This model is a 4-bit quantized version of the original zeroentropy/zembed-1 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: 8044.94 MB (BFLOAT16)
Quantized Footprint: 2594.96 MB (UINT8)
Memory Reduction: 67.7%

Usage

python

from transformers import AutoModel, AutoTokenizer

model_name = "zembed-1-bnb-4bit-nf4"
model = AutoModel.from_pretrained(
    "manu02/zembed-1-bnb-4bit-nf4",
)
tokenizer = AutoTokenizer.from_pretrained("manu02/zembed-1-bnb-4bit-nf4", use_fast=True)

manu02/zembed-1-bnb-4bit-nf4-dq

zembed-1 (Quantized)

Description

Quantization Details

Usage

manu02/
zembed-1-bnb-4bit-nf4-dq