VietPoet-Qwen3.5-9B-GGUF

VietPoet Qwen3.5-9B (GGUF)

A Qwen3.5-9B fine-tuned (QLoRA, 8,000 poems, 2 epochs) to write Vietnamese lục bát poems, exported for LM Studio / llama.cpp. It is meant to be used with the line-by-line sampler and rule checker from github.com/peterbuitho/ThoLucBat, whose local installation for Windows (VietPoet-win.zip) sets it up for you. Used on its own it writes the right shape but breaks the tone rules more often.

What it does and does not do

The larger sibling of VietPoet-Qwen3.5-4B-GGUF, trained the same way. In the author's tests (16-bit model on vLLM, 100 held-out "8 câu" prompts, line-by-line sampler with 16 samples per line) the 9B and the 4B are equivalent on the rules: rule score 0.994, 98% of poems fully valid. The raw 9B breaks the 6th/8th-syllable tone rule far less often than the 4B does (10% vs 38% of bát lines), so it needs the sampler less, but with the sampler the difference disappears. These GGUF files were not separately scored, and the 9B is slower and needs more memory than the 4B. Whether it writes better poems is not something these numbers can say.

These numbers measure form, not poetry. Poems are correct lục bát but the meaning is often loose or off-topic (training prompts only had the poem title as topic). Judge the poetry yourself.

Prompt format

Qwen chat format with thinking off, and the poem written by appending lines to the assistant turn:

<|im_start|>system
Bạn là nhà thơ Việt Nam chuyên sáng tác thơ lục bát.<|im_end|>
<|im_start|>user
Viết một bài thơ lục bát 8 câu về mùa thu quê em.<|im_end|>
<|im_start|>assistant
<think>

</think>

The request wording varies (several templates); they are in app/prompts.py in the GitHub repo.

Credits

• Training data: phamson02/vietnamese-poetry-corpus (CC BY 4.0), filtered to poems that pass a lục bát rule checker.
• Base model: Qwen/Qwen3.5-9B (Apache-2.0).
• Scoring idea: Vietnamese Poem Generation & the Prospect of Cross-Language Poem-to-Poem Translation (arXiv:2401.01078).