README.md
3.9 KB · 101 lines · markdown Raw
1 ---
2 license: apache-2.0
3 language:
4 - en
5 base_model:
6 - EssentialAI/rnj-1-instruct
7 tags:
8 - telecom
9 - telecommunications
10 - gsma
11 - fine-tuned
12 pipeline_tag: text-generation
13 ---
14
15 # OTel-LLM-8.3B-IT
16
17 **OTel-LLM-8.3B-IT** is a telecom-specialized language model fine-tuned on telecommunications domain data. It is part of the [OTel Family of Models](https://huggingface.co/collections/farbodtavakkoli/otel-llm), an open-source initiative to build industry-standard AI models for the global telecommunications sector.
18
19 ## Model Details
20
21 | Attribute | Value |
22 |-----------|-------|
23 | **Base Model** | [EssentialAI/rnj-1-instruct](https://huggingface.co/EssentialAI/rnj-1-instruct) |
24 | **Parameters** | 8.3B |
25 | **Training Method** | Full parameter fine-tuning |
26 | **Language** | English |
27 | **License** | Apache 2.0 |
28
29 ## Training Data
30
31 The model was trained on telecom-focused data curated by 100+ domain experts. Each source class was contributed by a specific institutional partner:
32
33 | Source | Contributor |
34 |---|---|
35 | arXiv telecom papers, 3GPP standards, telecom Wikipedia, telecom Common Crawl | Yale University |
36 | GSMA Permanent Reference Documents, Discover portal | GSMA |
37 | IETF RFC series | NetoAI |
38 | Industry whitepapers | Khalifa University |
39 | O-RAN specifications (working groups 1, 2, 4, 5, 6, 7, 8, 9, 10) | University of Leeds |
40 | O-RAN documents across working groups | The University of Texas at Dallas |
41
42 Released datasets: [OTel-LLM](https://huggingface.co/datasets/farbodtavakkoli/OTel-LLM), [OTel-Embedding](https://huggingface.co/datasets/farbodtavakkoli/OTel-Embedding), [OTel-Reranker](https://huggingface.co/datasets/farbodtavakkoli/OTel-Reranker), [OTel-Safety](https://huggingface.co/datasets/farbodtavakkoli/OTel-Safety).
43
44 ## Intended Use
45
46 The OTel model family is designed to power end-to-end Retrieval-Augmented Generation (RAG) pipelines for telecommunications. The three model types serve complementary roles:
47
48 1. **Embedding** — Retrieve relevant chunks from telecom specifications, standards, and documentation.
49 2. **Reranker** — Re-score and prioritize the retrieved chunks for relevance.
50 3. **LLM** — Generate accurate responses grounded in the retrieved context.
51
52 Users can deploy the full pipeline or use individual models independently based on their needs.
53
54 **Note:** The LLMs include abstention training — if the model does not receive sufficient context, it will decline to answer rather than hallucinate. This means the models are optimized for context-grounded generation, not open-ended question answering.
55
56 ## Related Models
57
58 ### Language Models
59 - [OTel LLM Collection](https://huggingface.co/collections/farbodtavakkoli/otel-llm)
60
61 ### Embedding Models
62 - [OTel Embedding Collection](https://huggingface.co/collections/farbodtavakkoli/otel-embedding)
63
64 ### Reranker Models
65 - [OTel Reranker Collection](https://huggingface.co/collections/farbodtavakkoli/otel-reranker)
66
67 ## Related Datasets
68
69 - [OTel-Embedding](https://huggingface.co/datasets/farbodtavakkoli/OTel-Embedding)
70 - [OTel-Safety](https://huggingface.co/datasets/farbodtavakkoli/OTel-Safety)
71 - [OTel-LLM](https://huggingface.co/datasets/farbodtavakkoli/OTel-LLM)
72 - [OTel-Reranker](https://huggingface.co/datasets/farbodtavakkoli/OTel-Reranker)
73
74 ## Training Infrastructure
75
76 - **Framework**: ScalarLM (GPU-agnostic)
77 - **Compute**: AMD and NVIDIA GPUs.
78
79 ## Project Resources
80
81 - **Project page:** https://huggingface.co/farbodtavakkoli
82 - **Code:** https://github.com/farbodtavakkoli/OTel
83 - **Media coverage list:** https://github.com/farbodtavakkoli/OTel/blob/main/docs/media_coverage.md
84
85 ## Citation
86
87
88 ```bibtex
89 @misc{otel_models_2026,
90 title = {OTel: Open Telco AI Datasets, Benchmarks, and Models},
91 author = {Tavakkoli, Farbod and others},
92 year = {2026},
93 note = {Open Telco (OTel) model release},
94 url = {https://huggingface.co/farbodtavakkoli}
95 }
96 ```
97
98 ## Contact
99
100 If you have any technical questions, please feel free to reach out to farbod.tavakkoli@att.com or farbodtavakoli@gmail.com
101