pplx-embed-v1-0.6b (Quantized)

Description

This model is a 4-bit quantized version of the original perplexity-ai/pplx-embed-v1-0.6b model, optimized for reduced memory usage while maintaining performance.

Quantization Details

• Quantization Type: 4-bit
• bnb_4bit_quant_type: nf4
• bnb_4bit_use_double_quant: True
• bnb_4bit_compute_dtype: bfloat16
• bnb_4bit_quant_storage: uint8
• Original Footprint: 2384.20 MB (FLOAT32)
• Quantized Footprint: 842.79 MB (UINT8)
• Memory Reduction: 64.7%

Usage

from transformers import AutoModel, AutoTokenizer

model_name = "pplx-embed-v1-0.6b-bnb-4bit-nf4"
model = AutoModel.from_pretrained(
    "manu02/pplx-embed-v1-0.6b-bnb-4bit-nf4",
)
tokenizer = AutoTokenizer.from_pretrained("manu02/pplx-embed-v1-0.6b-bnb-4bit-nf4", use_fast=True)