cd /news/large-language-models/2-184-to-170-184-one-epoch-on-two-r9… · home topics large-language-models article
[ARTICLE · art-127634] src=forum.level1techs.com ↗ pub= topic=large-language-models verified=true sentiment=↑ positive

2/184 to 170/184: one epoch on two R9700s took a 14B coder from useless to 5 points behind gpt-5.6-sol

A fine-tuned Qwen2.5-Coder-14B-Instruct model compiled 170 of 184 MQL5 test prompts after one epoch of training on the author's dataset, up from 2 of 184 for the untouched base model, according to a published benchmark card. The same 184 prompts scored 179 of 184 for gpt-5.6-sol, leaving the local 14B about 5 points behind the frontier API model. Training ran under ROCm on three AMD Radeon AI PRO R9700 cards across two machines totaling 96GB of VRAM, and the release includes the 184 prompts, scoring rules, per-item results, and a verifier script, but not the training data or tuned weights.

read2 min views6 publishedSep 12, 2026

Been a while. Last time I posted the dataset had passed 300k

entries and I had one 14B fine tune with 81% structural

correctness. Since then I did the thing I should have done from

the start, I built a proper benchmark and published it so people

can check my numbers instead of taking my word for it.

What got published:

184 test prompts for MQL5 (the language for algo trading bots),

the scoring rules, per item results for every model, and a

verifier script that checks the release hashes and recomputes the

headline numbers from the rows. A pass is zero compiler errors

and a built binary. Thats it. Compiling is a low bar and the card

says so, it does not mean the bot trades well.

The numbers:

Base Qwen2.5-Coder-14B-Instruct, untouched: 2 of 184 compiled. Same model after one epoch on my data: 170 of 184.

gpt-5.6-sol on the same 184 prompts: 179 of 184. So the local 14B went from useless to about 5 points behind a

frontier API model on my niche. One epoch. I am not going to

pretend thats the finish line but it is the first number I have

that somebody else can reproduce, and that matters more to me

than the number itself.

Hardware, since this is L1T:

Everything is still AMD and still local, but the layout changed

since April. The main rig is now a 7950X3D with 64GB and one

Radeon AI PRO R9700 in it, the 9800X3D and the 9070XT are out.

That box is where both local arms of the benchmark were run, base

model and tuned model, same card, same settings. The

Threadripper 2970WX has two R9700s now and does the training

under ROCm. So three R9700s total across two machines, 96GB of

VRAM, no CUDA anywhere. I fought ROCm plenty (some of you saw the

Windows thread) but it does the job and the numbers came out of

it.

Links:

Hugging Face:

GitHub mirror:

What is not in the release is the training data and the tuned

weights. What is in the release is every prompt and every result

row, so you can run your own model against the same 184 and

compare.

Caveat thats also on the card: the test prompts come from the

same generator family as the training data, so this measures how

good the model got at my kind of spec, not how it does on random

human written requests. I have a bigger private holdout in the

works and a version 1.1 of the card coming with those results.

Still grinding. If you are on AMD and want to compare notes on

ROCm training or eval, ask away.

MQL5 is a trademark of its owner. Not affiliated, independent

project.

── more in #large-language-models 4 stories · sorted by recency
── more on @qwen2.5-coder-14b-instruct 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/2-184-to-170-184-one…] indexed:0 read:2min 2026-09-12 ·