skills.rest
DeepLink-org
DeepLink-orgOFFICIAL
Explore all skills by this author →
WebsiteGitHub

ppu-audio-au

Evaluates speech understanding model inference accuracy and latency on PPU accelerators.

7|1|Updated Apr 17, 2026
Search Tags:#speech recognition#docker#ppu#language identification#accuracy evaluation#inference benchmark#ecapa-tdnn
One-click install
npx skills add https://github.com/DeepLink-org/DeepEval-Skills --skill ppu-audio-au-deeplink-org
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ppu-audio-au
Source: https://github.com/DeepLink-org/DeepEval-Skills/tree/main/skills/PPU/audio/ppu-audio-au
Command: npx skills add https://github.com/DeepLink-org/DeepEval-Skills --skill ppu-audio-au-deeplink-org

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Running language identification model benchmarks on PPU hardware requires manual container setup, dataset mounting, inference execution, and metric collection. This Skill automates the full evaluation workflow for the lang-id-voxlingua107-ecapa model, from Docker container launch to ACC accuracy reporting. ## Core Features & Use Cases - Automated Container Orchestration: Launches a privileged Docker container with dynamic PPU device discovery (/dev/alixpu*) and correct volume mounts for models, datasets, and configs. - Inference Benchmarking: Runs infer_runner.py against the foundation-lid LMDB dataset and captures per-sample inference time and accuracy. - Structured Metric Collection: Parses acc_report.json into a normalized result.json containing ACC, avg_inference_time, success_rate, and total_samples. - Use Case: A hardware evaluation engineer needs to verify that the ECAPA-TDNN language identification model meets accuracy targets on PPU cards; this Skill executes the full pipeline and produces comparable metrics. ## Quick Start Ask the agent to test the lang-id-voxlingua107-ecapa language identification model inference performance on PPU with the foundation-lid dataset.

Frequently Asked Questions about ppu-audio-au

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I benchmark a language identification model on PPU hardware?▼

Launch the provided Docker container with PPU devices mounted, then run infer_runner.py with the model directory, dataset config, and output paths. The Skill collects ACC accuracy and average inference time into acc_report.json automatically.

What models and datasets does PPU audio understanding evaluation support?▼

It currently supports the lang-id-voxlingua107-ecapa model, an ECAPA-TDNN language identification model covering 107 languages. The supported dataset is foundation-lid, stored as LMDB shards with a meta.json file.

What metrics are collected during speech model inference evaluation?▼

The evaluation collects ACC (language identification accuracy) as the primary metric, plus avg_inference_time per sample, success_rate, and total_samples. Results are written to acc_report.json and normalized into result.json.

Why does the PPU evaluation container fail to start?▼

Common causes include an existing container with the same name, missing /dev/alixpu* device files on the host, an unpulled image, or absent --privileged=true flag. Remove old containers with docker rm -f and verify device files with ls /dev/alixpu*.

How do I fix PPU out-of-memory errors during inference?▼

Check which PPU cards are occupied, then set ALIXPU_VISIBLE_DEVICES to an idle card before launching the Python script. Restart the inference run after switching the visible device.

Related Skills

Explore More Software Engineering Skills→
paunchygentpaunchygent

speech-model-finetuning-on-hemma

Fine-tune speech models on Hemma and Colab with GPU runtimes.

Community
Advanced
Open330Open330
1

benchmarking-ml-models

Benchmark ML models by measuring latency, throughput, memory, and accuracy.

Official
Intermediate
vikisingh23vikisingh23
4

load-testing

Run automated load and stress tests against endpoints to measure latency, throughput, and error rates.

Community
Intermediate
vikrant-projectvikrant-project
2

tensorrt-llm

Optimize large language model inference with NVIDIA TensorRT on GPU hardware.

Community
Advanced