Modelpublic

fivetech/Harbour

sourceHugging Faceapache-2.0updated 3mo agoView on Hugging Face

0likes40downloads

Model Card

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, 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 project — an open-source Clipper-compatible compiler — and FiveWin (FWH) GUI framework.

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