The database is under ODbL 1.0 and individual contents under DbCL 1.0. The adapted HINT3 view retains those terms. See the [verbatim original notice](licenses/HINT3-ODbL-DbCL.txt), [attribution and source links](licenses/HINT3-attribution.md), and [pinned file hashes](curation/v3/hint3_source_manifest.json). HINT3 labels are not silently relicensed as MIT or merged into the choice catalog. + +The [inherited source-acquisition manifest](audit/v3/inherited_source_manifest.json) preserves the earlier acquisition revisions and hashes. It is provenance for v3, not a separate downloadable FastBrain release. diff --git a/README.md b/README.md index a7af9b6fb42927fd28aa4e083f2a9330a7d7b59e..6cfbb371233b446181f107f7afae9096dac9f473 100644 --- a/README.md +++ b/README.md @@ -35,53 +35,6 @@ source_datasets: size_categories: - 100K [!WARNING] -> ### ⚠️ Provenance: not an original dataset -> -> **FastBrain-Instruct is NOT written from scratch.** It aggregates existing open datasets, reformats them into one choice/routing schema, removes finance content, deduplicates, and decontaminates against public decision benchmarks. All query text and labels come from the upstream sources below (except 25 maintainer-written DevOps templates). -> -> | Upstream source | Author | License | Rows (train / test) | -> | :--- | :--- | :--- | ---: | -> | [Schema-Guided Dialogue (SGD)](https://github.com/google-research-datasets/dstc8-schema-guided-dialogue) | Google Research (Rastogi et al., 2020) | CC BY-SA 4.0 | 34,455 / 3,889 | -> | [Nemotron-SFT-Agentic-v2](https://huggingface.co/datasets/nvidia/Nemotron-SFT-Agentic-v2), via the `tool_selection` subset of [AgentToolDecisions-180K](https://huggingface.co/datasets/MaziyarPanahi/AgentToolDecisions-180K) | NVIDIA; transformed by Maziyar Panahi | CC BY 4.0 | 32,500 / 3,548 | -> | [MASSIVE](https://huggingface.co/datasets/AmazonScience/massive) (en, via `mteb/amazon_massive_intent`) | Amazon Science (FitzGerald et al., 2022) | CC BY 4.0 | 14,264 / 1,531 | -> | [Glaive Function Calling v2](https://huggingface.co/datasets/glaiveai/glaive-function-calling-v2) | Glaive AI | Apache 2.0 | 11,575 / 1,305 | -> | [NL2Bash](https://huggingface.co/datasets/AnishJoshi/nl2bash-custom) | Lin et al., 2018 (HF mirror by AnishJoshi) | see source | 2,809 / 345 | -> | [Hermes Function Calling v1](https://huggingface.co/datasets/NousResearch/hermes-function-calling-v1) | NousResearch | Apache 2.0 | 1,015 / 132 | -> | [Git natural-language commands](https://huggingface.co/datasets/burak29/git-natural-language-commands) | burak29 | see source | 810 / 75 | -> | DevOps / terminal templates | maintainer | Apache 2.0 | 22 / 3 | -> -> Glaive v2 and Nemotron-SFT-Agentic-v2 are themselves synthetically generated. Users must comply with each upstream license; SGD is CC BY-SA 4.0 (share-alike). - ---- - -## 📌 What it is - -Each row is a single decision: given a `query` (a user request, or for agent rows the recent conversation) and a list of candidate `choices` (name + description), pick the correct `target`. The format targets non-autoregressive "System 1" decision heads (e.g., the `JointSchemaHead` design used by Cloudflare Clef / Clef-Flash) that score candidates in one forward pass. - -Two domains, balanced 50/50: -- **`code_mechanical`** — tool / function selection (Glaive, Hermes), multi-tool agent next-tool selection (Nemotron), terminal utility selection (NL2Bash), git subcommand selection, DevOps templates. -- **`daily_assistant`** — intent selection for voice-assistant and service dialogues (MASSIVE, SGD). - -No model trained on this dataset is released, and no benchmark results are claimed. - -## 🚫 Benchmark exclusion & decontamination - -This dataset is intended for training models that will be compared on the benchmarks used in the Cloudflare Clef / Jev Decision Index table. Therefore: - -1. **No benchmark is used as a source.** Excluded entirely: BFCL, ToolRet and all 35 of its source datasets (incl. APIGen/xLAM, ToolACE, ToolBench, Gorilla, MetaTool, API-Bank, …), API-Bank, When2Call (and derived data such as the When2Call part of AgentToolDecisions), BANKING77, CLINC150, BRIGHT, Amazon ESCI, PhishNChips. -2. **Text-level decontamination.** Every query (for agent rows: every user message) was checked against the texts of BFCL v3, ToolRet-Queries (35 subsets), When2Call (all files), API-Bank (all files), BANKING77 and CLINC150 (all configs). A row is removed on a normalized exact match **or** any shared 10-word n-gram. 2,674 rows were removed this way (e.g., Glaive: 899 ToolRet, 470 BFCL, 333 When2Call matches; SGD: 420 BFCL matches; MASSIVE: 212 CLINC150 matches). - -Not checked at text level: Home appliances (no public copy found), BRIGHT, Amazon ESCI, PhishNChips (different domains from all sources used). - -## 🔧 Processing - -1. **Extraction** - - Glaive v2: all 112,960 rows scanned; the user turn right before the first `` is the query; candidates are the tools listed in that sample's system prompt. - - Hermes v1: first human turn; candidates are that sample's tools. - - Nemotron agentic: conversation rendered as `role: content` lines (last 1,500 chars); candidates are all tools available to the agent (5–30). - - NL2Bash: only single, non-compound commands (no pipes, `;`, `&&`, substitutions, loops); target is the utility name; `find` capped at 1,500 rows. - - SGD: user turns where a service's active intent changes; candidates are the intents of that service plus intents from other services, with SGD's own intent descriptions. - - MASSIVE: all splits; description generated from the label name. -2. **Candidates**: at least 5 per row; when a sample has fewer own candidates, random distractors from the same source are added, excluding names that heavily overlap with the target (token Jaccard ≥ 0.5). -3. **Finance removal**: MASSIVE `qa_currency`, `qa_stock`; SGD `Banks_*` and `Payment_*` services. -4. **Deduplication**: normalized exact `query` duplicates removed; queries with conflicting labels dropped (1,410). -5. **Balancing**: the larger domain is downsampled (largest source first) to match the smaller one. -6. **Split**: random 90/10 over unique queries; no `query` appears in both splits. - -## 📊 Statistics - -| | Train | Test | Total | -| :--- | ---: | ---: | ---: | -| **Rows** | **97,450** | **10,828** | **108,278** | -| `code_mechanical` | — | — | 54,139 (50%) | -| `daily_assistant` | — | — | 54,139 (50%) | -| Choices per row | 5–30 (median 5) | | | -| Distinct targets | | | 5,681 | - -## ⚠️ Limitations - -- **Easy distractors in some sources**: Glaive and NL2Bash rows usually have only one "own" candidate, so the other 4 are random and often unrelated. -- **Near-duplicates remain**: templated queries (Glaive, SGD) differing only in entities are kept; only exact duplicates were removed. -- **Weak descriptions**: MASSIVE descriptions are generated from label names (e.g., "User intention to play music"); Nemotron tool descriptions are kept as-is. -- **NL2Bash skew**: `find` dominates the terminal subset even after capping. -- **Decontamination is lexical**: paraphrased benchmark items would not be caught. - ---- - -## 🛠️ Quickstart - -```python -from datasets import load_dataset - -ds = load_dataset("TypeScript/FastBrain-Instruct") -print(ds["train"][0]) -``` - -| Field | Type | Description | -| :--- | :--- | :--- | -| `id` | string | Row id | -| `domain` | string | `code_mechanical` or `daily_assistant` | -| `category` | string | e.g. `api_tool_routing`, `agent_tool_selection`, `terminal_command`, `service_dialogue` | -| `query` | string | User request (or rendered recent conversation for agent rows) | -| `target` | string | Correct choice name | -| `choices` | list of `{name, description}` | Candidates, including `target` | -| `source` | string | Upstream source | - -## ⚖️ Upstream Attribution - -- **Google Research** — Schema-Guided Dialogue (Rastogi et al., AAAI 2020), CC BY-SA 4.0. -- **NVIDIA** — Nemotron-SFT-Agentic-v2, CC BY 4.0; decision-format transformation by **Maziyar Panahi** (AgentToolDecisions-180K). -- **Amazon Science** — MASSIVE (FitzGerald et al., 2022), CC BY 4.0. -- **Glaive AI** — Glaive Function Calling v2. -- **NousResearch** — Hermes Function Calling v1. -- **Lin et al.** — NL2Bash (LREC 2018). -- **burak29** — Git natural-language commands. - -## 📜 Citation - -```bibtex -@dataset{fastbrain_instruct_2026, - author = {TypeScript}, - title = {FastBrain-Instruct: Aggregated Fine-Tuning Corpus for System 1 Decision Models}, - year = {2026}, - publisher = {Hugging Face}, - howpublished = {https://huggingface.co/datasets/TypeScript/FastBrain-Instruct} -} -``` diff --git a/docs/v2-data-card.md b/docs/v2-data-card.md deleted file mode 100644 index f1479deb2edc3e35936278965d81af1a351be9b0..0000000000000000000000000000000000000000 --- a/docs/v2-data-card.md +++ /dev/null @@ -1,145 +0,0 @@ ---- -language: -- en -license: other -license_name: mixed-upstream-licenses -license_link: https://huggingface.co/datasets/TypeScript/FastBrain-Instruct/blob/main/LICENSE.md -task_categories: -- text-classification -- zero-shot-classification -task_ids: -- intent-classification -- multi-class-classification -tags: -- fastbrain -- tool-use -- function-calling -- intent-classification -- multi-label -- synthetic -- aggregated -source_datasets: -- glaiveai/glaive-function-calling-v2 -- NousResearch/hermes-function-calling-v1 -- nvidia/Nemotron-SFT-Agentic-v2 -- MaziyarPanahi/AgentToolDecisions-180K -- AnishJoshi/nl2bash-custom -- burak29/git-natural-language-commands -- AmazonScience/massive -- google-research-datasets/schema_guided_dstc8 -- WillHeld/mtop -- WillHeld/top_v2 -- prem-research/Funcdex-MT-Function-Calling -size_categories: -- 100K