README.md

---
base_model: Qwen/Qwen3.5-35B-A3B
library_name: peft
pipeline_tag: text-generation
tags:
- harbour
- fivewin
- fwh
- lora
- sft
- transformers
- trl
- unsloth
- code-generation
- xbase
- clipper
language:
- en
- es
license: apache-2.0
---

# Harbour/FWH Coder — Qwen3.5-35B-A3B LoRA v2

> **Training code:** https://github.com/FiveTechSoft/finetune

LoRA adapter fine-tuned on **5,004 compilable Harbour and FiveWin (FWH) examples** for code generation. Built on top of [Qwen3.5-35B-A3B](https://huggingface.co/Qwen/Qwen3.5-35B-A3B), a 35B Mixture-of-Experts model with 256 experts (8 active per token).

## What's New in v2

| | v1 | **v2** |
|---|---|---|
| **Dataset** | 996 entries (7 categories) | **5,004 entries** (8 categories) |
| **Training time** | 3h 43min | **~8h** |
| **Eval loss** | 0.5957 | **0.4790** (↓20%) |
| **Train loss** | 0.6456 | **0.4211** (↓35%) |
| **Format** | messages (chat) | instruction/output |
| **Learning rate** | 1e-4 | 8e-5 (conservative) |
| **Epochs** | 3 | 2 |

### Key improvements

- **5x more training data** — expanded from 996 to 5,004 unique, compilable examples
- **Better loss convergence** — 35% lower train loss, 20% lower eval loss
- **More conservative training** — lower learning rate preserves base model capabilities
- **FiveWin (FWH) coverage** — added FiveWin GUI framework examples
- **Verified code** — all examples verified with Harbour v3.2.0dev compiler

## Dataset

Training data sourced from the [Harbour](https://harbour.github.io/) project — an open-source Clipper-compatible compiler — and [FiveWin](https://fivewin.com/) (FWH) GUI framework.

### Categories

| Category | Count | Description |
|---|---|---|
| contrib | 583 | Contribution libraries (network, database, graphics, security...) |
| rtl | 80 | Harbour Runtime Library |
| include | 59 | Header files with constants/macros |
| tests | 225 | Test programs |
| extras | 25 | Extra libraries |
| utils | 13 | Utility programs |
| fwh | ~500+ | FiveWin GUI framework examples |
| low-level C | 500+ | HB_FUNC C extension wrappers |

### Format

```json
{
  "instruction": "Write a Harbour function that creates a 2D array...",
  "input": "",
  "system": "You are an expert Harbour programmer...",
  "output": "FUNCTION CreateTable()\n   LOCAL aTable := {}\n   ...",
  "task_type": "code_generation"
}
```

## Training Details

### Hardware

- **Device:** NVIDIA GB10 Grace Blackwell Superchip (DGX Spark)
- **Architecture:** ARM aarch64 (10 NVIDIA Grace CPU cores + Blackwell GPU)
- **RAM:** 121 GB unified memory (CPU + GPU shared)
- **OS:** Ubuntu 24.04.4 LTS (aarch64)
- **Training time:** 7h 49min (564 steps, 2 epochs over 5,004 samples)

### Hyperparameters

| Parameter | Value |
|---|---|
| Base model | Qwen3.5-35B-A3B (MoE, 256 experts) |
| Method | QLoRA (4-bit) |
| LoRA rank | 8 |
| LoRA alpha | 16 |
| LoRA targets | q/k/v/o/gate/up/down_proj |
| Epochs | 2 |
| Learning rate | 8e-5 |
| LR scheduler | cosine |
| Warmup ratio | 0.05 |
| Batch size | 1 (effective: 16 via grad accum) |
| Max seq length | 1024 |
| Optimizer | adamw_8bit |

### Framework

- Unsloth 2026.6.8
- Transformers 5.5.0
- PEFT 0.19.1
- PyTorch 2.12.1

## How to Use

### With PEFT + Transformers

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

base_model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen3.5-35B-A3B",
    load_in_4bit=True,
    device_map="auto",
)
model = PeftModel.from_pretrained(base_model, "fivetech/Harbour")
tokenizer = AutoTokenizer.from_pretrained("fivetech/Harbour")

prompt = "Write a Harbour function that splits a CSV string into an array."
messages = [
    {"role": "system", "content": "You are an expert Harbour programmer. Write compilable code."},
    {"role": "user", "content": prompt},
]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer([text], return_tensors="pt").to(model.device)

output = model.generate(**inputs, max_new_tokens=1500, temperature=0.2)
print(tokenizer.decode(output[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))
```

### With Ollama (merge + quantize first)

```bash
# Export to GGUF
python -m unsloth.save_pretrained_gguf model_output/ ./tokenizer/ q4_k_m

# Then use with Ollama
ollama create harbour-coder -f Modelfile
```

## Evaluation

Evaluated on 100 Harbour programming tests (Arrays, OOP, Functions, Database, File I/O, Control flow):

- **Compilation pass rate:** TBD (running test battery)
- **Categories tested:** Arrays (48), OOP (22), Other (9), Functions (8), Database (7), File I/O (4), Control (2)

## License

Apache 2.0

## Model Card Contact

fivetech — https://github.com/fivetechsoft/finetune