LlamaFactory

Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

Runssource: GitHubhomepagePythonApache-2.0commit d6bb97ddff5d

Python, Apache-2.0 licensed. The project labels itself: agent, ai, deepseek, fine tuning, gemma, gpt, instruction tuning and large language models.

LlamaFactory runs. An Argusic agent installed it in 30 minutes and hit no errors on a clean machine with no GPU, and the whole session was recorded.

LLaMA Factory 0.9.6.dev0 is installed with all dependencies, CLI responds to all subcommands, LlamaBoard Web UI serves HTTP 200 on port 7860, and 325 automated tests pass on CPU (GPU-dependent tests skipped).

What the agent ended up with on a clean machine, in its own words. How this is measured

Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 97.3 of 100 (the mean of this project's run scores).

At a glance

verdict
Runs
Argusic Score
97.3 / 100
cost of the verifying run
$0.06 (measured)
recorded runs
3
last tested
stars
74,987
forks
9,179
open issues
1,154
watchers
348
size
14 MB
created
last push

Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.

What was measured

From the run that decided this verdict. Every line is a recorded value, not an estimate.

Install time
30 minutes
Cold machine to finish
14 minutes
Errors hit and fixed
0 hit, 0 fixed with no human help
How the result was proved
pip install -e . succeeded; llamafactory-cli version printed version 0.9.6.dev0; pytest tests/ (254 passed, 14 skipped, 2 xfailed, 4 xpassed) and tests_v1/ (71 passed, 9 skipped, 1 xfailed); curl http://127.0.0.1:7860 returned HTTP 200 from LlamaBoard Web UI
Model tokens used
119,274
Exact commit tested
d6bb97ddff5d
Container image
sha256:cdd920bce783

What was not measured

The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.

Peak GPU memory
this ran on a CPU machine, so there was no GPU to sample
Behaviour over a long run
each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load

Watch the session

The whole run, from a clean machine, at the speed it actually happened. Nothing is cut and nothing is sped up.

A recording of attempt 2, run , the attempt that decided the verdict above. The terminal recording it was made from is the evidence, stored with a sha256 fingerprint; the video is a rendering of it. Nothing loads from the video host until you press play.

The end of the recording

The last lines of the terminal session that decided this verdict, exactly as they were recorded.

exec
/bin/sh -lc 'echo "Total elapsed: $(date -d @$(echo "$(date +%s) - $(stat -c %Y /proc/1/cmdline 2>/…
 succeeded in 0ms:
/bin/sh: 1: bc: not found
Total elapsed: 1789360132
cleaned
exec
/bin/sh -lc 'date +%s' in /work/repo
 succeeded in 0ms:
1789360137
tokens used
119,274

Replay the whole session, every command from a clean machine to this point.

How it was tested

One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.

Strengths and limits

Measured facts, not opinions. How this is written.

What went well

  • Reached a running state on a clean machine, with the session recorded.
  • Nothing broke on the way: zero errors between clone and running.
  • Ran without a GPU, so it does not need one to start.
  • Licensed Apache-2.0, as reported by its host.
  • Measured 3 times, so the result is not a one-off.

What did not, or is not known

  • Took 30 minutes to install, slower than the median of the 32 comparable projects Argusic has measured.

Of the 33 agent projects Argusic has installed and timed, LlamaFactory was the 30th fastest to reach a running state, and 25 of 33 reached one at all.

What it is a reasonable choice for

  • Trying it on a laptop or a small server: it reached a running state without a GPU.
  • Following its own README: the documented path worked with nothing to fix.

Also tested, in the same area

Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.

Run history

tested byattemptoutcomescorecost USDwhen
Argusic Runner3/3Runs with mocks92.000.15
Argusic Runner2/3Runs100.000.06
Argusic Runner1/3Runs100.000.24

Topics (from GitHub)

agentaideepseekfine-tuninggemmagptinstruction-tuninglarge-language-modelsllamallama3llmloramoenlppeftqloraquantizationqwenrlhftransformers

Embed the badge

Markdown for the project README. It links back here; terms on the terms page.

[![Tested by Argusic](https://argusic.com/badge/LlamaFactory.svg)](https://argusic.com/subject/llamafactory)

Questions

Does LlamaFactory run?
Yes. LlamaFactory runs. Argusic installed and launched it on a clean machine in 30 minutes, hitting 0 errors on the way, and recorded the session.
How did Argusic test LlamaFactory?
On a fresh, disposable machine, with every command recorded and the repository pinned at commit d6bb97ddff5d. 3 attempts are recorded, and the full method is on the methodology page.
What did it cost to test LlamaFactory?
The run that produced this verdict cost $0.06: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
How long does LlamaFactory take to install?
30 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
Does LlamaFactory need a GPU?
Not to start. Argusic reached a running state on a machine with no GPU.
What goes wrong when installing LlamaFactory?
Nothing did, in the recorded run: zero errors between clone and running.
How does LlamaFactory compare with the alternatives?
Of the 33 agent projects Argusic has installed and timed, LlamaFactory was the 30th fastest to reach a running state, and 25 of 33 reached one at all.
Where is the evidence for LlamaFactory?
All 3 recorded runs are on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.

Discussion