Catalog
huggingface/huggingface-community-evals

huggingface

huggingface-community-evals

Run evaluations for Hugging Face Hub models using inspect-ai and lighteval on local hardware. Use for backend selection, local GPU evals, and choosing between vLLM / Transformers / accelerate. Not for HF Jobs orchestration, model-card PRs, .eval_results publication, or community-evals automation.

global
New~1.6k
v1.0Saved Jul 11, 2026

Overview

This skill is for running evaluations against models on the Hugging Face Hub on local hardware.

It covers:

  • inspect-ai with local inference
  • lighteval with local inference
  • choosing between vllm, Hugging Face Transformers, and accelerate
  • smoke tests, task selection, and backend fallback strategy

It does not cover:

  • Hugging Face Jobs orchestration
  • model-card or model-index edits
  • README table extraction
  • Artificial Analysis imports
  • .eval_results generation or publishing
  • PR creation or community-evals automation

If the user wants to run the same eval remotely on Hugging Face Jobs, hand off to the hugging-face-jobs skill and pass it one of the local scripts in this skill.

If the user wants to publish results into the community evals workflow, stop after generating the evaluation run and hand off that publishing step to ~/code/community-evals.

All paths below are relative to the directory containing this SKILL.md.

When To Use Which Script

Use case Script
Local inspect-ai eval on a Hub model via inference providers scripts/inspect_eval_uv.py
Local GPU eval with inspect-ai using vllm or Transformers scripts/inspect_vllm_uv.py
Local GPU eval with lighteval using vllm or accelerate scripts/lighteval_vllm_uv.py
Extra command patterns examples/USAGE_EXAMPLES.md

Prerequisites

  • Prefer uv run for local execution.
  • Set HF_TOKEN for gated/private models.
  • For local GPU runs, verify GPU access before starting:
uv --version
printenv HF_TOKEN >/dev/null
nvidia-smi

If nvidia-smi is unavailable, either:

  • use scripts/inspect_eval_uv.py for lighter provider-backed evaluation, or
  • hand off to the hugging-face-jobs skill if the user wants remote compute.

Core Workflow

  1. Choose the evaluation framework.
    • Use inspect-ai when you want explicit task control and inspect-native flows.
    • Use lighteval when the benchmark is naturally expressed as a lighteval task string, especially leaderboard-style tasks.
  2. Choose the inference backend.
    • Prefer vllm for throughput on supported architectures.
    • Use Hugging Face Transformers (--backend hf) or accelerate as compatibility fallbacks.
  3. Start with a smoke test.
    • inspect-ai: add --limit 10 or similar.
    • lighteval: add --max-samples 10.
  4. Scale up only after the smoke test passes.
  5. If the user wants remote execution, hand off to hugging-face-jobs with the same script + args.

Quick Start

Option A: inspect-ai with local inference providers path

Best when the model is already supported by Hugging Face Inference Providers and you want the lowest local setup overhead.

uv run scripts/inspect_eval_uv.py \
  --model meta-llama/Llama-3.2-1B \
  --task mmlu \
  --limit 20

Use this path when:

  • you want a quick local smoke test
  • you do not need direct GPU control
  • the task already exists in inspect-evals

Option B: inspect-ai on Local GPU

Best when you need to load the Hub model directly, use vllm, or fall back to Transformers for unsupported architectures.

Local GPU:

uv run scripts/inspect_vllm_uv.py \
  --model meta-llama/Llama-3.2-1B \
  --task gsm8k \
  --limit 20

Transformers fallback:

uv run scripts/inspect_vllm_uv.py \
  --model microsoft/phi-2 \
  --task mmlu \
  --backend hf \
  --trust-remote-code \
  --limit 20

Option C: lighteval on Local GPU

Best when the task is naturally expressed as a lighteval task string, especially Open LLM Leaderboard style benchmarks.

Local GPU:

uv run scripts/lighteval_vllm_uv.py \
  --model meta-llama/Llama-3.2-3B-Instruct \
  --tasks "leaderboard|mmlu|5,leaderboard|gsm8k|5" \
  --max-samples 20 \
  --use-chat-template

accelerate fallback:

uv run scripts/lighteval_vllm_uv.py \
  --model microsoft/phi-2 \
  --tasks "leaderboard|mmlu|5" \
  --backend accelerate \
  --trust-remote-code \
  --max-samples 20

Remote Execution Boundary

This skill intentionally stops at local execution and backend selection.

If the user wants to:

  • run these scripts on Hugging Face Jobs
  • pick remote hardware
  • pass secrets to remote jobs
  • schedule recurring runs
  • inspect / cancel / monitor jobs

then switch to the hugging-face-jobs skill and pass it one of these scripts plus the chosen arguments.

Task Selection

inspect-ai examples:

  • mmlu
  • gsm8k
  • hellaswag
  • arc_challenge
  • truthfulqa
  • winogrande
  • humaneval

lighteval task strings use suite|task|num_fewshot:

  • leaderboard|mmlu|5
  • leaderboard|gsm8k|5
  • leaderboard|arc_challenge|25
  • lighteval|hellaswag|0

Multiple lighteval tasks can be comma-separated in --tasks.

Backend Selection

  • Prefer inspect_vllm_uv.py --backend vllm for fast GPU inference on supported architectures.
  • Use inspect_vllm_uv.py --backend hf when vllm does not support the model.
  • Prefer lighteval_vllm_uv.py --backend vllm for throughput on supported models.
  • Use lighteval_vllm_uv.py --backend accelerate as the compatibility fallback.
  • Use inspect_eval_uv.py when Inference Providers already cover the model and you do not need direct GPU control.

Hardware Guidance

Model size Suggested local hardware
< 3B consumer GPU / Apple Silicon / small dev GPU
3B - 13B stronger local GPU
13B+ high-memory local GPU or hand off to hugging-face-jobs

For smoke tests, prefer cheaper local runs plus --limit or --max-samples.

Troubleshooting

  • CUDA or vLLM OOM:
    • reduce --batch-size
    • reduce --gpu-memory-utilization
    • switch to a smaller model for the smoke test
    • if necessary, hand off to hugging-face-jobs
  • Model unsupported by vllm:
    • switch to --backend hf for inspect-ai
    • switch to --backend accelerate for lighteval
  • Gated/private repo access fails:
    • verify HF_TOKEN
  • Custom model code required:
    • add --trust-remote-code

Examples

See:

  • examples/USAGE_EXAMPLES.md for local command patterns
  • scripts/inspect_eval_uv.py
  • scripts/inspect_vllm_uv.py
  • scripts/lighteval_vllm_uv.py
Files6
6 files · 34.1 KB

Select a file to preview

Overall Score

82/100

Grade

B

Good

Safety

80

Quality

85

Clarity

85

Completeness

78

Summary

This skill provides structured guidance for running local evaluations of Hugging Face Hub models using `inspect-ai` and `lighteval` frameworks. It covers backend selection (vLLM, Transformers, accelerate, inference providers), GPU configuration, smoke testing patterns, and hardware-aware troubleshooting. The skill intentionally stops at local execution and directs users to other skills for remote orchestration or result publishing.

Static Analysis Findings

1 finding

Patterns detected by deterministic static analysis before AI scoring. Hover over any finding code for detailed information and remediation guidance.

Command Injection
SEC-011Dynamic Shell Eval3x in 2 files

Shell eval/exec of dynamic content

scripts/inspect_vllm_uv.pyeval"2x
scripts/inspect_eval_uv.pyeval"

Detected Capabilities

shell execution (subprocess.run)environment variable read (HF_TOKEN)argument parsing and CLI invocationfile read (task introspection)subprocess spawning (inspect, lighteval, vllm commands)

Trigger Keywords

Phrases that MCP clients use to match this skill to user intent.

local model evaluationinspect-ai benchmarklighteval task runningvllm gpu inferencehuggingface eval backend

Risk Signals

INFO

eval() not detected in actual code analysis

scripts/inspect_vllm_uv.py, scripts/lighteval_vllm_uv.py, scripts/inspect_eval_uv.py
INFO

No hardcoded credentials or secrets

All files
INFO

HF_TOKEN environment variable read for authentication

All Python scripts (setup_environment functions)
INFO

subprocess.run() with check=True for error handling

scripts/inspect_vllm_uv.py:run_inspect_vllm, run_inspect_hf; scripts/lighteval_vllm_uv.py:run_lighteval_vllm, run_lighteval_accelerate; scripts/inspect_eval_uv.py:main
WARNING

Dynamic command construction from user arguments

All Python scripts (cmd list building)
WARNING

trust-remote-code flag passed to models without validation

scripts/inspect_vllm_uv.py, scripts/lighteval_vllm_uv.py, scripts/inspect_eval_uv.py

Referenced Domains

External domains referenced in skill content, detected by static analysis.

huggingface.cowww.apache.org

Use Cases

  • Run local GPU evaluations on Hugging Face models using inspect-ai
  • Execute lighteval benchmarks with vLLM or accelerate backends
  • Test evaluation scripts with smoke tests before scaling up
  • Choose between vLLM, Transformers, and accelerate backends for model compatibility
  • Troubleshoot OOM and backend compatibility issues in local evaluation runs

Quality Notes

  • Strong documentation with clear scope boundaries: explicitly states what it covers (local evals, backend selection) and what it does NOT cover (Jobs orchestration, publication, model cards)
  • Well-structured skill.md with clear workflow sections: prerequisites, core workflow, quick-start patterns, hardware guidance, troubleshooting
  • Practical examples for each framework (inspect-ai, lighteval) and backend (vLLM, Transformers, accelerate, inference providers)
  • Supporting examples file (USAGE_EXAMPLES.md) provides concrete command patterns
  • Clean hand-off boundaries: explicitly directs users to hugging-face-jobs skill for remote execution and community-evals for publishing
  • Python scripts have good error handling with subprocess.CalledProcessError and descriptive exit messages
  • Each script has PEP 723 inline dependencies (requires-python, dependencies) for reproducibility
  • Task selection and hardware guidance tables provide practical decision trees
  • Minor issue: SEC-011 false positives in pattern detection — strings contain 'eval' (e.g., 'eval' in 'inspect eval') but not actual eval() calls; this is benign
Model: claude-haiku-4-5-20251001Analyzed: Jul 11, 2026

Reviews

Add this skill to your library to leave a review.

No reviews yet

Be the first to share your experience.

Add huggingface/huggingface-community-evals to your library

Command Palette

Search for a command to run...