Copy of farbodtavakkoli/OTel-Embedding at 91662ce

.gitattributes

@@ -58,3 +58,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+rod_farbod_merged_with_qwen3_reranker_negs.jsonl filter=lfs diff=lfs merge=lfs -text
+OTel-Embedding-Data.jsonl filter=lfs diff=lfs merge=lfs -text

OTel-Embedding-Data.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77bdd4d5d78a9482cfdfe8704f23c72ab6e17d47a11f2e5d68db9a9302f7e81d
+size 950705678

README.md

@@ -0,0 +1,118 @@
+---
+license: apache-2.0
+task_categories:
+  - sentence-similarity
+  - feature-extraction
+language:
+  - en
+tags:
+  - telecommunications
+  - telecom
+  - 3gpp
+  - oran
+  - gsma
+  - rag
+  - embedding
+  - retrieval
+  - bi-encoder
+  - mnrl
+  - domain-specific
+  - otel
+size_categories:
+  - 100K<n<1M
+pretty_name: OTel-Embedding
+dataset_info:
+  features:
+    - name: anchor
+      dtype: string
+    - name: positive
+      dtype: string
+    - name: negative_1
+      dtype: string
+    - name: negative_2
+      dtype: string
+    - name: negative_3
+      dtype: string
+    - name: negative_4
+      dtype: string
+    - name: negative_5
+      dtype: string
+---
+
+# OTel-Embedding
+
+## Dataset Summary
+
+OTel-Embedding is a contrastive learning dataset for training bi-encoder retrieval models using Multiple Negatives Ranking Loss (MNRL). It is part of the Open Telco (OTel) AI project, the largest open-source AI initiative in telecommunications, curated by over 100 domain experts from industry and academia.
+
+Each sample consists of an anchor query, one positive passage, and five hard negative passages. The dataset is designed to train embedding models that can accurately retrieve relevant telecom documentation given a natural language query, forming the retrieval component of end-to-end RAG pipelines in telecommunications.
+
+## Dataset Structure
+
+Each sample contains the following fields:
+
+| Field | Type | Description |
+|---|---|---|
+| `anchor` | string | The query or information need used as the anchor in contrastive learning. |
+| `positive` | string | A passage that is relevant to the anchor query. |
+| `negative_1` | string | A hard negative passage that is topically related but does not answer the anchor query. |
+| `negative_2` | string | A hard negative passage. |
+| `negative_3` | string | A hard negative passage. |
+| `negative_4` | string | A hard negative passage. |
+| `negative_5` | string | A hard negative passage. |
+
+The five hard negatives per sample are selected to be semantically proximate to the anchor but factually non-relevant, forcing the model to learn fine-grained distinctions within the telecom domain.
+
+## Data Sources and Contributor Attribution
+
+The OTel corpus was curated by 100+ domain experts from industry and academia. Each source class was contributed by a specific institutional partner:
+
+| Source | Contributor | Raw samples |
+|---|---|---:|
+| arXiv telecom papers, 3GPP standards, telecom Wikipedia, telecom Common Crawl pages | Yale University | 681,172 |
+| GSMA Permanent Reference Documents (PRDs), Discover portal, mixed telecom documents | GSMA | 158,006 |
+| IETF RFC series | NetoAI | 100,751 |
+| Industry whitepapers | Khalifa University | 62,000 |
+| O-RAN specifications across working groups 1, 2, 4, 5, 6, 7, 8, 9, 10 | University of Leeds | 58,565 |
+| O-RAN documents across working groups | The University of Texas at Dallas | 42,000 |
+| **Total raw samples** | | **~1,102,494** |
+
+After the four-stage cleaning pipeline (described below), the corpus was reduced to **326,767** higher-confidence examples across the OTel dataset family.
+
+## Data Collection and Processing
+
+The raw corpus comprised approximately 1.1 million samples. A four-stage cleaning pipeline reduced this to 326K samples across the OTel dataset family:
+
+1. **Heuristic filtering**: Rule-based removal of malformed, duplicated, or low-quality entries.
+2. **Reranking filtering**: Relevance scoring with a cross-encoder reranker to discard weakly aligned query-passage pairs.
+3. **Embedding filtering**: Semantic quality assessment with a strong embedding model to remove outliers and near-duplicates in embedding space.
+4. **Deduplication**: Final deduplication pass to ensure sample uniqueness.
+
+Hard negatives were mined from the corpus using dense retrieval to identify passages that are close in embedding space to the positive but do not satisfy the information need of the anchor.
+
+## Intended Use
+
+OTel-Embedding is intended for training bi-encoder embedding models for telecom-domain retrieval. Target applications include:
+
+- Fine-tuning sentence transformer models for telecom document retrieval using MNRL or similar contrastive objectives.
+- Building the retrieval stage of RAG pipelines over telecom standards and specifications.
+- Benchmarking domain-adapted embedding models against general-purpose alternatives on telecom retrieval tasks.
+
+## Limitations
+
+- **Domain scope**: The dataset covers telecommunications standards and specifications. Embedding models trained on this data are optimized for telecom retrieval and may not generalize well to other domains.
+- **Language**: English only.
+- **Negative quality**: Hard negatives are mined automatically and may include occasional false negatives (passages that are in fact relevant to the anchor).
+- **Temporal coverage**: The source documents reflect standards and specifications available at the time of collection. Newer revisions may not be represented.
+
+## Citation
+
+```bibtex
+@misc{otel_data_2026,
+  title  = {{OTel}: Open Telco AI Datasets, Benchmarks, and Models},
+  author = {Tavakkoli, Farbod and others},
+  year   = {2026},
+  note   = {Open Telco (OTel) dataset release},
+  url    = {https://huggingface.co/farbodtavakkoli}
+}
+```

croissant.json

@@ -0,0 +1,252 @@
+{
+  "@context": {
+    "@language": "en",
+    "@vocab": "https://schema.org/",
+    "arrayShape": "cr:arrayShape",
+    "citeAs": "cr:citeAs",
+    "column": "cr:column",
+    "conformsTo": "dct:conformsTo",
+    "containedIn": "cr:containedIn",
+    "cr": "http://mlcommons.org/croissant/",
+    "data": {
+      "@id": "cr:data",
+      "@type": "@json"
+    },
+    "dataBiases": "rai:dataBiases",
+    "dataCollection": "cr:dataCollection",
+    "dataType": {
+      "@id": "cr:dataType",
+      "@type": "@vocab"
+    },
+    "dct": "http://purl.org/dc/terms/",
+    "extract": "cr:extract",
+    "field": "cr:field",
+    "fileProperty": "cr:fileProperty",
+    "fileObject": "cr:fileObject",
+    "fileSet": "cr:fileSet",
+    "format": "cr:format",
+    "includes": "cr:includes",
+    "isArray": "cr:isArray",
+    "isLiveDataset": "cr:isLiveDataset",
+    "jsonPath": "cr:jsonPath",
+    "key": "cr:key",
+    "md5": "cr:md5",
+    "parentField": "cr:parentField",
+    "path": "cr:path",
+    "personalSensitiveInformation": "rai:personalSensitiveInformation",
+    "recordSet": "cr:recordSet",
+    "references": "cr:references",
+    "regex": "cr:regex",
+    "repeated": "cr:repeated",
+    "replace": "cr:replace",
+    "sc": "https://schema.org/",
+    "separator": "cr:separator",
+    "source": "cr:source",
+    "subField": "cr:subField",
+    "transform": "cr:transform",
+    "rai": "http://mlcommons.org/croissant/RAI/",
+    "dataLimitations": "rai:dataLimitations",
+    "dataUseCases": "rai:dataUseCases",
+    "dataSocialImpact": "rai:dataSocialImpact",
+    "hasSyntheticData": "rai:hasSyntheticData",
+    "prov": "http://www.w3.org/ns/prov#",
+    "wasGeneratedBy": "prov:wasGeneratedBy"
+  },
+  "@type": "sc:Dataset",
+  "name": "OTel-Embedding",
+  "description": "OTel-Embedding is a contrastive learning dataset for training bi-encoder retrieval models using Multiple Negatives Ranking Loss (MNRL) in the telecommunications domain.",
+  "alternateName": [
+    "farbodtavakkoli/OTel-Embedding",
+    "OTel-Embedding"
+  ],
+  "conformsTo": "http://mlcommons.org/croissant/1.1",
+  "creator": {
+    "@type": "Person",
+    "name": "Farbod Tavakkoli",
+    "url": "https://huggingface.co/farbodtavakkoli"
+  },
+  "keywords": [
+    "sentence-similarity",
+    "feature-extraction",
+    "English",
+    "apache-2.0",
+    "100K - 1M",
+    "telecommunications",
+    "telecom",
+    "3gpp",
+    "oran",
+    "gsma",
+    "embedding",
+    "retrieval",
+    "bi-encoder",
+    "mnrl",
+    "otel",
+    "Croissant"
+  ],
+  "license": "https://choosealicense.com/licenses/apache-2.0/",
+  "url": "https://huggingface.co/datasets/farbodtavakkoli/OTel-Embedding",
+  "distribution": [
+    {
+      "@type": "cr:FileObject",
+      "@id": "repo",
+      "name": "repo",
+      "description": "The Hugging Face git repository.",
+      "contentUrl": "https://huggingface.co/datasets/farbodtavakkoli/OTel-Embedding/tree/refs%2Fconvert%2Fparquet",
+      "encodingFormat": "git+https",
+      "sha256": "https://github.com/mlcommons/croissant/issues/80"
+    },
+    {
+      "@type": "cr:FileSet",
+      "@id": "parquet-files-for-config-default",
+      "containedIn": {
+        "@id": "repo"
+      },
+      "encodingFormat": "application/x-parquet",
+      "includes": "default/*/*.parquet"
+    }
+  ],
+  "recordSet": [
+    {
+      "@type": "cr:RecordSet",
+      "@id": "default",
+      "name": "default",
+      "field": [
+        {
+          "@type": "cr:Field",
+          "@id": "default/split",
+          "name": "split",
+          "description": "The dataset split (e.g., train, test, validation).",
+          "dataType": "sc:Text",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-default"
+            },
+            "extract": {
+              "fileProperty": "fullpath"
+            },
+            "transform": {
+              "regex": "default/(?P<split>[^/]*)/.*"
+            }
+          }
+        },
+        {
+          "@type": "cr:Field",
+          "@id": "default/anchor",
+          "name": "anchor",
+          "description": "The anchor text (query).",
+          "dataType": "sc:Text",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-default"
+            },
+            "extract": {
+              "column": "anchor"
+            }
+          }
+        },
+        {
+          "@type": "cr:Field",
+          "@id": "default/positive",
+          "name": "positive",
+          "description": "The positive passage.",
+          "dataType": "sc:Text",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-default"
+            },
+            "extract": {
+              "column": "positive"
+            }
+          }
+        },
+        {
+          "@type": "cr:Field",
+          "@id": "default/negative_1",
+          "name": "negative_1",
+          "description": "Hard negative passage 1.",
+          "dataType": "sc:Text",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-default"
+            },
+            "extract": {
+              "column": "negative_1"
+            }
+          }
+        },
+        {
+          "@type": "cr:Field",
+          "@id": "default/negative_2",
+          "name": "negative_2",
+          "description": "Hard negative passage 2.",
+          "dataType": "sc:Text",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-default"
+            },
+            "extract": {
+              "column": "negative_2"
+            }
+          }
+        },
+        {
+          "@type": "cr:Field",
+          "@id": "default/negative_3",
+          "name": "negative_3",
+          "description": "Hard negative passage 3.",
+          "dataType": "sc:Text",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-default"
+            },
+            "extract": {
+              "column": "negative_3"
+            }
+          }
+        },
+        {
+          "@type": "cr:Field",
+          "@id": "default/negative_4",
+          "name": "negative_4",
+          "description": "Hard negative passage 4.",
+          "dataType": "sc:Text",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-default"
+            },
+            "extract": {
+              "column": "negative_4"
+            }
+          }
+        },
+        {
+          "@type": "cr:Field",
+          "@id": "default/negative_5",
+          "name": "negative_5",
+          "description": "Hard negative passage 5.",
+          "dataType": "sc:Text",
+          "source": {
+            "fileSet": {
+              "@id": "parquet-files-for-config-default"
+            },
+            "extract": {
+              "column": "negative_5"
+            }
+          }
+        }
+      ]
+    }
+  ],
+  "rai:dataLimitations": "The dataset covers telecommunications standards and specifications only. It is English-only and reflects standards available at the time of collection. Performance of models trained on this data should not be generalized beyond the telecom domain.",
+  "rai:dataBiases": "The dataset may overrepresent certain telecom subdomains (e.g., 3GPP, O-RAN) relative to others due to source availability. Widely documented standards topics are better represented than niche or emerging telecom areas.",
+  "rai:personalSensitiveInformation": "The dataset does not contain personal or sensitive information. All source material is drawn from publicly available technical standards, specifications, and academic publications.",
+  "rai:dataUseCases": "Training, Fine Tuning",
+  "rai:dataSocialImpact": "The dataset supports open and reproducible telecom AI research, reducing barriers to domain-specific model development. The primary risk is potential misuse of domain-specialized models to generate misleading telecom documentation.",
+  "rai:hasSyntheticData": "The training data includes synthetically generated question-answer pairs derived from source documents through automated pipelines, including LLM-assisted answer recreation and fact-grounded verification.",
+  "rai:dataCollection": "Source datasets and collection methodology: the OTel corpus is derived from publicly available telecom sources, including 3GPP technical specifications, GSMA Permanent Reference Documents (PRDs) and Discover portal materials, O-RAN Alliance specifications across working groups 1, 2, 4, 5, 6, 7, 8, 9, and 10, IETF Request for Comments (RFC) series, telecom academic papers from arXiv, telecom Wikipedia articles, telecom-related Common Crawl pages, and industry whitepapers. The OTel datasets release only derived question-answer pairs and contrastive examples, not the raw source documents. Data was collected by 100+ domain experts across industry and academia and processed through a six-phase enrichment and four-stage cleaning pipeline.",
+  "prov:wasGeneratedBy": {
+    "@type": "prov:Activity",
+    "name": "OTel Data Curation Pipeline",
+    "description": "A six-phase enrichment and four-stage cleaning pipeline curated by 100+ domain experts from AT&T, AMD, GSMA, Microsoft, RelationalAI, MLCommons, Mantis NLP, Essential AI, Yale University, Khalifa University, University of Leeds, The University of Texas at Dallas, Purdue University, and NetoAI. The raw corpus of approximately 1.1M samples was cleaned to approximately 326K samples through heuristic filtering, cross-encoder reranking filtering, embedding-based semantic filtering, and deduplication."
+  }
+}