---
license: apache-2.0
base_model: Qwen/Qwen1.5-0.5B
library_name: peft
language:
  - en
tags:
  - lora
  - peft
  - qwen
  - edge-ai
  - edge-impulse
  - documentation
  - code-generation
  - conversational
pipeline_tag: text-generation
widget:
  - text: "How can I deploy an Edge Impulse model to an Arduino device?"
    example_title: "Edge AI deployment"
space: eoinedge/edgeai-docs-embedding-qwen1.5-0.5b-instruct-space
---

# edgeai-docs-embedding-qwen1.5-0.5b-instruct

A lightweight LoRA adapter fine-tuned on **1,794 Edge Impulse / Edge AI MDX documentation files** from the [Edge Impulse documentation](https://docs.edgeimpulse.com), built on top of [`Qwen/Qwen1.5-0.5B`](https://huggingface.co/Qwen/Qwen1.5-0.5B).

Optimized for:
- answering developer questions about Edge Impulse Studio, SDKs, APIs, and tooling
- summarizing technical documentation and tutorials
- generating code snippets for edge ML workflows
- lightweight local/edge deployment with PEFT adapters

> **Larger variants in training:** [1.5B](https://huggingface.co/eoinedge/edgeai-qwen2.5coder-1.5b-lora) · [7B](https://huggingface.co/eoinedge/edgeai-qwen2.5coder-7b-lora) (Qwen2.5-Coder base)

---

## Model Summary

`edgeai-docs-embedding-qwen1.5-0.5b-instruct` is a PEFT LoRA adapter trained for documentation-focused text generation and conversational support over Edge Impulse / Edge AI knowledge.

### Use cases
- Documentation Q&A for Edge Impulse developers
- Technical explanation of Studio workflows, SDK usage, and hardware deployment
- Generating sample code for API, CLI, and Python SDK integrations
- Retrieval-augmented generation (RAG) over Edge AI docs

---

## Model Details

| Property | Value |
|---|---|
| Base model | `Qwen/Qwen1.5-0.5B` |
| Adapter type | LoRA (PEFT) |
| LoRA rank (`r`) | 8 |
| LoRA alpha | 32 |
| Target modules | `q_proj`, `v_proj` |
| Task type | CAUSAL_LM |
| Trainable parameters | ~786K (0.17% of base) |
| Training epochs | 3 |
| Batch size | 4 (× grad accum 2 = effective 8) |
| Learning rate | 3e-4 |
| Max sequence length | 512 tokens |
| Training hardware | Apple M1 Pro (MPS, fp16) |
| Precision | float16 |

---

## Training Data

| Stat | Value |
|---|---|
| Source | [Edge Impulse documentation](https://docs.edgeimpulse.com) |
| File format | MDX (Markdown + JSX components) |
| Total files | 1,794 `.mdx` files |
| Preprocessing | Stripped frontmatter, imports, JSX tags; unwrapped code fences; flattened links |
| Chunk size | 512 tokens |

Topics covered: Studio projects, datasets, data ingestion, DSP and transformation blocks, learning and processing blocks, model deployment, Python SDK, REST API, CLI tools, and edge inference.

---

## Evaluation

### QA evaluation
- Dataset: 5 fixed developer-style prompts
- Base avg keyword count: **8.2**
- Adapter avg keyword count: **6.8**
- Code snippet presence: **5/5** for both base and adapter

### Perplexity on Edge AI samples
- Test corpus: 30 sample Edge AI documentation files
- Base mean perplexity: **11.53**
- Adapter mean perplexity: **12.02**
- Adapter wins: **4 / 30 documents**

> These metrics are from small validation samples and should be interpreted as a lightweight benchmark rather than a full production evaluation.

---

## Tutorials

- [Offline SLMs for Edge AI Development — Part 1: Qwen LoRA Adapter Fine-Tuned on Edge Impulse Docs](https://docs.edgeimpulse.com/projects/expert-network/integrating-slms-on-linux)
- [Offline SLMs for Edge AI Development — Part 2: RAG as an Enhancement for Fine-Tuned Models with FAISS](https://docs.edgeimpulse.com/projects/expert-network/rag-docs-assistant-faiss-qwen)
- [Offline SLMs for Edge AI Development — Part 3: Agentic Coding with an Arduino Fine-Tuned Adapter via llama.cpp and OpenCode](https://docs.edgeimpulse.com/projects/expert-network/opencode-offline-coding-assistant)

---

## Usage

### Load with PEFT

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

BASE_MODEL = "Qwen/Qwen1.5-0.5B"
ADAPTER = "eoinedge/edgeai-docs-embedding-qwen1.5-0.5b-instruct"

device = "cuda" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu")

tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)
base_model = AutoModelForCausalLM.from_pretrained(
    BASE_MODEL,
    torch_dtype=torch.float16 if device != "cpu" else torch.float32,
    device_map=device,
)
model = PeftModel.from_pretrained(base_model, ADAPTER)
model.eval()
```

### Text generation pipeline

```python
from transformers import pipeline

pipe = pipeline("text-generation", model="eoinedge/edgeai-docs-embedding-qwen1.5-0.5b-instruct")
print(pipe([{"role": "user", "content": "How do I use the Edge Impulse Python SDK to upload data?"}]))
```

---

## Example prompts

| Task | Prompt |
|---|---|
| Concept explanation | `What is a DSP block in Edge Impulse?` |
| API usage | `How do I use the Edge Impulse Python SDK to upload data?` |
| Deployment | `How do I deploy a model to an Arduino Nano 33 BLE Sense?` |
| Code generation | `Write Python code to collect IMU data and upload it to Edge Impulse.` |
| Troubleshooting | `Why is my Edge Impulse model showing high latency?` |

---

## Limitations

- Based on a 0.5B base model — may struggle with long multi-step reasoning
- Training data covers Edge Impulse docs as of mid-2026; newer features may be missing
- May hallucinate or fabricate undocumented APIs or block behavior
- Not validated for safety-critical or production use
- Validate generated code before deploying on hardware

---

## Related models

| Model | Base | Status |
|---|---|---|
| This model | Qwen/Qwen1.5-0.5B | ✅ Available |
| [eoinedge/edgeai-qwen2.5coder-1.5b-lora](https://huggingface.co/eoinedge/edgeai-qwen2.5coder-1.5b-lora) | Qwen2.5-Coder-1.5B-Instruct | 🔄 Training |
| [eoinedge/edgeai-qwen2.5coder-7b-lora](https://huggingface.co/eoinedge/edgeai-qwen2.5coder-7b-lora) | Qwen2.5-Coder-7B-Instruct | 🔄 Training |
| [eoinedge/arduino-qwen0.5-lora](https://huggingface.co/eoinedge/arduino-qwen0.5-lora) | Qwen/Qwen1.5-0.5B | ✅ Available |

---

## Citation

```bibtex
@misc{edgeai-docs-embedding-qwen1.5-0.5b-instruct,
  author       = {Jordan, Eoin},
  title        = {edgeai-docs-embedding-qwen1.5-0.5b-instruct},
  year         = {2026},
  publisher    = {Hugging Face},
  howpublished = {\url{https://huggingface.co/eoinedge/edgeai-docs-embedding-qwen1.5-0.5b-instruct}}
}
```