How to calculate training efficiency ,i.e tokens/sec of step 1 fine tuning of llama2 model ?

Open

#923 opened on Aug 29, 2024 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty: 2/5
Estimated time: 1-3 hours
Newbie friendliness: 25/100

Issue type: Documentation
Clarity: Needs clarification
Activity status: Stale
Tech stack: python
Domain: machine-learning, performance

Research direction

Start with the linked 7b_logfile.txt and the step 1 fine-tuning logs mentioned in the issue. Identify which logged values represent processed tokens and elapsed training time, then verify the resulting efficiency against the run configuration. Done means providing a reproducible calculation and explaining which log entries support it.

Written by the indexing model from the issue text.

Description

hi all,
I'm executing step1 fine tuning of llama2 7b model on DGX system . mentioned below is the log generated . I need help in calculating the tokens/sec i.e training efficiency of the model . Can please someone help me with that.
7b_logfile.txt

Dominant language: Python
Stars: 6.8k
Forks: 1.1k
Avg merge: 2d 16h
Merged PRs (30d): 1

Contributor guide

No contributing guide indexed for this repository

First steps

Read the whole issue, then the project's contributing guide.
Comment on the issue to say you are picking it up — it saves two people doing the same work.
Fork the repository and make your change on a branch.
Open a pull request that references the issue number.

Similar issues

PlaywrightRasterizer sizes the viewport from the root attributes, ignoring an inline style Open

bug priority:low

Difficulty 2/5 1-3 hours Newbie friendliness 76/100

CyberAgent/psd2svg#436 · Opened Sep 25, 2026
[Bug]: Linux launcher entry doesn't claim the window's app id — dock/taskbar shows a placeholder icon Open

area/install-update comp/cli comp/desktop P3 sweeper:risk-compatibility type/bug

Difficulty 2/5 1-3 hours Newbie friendliness 86/100

NousResearch/hermes-agent#122386 · 1 comment · Opened Sep 25, 2026
Symbol truncation can split multi-byte UTF-8 runes and cause valid profiles to be rejected Open

ai-generated

Difficulty 2/5 1-3 hours Newbie friendliness 78/100

grafana/pyroscope#5691 · Opened Sep 25, 2026
Router doesn't proxy /reset_prefix_cache, unlike /sleep//wake_up//is_sleeping Open

Difficulty 2/5 1-3 hours Newbie friendliness 78/100

vllm-project/production-stack#1105 · Opened Sep 25, 2026
hookify: hooks decode stdin with the Windows ANSI codepage — non-ASCII payloads silently corrupt and block rules fail open Open

Difficulty 2/5 1-3 hours Newbie friendliness 88/100

anthropics/claude-plugins-official#6306 · Opened Sep 25, 2026

More Python issues