telecom-oss-8b-merged/README.md

---
base_model: AliMaatouk/LLama-3-8B-Tele
license: llama3
language:
- en
tags:
- telecom
- oss
- bss
- tmf
- tmforum
- etom
- sid
- llama-3
- merged
pipeline_tag: text-generation
library_name: transformers
---

# Telecom OSS/BSS Domain LLM (Merged Standalone)

**Built with Meta Llama 3.**

A standalone 8B model merging the [`Tapask/telecom-oss-8b`](https://huggingface.co/Tapask/telecom-oss-8b) LoRA adapter into its base [`AliMaatouk/LLama-3-8B-Tele`](https://huggingface.co/AliMaatouk/LLama-3-8B-Tele). Specialised for **TMF Frameworx** (eTOM, SID, Open APIs) and OSS/BSS telecom operations. No PEFT runtime required — load and use like any Llama-3 model.

Two flavours of the same fine-tune:
- **Standalone (this repo)** — single load, simpler for inference
- **[Adapter-only](https://huggingface.co/Tapask/telecom-oss-8b)** — 670 MB, needs the base model at load time (smaller download)

## Model summary

| | |
|---|---|
| **Architecture** | Llama-3 8B (transformers-native, fp16 safetensors) |
| **Origin** | `AliMaatouk/LLama-3-8B-Tele` + QLoRA fine-tune (r=64, α=128, dropout=0.05) |
| **Fine-tune target modules** | `q_proj`, `k_proj`, `v_proj`, `o_proj`, `gate_proj`, `up_proj`, `down_proj` |
| **Training data** | 18,779 synthetic instruction–response pairs across 8 TMF-aligned categories |
| **Training config** | 3 epochs · effective batch 16 · seq 4096 · cosine LR (peak 2e-4) · bf16 · gradient checkpointing |
| **Training hardware** | NVIDIA A100 SXM4 80GB · ~8.3 h wall time |
| **Best eval loss** | **0.8438** (epoch 2.56) — `load_best_model_at_end=True` |
| **Sharded safetensors** | 5 × ~3-4 GB files (~16.1 GB total) |

## Intended use

Domain-specialised completions for:

- **TMF Open API** payload generation (TMF620–TMF700 suite)
- **eTOM** process decomposition (Fulfillment / Assurance / Billing end-to-end flows)
- **SID** entity relationship reasoning (ProductOffering → Service → Resource hierarchies, Party/Role patterns, characteristic specifications)
- **Inventory reconciliation** (discovery–inventory mismatches, phantom/orphan resources)
- **OSS/BSS architecture** decisions (ODA components, build-vs-buy, MANO choices)
- **Fault-to-inventory correlation** (service impact from topology traversal)
- **TMF spec Q&A** (technical knowledge retrieval)
- **Integration code** (TMF-compliant Python clients)

### How to use

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Tapask/telecom-oss-8b-merged"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype="auto", device_map="auto")
model.eval()

prompt = """Below is an instruction that describes a task related to telecom OSS/BSS systems, TMF Frameworx, or network operations. Write a response that appropriately completes the request.

### Instruction:
Generate a TMF641 service order payload for a 5G network slice with URLLC characteristics targeting an enterprise IoT customer.

### Response:
"""

inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
output = model.generate(**inputs, max_new_tokens=1024, temperature=0.3, do_sample=True)
print(tokenizer.decode(output[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))
```

Uses the **Alpaca prompt template** the model was trained with. Keep the `### Instruction: / ### Response:` markers exactly.

### Deploying with Ollama / llama.cpp

This repo is fp16 safetensors. For Ollama/llama.cpp, convert to GGUF:

```bash
git clone https://github.com/ggerganov/llama.cpp && cd llama.cpp
pip install -r requirements/requirements-convert_hf_to_gguf.txt
python convert_hf_to_gguf.py /path/to/downloaded/telecom-oss-8b-merged \
    --outfile telecom-oss-8b.f16.gguf --outtype f16
./llama-quantize telecom-oss-8b.f16.gguf telecom-oss-8b.Q4_K_M.gguf Q4_K_M
```

Then create an Ollama Modelfile with the Llama-3 chat template and `FROM ./telecom-oss-8b.Q4_K_M.gguf`.

## Training data

18,779 instruction–response pairs generated programmatically via [Claude API](https://www.anthropic.com/), [Kimi K2.5 on Ollama Cloud](https://ollama.com/), and [GLM-5 on Ollama Cloud](https://ollama.com/), prompted with 8 category-specific TMF expert personas (system prompts + 4–5 batch variants each). Distribution:

| # | Category | Pairs | Primary model |
|---|---|---:|---|
| 1 | TMF Open API Payloads | 2,962 | GLM-5 |
| 2 | eTOM Process Decomposition | 1,967 | GLM-5 |
| 3 | SID Entity Reasoning | 1,963 | Kimi K2.5 |
| 4 | Inventory Reconciliation | 2,962 | Kimi K2.5 |
| 5 | OSS/BSS Architecture | 1,893 | Kimi K2.5 |
| 6 | Fault-to-Inventory Correlation | 1,929 | GLM-5 |
| 7 | TMF Spec Q&A | 2,875 | Kimi K2.5 (after GLM-5 hit 54% dedup rate) |
| 8 | TMF Integration Code Generation | 2,228 | GLM-5 |

Splits (seed 42): **16,901 train / 939 val / 939 test.**

Quality passes applied:
- MD5-hash deduplication on `instruction` field
- Category-aware soft validators (TMF API reference presence, SID entity coverage, eTOM term coverage, JSON validity for payload categories)
- Refusal-pattern scrubbing (`I cannot`, `As an AI`, etc. removed)
- Type coercion for 297 pairs where source models emitted `output` as nested JSON objects instead of JSON strings

## Evaluation loss trajectory

| Epoch | Eval loss |
|---|---|
| 2.27 | 0.8545 |
| 2.37 | 0.8440 |
| 2.46 | 0.8447 |
| **2.56** | **0.8438** ← best, used for merge |
| 2.65 | 0.8479 |
| 2.75 | 0.8478 |

Loss plateaued and began ticking up after epoch 2.56 — classic mild overfitting signal. `load_best_model_at_end=True` ensured the merged model corresponds to the epoch 2.56 region.

## Limitations

- **Synthetic data provenance** — training pairs were generated by LLMs (Claude, Kimi K2.5, GLM-5) prompted with TMF expert personas. Content is stylistically consistent with TMF specs but **not validated line-by-line against official TMF Open API documents**. Treat outputs as starting points, not canonical.
- **Inner-JSON flaws** — ~15% of category-1 pairs had minor inner-JSON issues (unescaped quotes inside payload strings). Not filtered out for training.
- **Category 8 undertrained** — TMF Code Generation category ended at 74% of its 3,000-pair target due to narrow topic space and dedup loss. Code-generation quality is the weakest axis.
- **Domain scope** — the model is narrow. General-purpose conversation, math, or code outside TMF integration will be no better (and often worse) than the base.
- **Standards currency** — trained against TMF Open API versions current as of the prompt design (~v4/v5 dominant). May cite outdated endpoint paths for newer TMF releases.

## Intended use — restrictions

Follows the [Llama 3 Community License](https://llama.meta.com/llama3/license/) and [Acceptable Use Policy](https://llama.meta.com/llama3/use-policy/). Intended for:

- Domain research, prototyping, and educational use
- Assistant-style answers to TMF/OSS/BSS engineering questions
- Starter payload generation (to be reviewed before use in production)

Not suitable for:
- Generating production systems config without human review
- Compliance-sensitive deployments (TMF spec accuracy is not guaranteed)
- Any of the prohibited uses in the Llama 3 AUP

## License

- Model weights: inherit **Llama 3 Community License** from the base model `meta-llama/Meta-Llama-3-8B`
- "Built with Meta Llama 3" attribution required (see top of this card)
- Note that Llama 3 license restricts some commercial uses (700M+ MAU clause) and prohibited use cases — consult the license before redistribution

## Acknowledgements

- **Meta AI** — Llama 3 base model
- **Ali Maatouk** — telecom-pretrained continuation [`AliMaatouk/LLama-3-8B-Tele`](https://huggingface.co/AliMaatouk/LLama-3-8B-Tele)
- **Anthropic, Moonshot AI, Zhipu AI** — Claude, Kimi K2.5, GLM-5 (used to generate synthetic training data)
- **TMForum** — the eTOM, SID, and Open API standards this model targets

## Citation

```
@misc{tapask_telecom_oss_8b_merged_2026,
  title        = {Telecom OSS/BSS Domain LLM (Merged, based on LLama-3-8B-Tele)},
  author       = {Tapas},
  year         = {2026},
  howpublished = {\url{https://huggingface.co/Tapask/telecom-oss-8b-merged}},
}
```