Modelpublic

ranjankn/gpt2-large-agnews-quantization-bitsandbytes

sourceHugging Facemitupdated 1y agoView on Hugging Face

0likes

Model Card

GPT-2 Large AG News - bitsandbytes Quantization Results

Quantization results using bitsandbytes library.

Methods

—INT8 (8-bit)
—NF4 (4-bit)

Results

See task_2_3_results.json for complete metrics.

Base model: ranjan56cse/gpt2-large-agnews-classifier