Bitsandbytes documentation

4-bit quantization

Hugging Face's logo
Join the Hugging Face community

and get access to the augmented documentation experience

to get started

4-bit quantization

QLoRA is a finetuning method that quantizes a model to 4-bits and adds a set of low-rank adaptation (LoRA) weights to the model and tuning them through the quantized weights. This method also introduces a new data type, 4-bit NormalFloat (LinearNF4) in addition to the standard Float4 data type (LinearFP4). LinearNF4 is a quantization data type for normally distributed data and can improve performance.

Linear4bit

class bitsandbytes.nn.Linear4bit

< >

( input_features output_features bias = True compute_dtype = None compress_statistics = True quant_type = 'fp4' quant_storage = torch.uint8 device = None )

This class is the base module for the 4-bit quantization algorithm presented in QLoRA. QLoRA 4-bit linear layers uses blockwise k-bit quantization under the hood, with the possibility of selecting various compute datatypes such as FP4 and NF4.

In order to quantize a linear layer one should first load the original fp16 / bf16 weights into the Linear4bit module, then call quantized_module.to("cuda") to quantize the fp16 / bf16 weights.

Example:

import torch
import torch.nn as nn

import bitsandbytes as bnb
from bnb.nn import Linear4bit

fp16_model = nn.Sequential(
    nn.Linear(64, 64),
    nn.Linear(64, 64)
)

quantized_model = nn.Sequential(
    Linear4bit(64, 64),
    Linear4bit(64, 64)
)

quantized_model.load_state_dict(fp16_model.state_dict())
quantized_model = quantized_model.to(0) # Quantization happens here

__init__

< >

( input_features output_features bias = True compute_dtype = None compress_statistics = True quant_type = 'fp4' quant_storage = torch.uint8 device = None )

Parameters

  • input_features (str) — Number of input features of the linear layer.
  • output_features (str) — Number of output features of the linear layer.
  • bias (bool, defaults to True) — Whether the linear class uses the bias term as well.

Initialize Linear4bit class.

LinearFP4

[[autdodoc]] bitsandbytes.nn.LinearFP4

  • init

LinearNF4

class bitsandbytes.nn.LinearNF4

< >

( input_features output_features bias = True compute_dtype = None compress_statistics = True quant_storage = torch.uint8 device = None )

Implements the NF4 data type.

Constructs a quantization data type where each bin has equal area under a standard normal distribution N(0, 1) that is normalized into the range [-1, 1].

For more information read the paper: QLoRA: Efficient Finetuning of Quantized LLMs (https://arxiv.org/abs/2305.14314)

Implementation of the NF4 data type in bitsandbytes can be found in the create_normal_map function in the functional.py file: https://github.com/TimDettmers/bitsandbytes/blob/main/bitsandbytes/functional.py#L236.

__init__

< >

( input_features output_features bias = True compute_dtype = None compress_statistics = True quant_storage = torch.uint8 device = None )

Parameters

  • input_features (str) — Number of input features of the linear layer.
  • output_features (str) — Number of output features of the linear layer.
  • bias (bool, defaults to True) — Whether the linear class uses the bias term as well.

Params4bit

class bitsandbytes.nn.Params4bit

< >

( data: Optional = None requires_grad = False quant_state: Optional = None blocksize: int = 64 compress_statistics: bool = True quant_type: str = 'fp4' quant_storage: dtype = torch.uint8 module: Optional = None bnb_quantized: bool = False )

__init__

( *args **kwargs )

Initialize self. See help(type(self)) for accurate signature.

< > Update on GitHub