furiosa-ai/bert-large-uncased-INT8-MLPerf

Model Overview

• Model Architecture: Bert
  • Input: Text
  • Output: Text
• Model Optimizations:
• Maximum Context Length: 384 tokens
• Intended Use Cases: Intended for commercial and non-commercial use. Same as google/bert-large-uncase, this models is intended for question-answering.
• Release Date: 04/12/2025
• Version: v2025.2
• License(s): Apache License 2.0
• Supported Inference Engine(s): Furiosa LLM
• Supported Hardware Compatibility: FuriosaAI RNGD
• Preferred Operating System(s): Linux
• Quantization:
  • Tool: Furiosa Model Compressor v0.6.2, included in Furiosa SDK 2025.2
  • Weight: int8, Activation: int8, KV cache: int8
  • Calibration: SQuAD v1.1 dataset (instruction), 100 samples

Description:

This model is the pre-compiled version of the google/bert-large-uncase, which is an embedding model that uses an optimized transformer architecture.

Usage

MLPerf Benchmark using RNGD

Follow the example command below after installing furiosa-mlperf and its prerequisites.

furiosa-mlperf bert-offline furiosa-ai/bert-large-uncased-INT8-MLPerf ./mlperf-result