Quantize π€ Transformers models
bitsandbytes
Integration
π€ Transformers is closely integrated with most used modules on bitsandbytes
. You can load your model in 8-bit precision with few lines of code.
This is supported by most of the GPU hardwares since the 0.37.0
release of bitsandbytes
.
Learn more about the quantization method in the LLM.int8() paper, or the blogpost about the collaboration.
Since its 0.39.0
release, you can load any model that supports device_map
using 4-bit quantization, leveraging FP4 data type.
If you want to quantize your own pytorch model, check out this documentation from π€ Accelerate library.
Here are the things you can do using bitsandbytes
integration
FP4 quantization
Requirements
Make sure that you have installed the requirements below before running any of the code snippets below.
Latest
bitsandbytes
librarypip install bitsandbytes>=0.39.0
Install latest
accelerate
from sourcepip install git+https://github.com/huggingface/accelerate.git
Install latest
transformers
from sourcepip install git+https://github.com/huggingface/transformers.git
Load a large model in 4bit
By using load_in_4bit=True
when calling the .from_pretrained
method, you can divide your memory use by 4 (roughly).
# pip install transformers accelerate bitsandbytes
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "bigscience/bloom-1b7"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_4bit=True)
Note that once a model has been loaded in 4-bit it is currently not possible to push the quantized weights on the Hub. Note also that you cannot train 4-bit weights as this is not supported yet. However you can use 4-bit models to train extra parameters, this will be covered in the next section.
Load a large model in 8bit
You can load a model by roughly halving the memory requirements by using load_in_8bit=True
argument when calling .from_pretrained
method
# pip install transformers accelerate bitsandbytes
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "bigscience/bloom-1b7"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_8bit=True)
Then, use your model as you would usually use a PreTrainedModel.
You can check the memory footprint of your model with get_memory_footprint
method.
print(model.get_memory_footprint())
With this integration we were able to load large models on smaller devices and run them without any issue.
Note that once a model has been loaded in 8-bit it is currently not possible to push the quantized weights on the Hub except if you use the latest transformers
and bitsandbytes
. Note also that you cannot train 8-bit weights as this is not supported yet. However you can use 8-bit models to train extra parameters, this will be covered in the next section.
Note also that device_map
is optional but setting device_map = 'auto'
is prefered for inference as it will dispatch efficiently the model on the available ressources.
Advanced usecases
Here we will cover some advanced usecases you can perform with FP4 quantization
Change the compute dtype
The compute dtype is used to change the dtype that will be used during computation. For example, hidden states could be in float32
but computation can be set to bf16 for speedups. By default, the compute dtype is set to float32
.
import torch
from transformers import BitsAndBytesConfig
quantization_config = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_compute_dtype=torch.bfloat16)
Using NF4 (Normal Float 4) data type
You can also use the NF4 data type, which is a new 4bit datatype adapted for weights that have been initialized using a normal distribution. For that run:
from transformers import BitsAndBytesConfig
nf4_config = BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_quant_type="nf4",
)
model_nf4 = AutoModelForCausalLM.from_pretrained(model_id, quantization_config=nf4_config)
Use nested quantization for more memory efficient inference
We also advise users to use the nested quantization technique. This saves more memory at no additional performance - from our empirical observations, this enables fine-tuning llama-13b model on an NVIDIA-T4 16GB with a sequence length of 1024, batch size of 1 and gradient accumulation steps of 4.
from transformers import BitsAndBytesConfig
double_quant_config = BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_use_double_quant=True,
)
model_double_quant = AutoModelForCausalLM.from_pretrained(model_id, quantization_config=double_quant_config)
Push quantized models on the π€ Hub
You can push a quantized model on the Hub by naively using push_to_hub
method. This will first push the quantization configuration file, then push the quantized model weights.
Make sure to use bitsandbytes>0.37.2
(at this time of writing, we tested it on bitsandbytes==0.38.0.post1
) to be able to use this feature.
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("bigscience/bloom-560m", device_map="auto", load_in_8bit=True)
tokenizer = AutoTokenizer.from_pretrained("bigscience/bloom-560m")
model.push_to_hub("bloom-560m-8bit")
Pushing 8bit models on the Hub is strongely encouraged for large models. This will allow the community to benefit from the memory footprint reduction and loading for example large models on a Google Colab.
Load a quantized model from the π€ Hub
You can load a quantized model from the Hub by using from_pretrained
method. Make sure that the pushed weights are quantized, by checking that the attribute quantization_config
is present in the model configuration object.
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("{your_username}/bloom-560m-8bit", device_map="auto")
Note that in this case, you donβt need to specify the arguments load_in_8bit=True
, but you need to make sure that bitsandbytes
and accelerate
are installed.
Note also that device_map
is optional but setting device_map = 'auto'
is prefered for inference as it will dispatch efficiently the model on the available ressources.
Advanced usecases
This section is intended to advanced users, that want to explore what it is possible to do beyond loading and running 8-bit models.
Offload between cpu
and gpu
One of the advanced usecase of this is being able to load a model and dispatch the weights between CPU
and GPU
. Note that the weights that will be dispatched on CPU will not be converted in 8-bit, thus kept in float32
. This feature is intended for users that want to fit a very large model and dispatch the model between GPU and CPU.
First, load a BitsAndBytesConfig
from transformers
and set the attribute llm_int8_enable_fp32_cpu_offload
to True
:
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
quantization_config = BitsAndBytesConfig(llm_int8_enable_fp32_cpu_offload=True)
Letβs say you want to load bigscience/bloom-1b7
model, and you have just enough GPU RAM to fit the entire model except the lm_head
. Therefore write a custom device_map as follows:
device_map = {
"transformer.word_embeddings": 0,
"transformer.word_embeddings_layernorm": 0,
"lm_head": "cpu",
"transformer.h": 0,
"transformer.ln_f": 0,
}
And load your model as follows:
model_8bit = AutoModelForCausalLM.from_pretrained(
"bigscience/bloom-1b7",
device_map=device_map,
quantization_config=quantization_config,
)
And thatβs it! Enjoy your model!
Play with llm_int8_threshold
You can play with the llm_int8_threshold
argument to change the threshold of the outliers. An βoutlierβ is a hidden state value that is greater than a certain threshold.
This corresponds to the outlier threshold for outlier detection as described in LLM.int8()
paper. Any hidden states value that is above this threshold will be considered an outlier and the operation on those values will be done in fp16. Values are usually normally distributed, that is, most values are in the range [-3.5, 3.5], but there are some exceptional systematic outliers that are very differently distributed for large models. These outliers are often in the interval [-60, -6] or [6, 60]. Int8 quantization works well for values of magnitude ~5, but beyond that, there is a significant performance penalty. A good default threshold is 6, but a lower threshold might be needed for more unstable models (small models, fine-tuning).
This argument can impact the inference speed of the model. We suggest to play with this parameter to find which one is the best for your usecase.
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
model_id = "bigscience/bloom-1b7"
quantization_config = BitsAndBytesConfig(
llm_int8_threshold=10,
)
model_8bit = AutoModelForCausalLM.from_pretrained(
model_id,
device_map=device_map,
quantization_config=quantization_config,
)
tokenizer = AutoTokenizer.from_pretrained(model_id)
Skip the conversion of some modules
Some models has several modules that needs to be not converted in 8-bit to ensure stability. For example Jukebox model has several lm_head
modules that should be skipped. Play with llm_int8_skip_modules
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
model_id = "bigscience/bloom-1b7"
quantization_config = BitsAndBytesConfig(
llm_int8_skip_modules=["lm_head"],
)
model_8bit = AutoModelForCausalLM.from_pretrained(
model_id,
device_map=device_map,
quantization_config=quantization_config,
)
tokenizer = AutoTokenizer.from_pretrained(model_id)
Fine-tune a model that has been loaded in 8-bit
With the official support of adapters in the Hugging Face ecosystem, you can fine-tune models that have been loaded in 8-bit.
This enables fine-tuning large models such as flan-t5-large
or facebook/opt-6.7b
in a single google Colab. Please have a look at peft
library for more details.
Note that you donβt need to pass device_map
when loading the model for training. It will automatically load your model on your GPU. You can also set the device map to a specific device if needed (e.g. cuda:0
, 0
, torch.device('cuda:0')
). Please note that device_map=auto
should be used for inference only.
BitsAndBytesConfig
class transformers.BitsAndBytesConfig
< source >( load_in_8bit = False load_in_4bit = False llm_int8_threshold = 6.0 llm_int8_skip_modules = None llm_int8_enable_fp32_cpu_offload = False llm_int8_has_fp16_weight = False bnb_4bit_compute_dtype = None bnb_4bit_quant_type = 'fp4' bnb_4bit_use_double_quant = False **kwargs )
Parameters
-
load_in_8bit (
bool
, optional, defaults toFalse
) — This flag is used to enable 8-bit quantization with LLM.int8(). -
load_in_4bit (
bool
, optional, defaults toFalse
) — This flag is used to enable 4-bit quantization by replacing the Linear layers with FP4/NF4 layers frombitsandbytes
. -
llm_int8_threshold (
float
, optional, defaults to 6) — This corresponds to the outlier threshold for outlier detection as described inLLM.int8() : 8-bit Matrix Multiplication for Transformers at Scale
paper: https://arxiv.org/abs/2208.07339 Any hidden states value that is above this threshold will be considered an outlier and the operation on those values will be done in fp16. Values are usually normally distributed, that is, most values are in the range [-3.5, 3.5], but there are some exceptional systematic outliers that are very differently distributed for large models. These outliers are often in the interval [-60, -6] or [6, 60]. Int8 quantization works well for values of magnitude ~5, but beyond that, there is a significant performance penalty. A good default threshold is 6, but a lower threshold might be needed for more unstable models (small models, fine-tuning). -
llm_int8_skip_modules (
List[str]
, optional) — An explicit list of the modules that we do not want to convert in 8-bit. This is useful for models such as Jukebox that has several heads in different places and not necessarily at the last position. For example forCausalLM
models, the lastlm_head
is kept in its originaldtype
. -
llm_int8_enable_fp32_cpu_offload (
bool
, optional, defaults toFalse
) — This flag is used for advanced use cases and users that are aware of this feature. If you want to split your model in different parts and run some parts in int8 on GPU and some parts in fp32 on CPU, you can use this flag. This is useful for offloading large models such asgoogle/flan-t5-xxl
. Note that the int8 operations will not be run on CPU. -
llm_int8_has_fp16_weight (
bool
, optional, defaults toFalse
) — This flag runs LLM.int8() with 16-bit main weights. This is useful for fine-tuning as the weights do not have to be converted back and forth for the backward pass. -
bnb_4bit_compute_dtype (
torch.dtype
or str, optional, defaults totorch.float32
) — This sets the computational type which might be different than the input time. For example, inputs might be fp32, but computation can be set to bf16 for speedups. -
bnb_4bit_quant_type (
str
, {fp4, nf4}, defaults tofp4
) — This sets the quantization data type in the bnb.nn.Linear4Bit layers. Options are FP4 and NF4 data types which are specified byfp4
ornf4
. -
bnb_4bit_use_double_quant (
bool
, optional, defaults toFalse
) — This flag is used for nested quantization where the quantization constants from the first quantization are quantized again. -
kwargs (
Dict[str, Any]
, optional) — Additional parameters from which to initialize the configuration object.
This is a wrapper class about all possible attributes and features that you can play with a model that has been
loaded using bitsandbytes
.
This replaces load_in_8bit
or load_in_4bit
therefore both options are mutually exclusive.
Currently only supports LLM.int8()
, FP4
, and NF4
quantization. If more methods are added to bitsandbytes
,
then more arguments will be added to this class.
from_dict
< source >( config_dict return_unused_kwargs **kwargs ) β BitsAndBytesConfig
Parameters
-
config_dict (
Dict[str, Any]
) — Dictionary that will be used to instantiate the configuration object. -
return_unused_kwargs (
bool
) — Whether or not to return a list of unused keyword arguments. Used forfrom_pretrained
method inPreTrainedModel
. -
kwargs (
Dict[str, Any]
) — Additional parameters from which to initialize the configuration object.
Returns
The configuration object instantiated from those parameters.
Instantiates a BitsAndBytesConfig from a Python dictionary of parameters.
Returns True
if the model is quantizable, False
otherwise.
Safety checker that arguments are correct - also replaces some NoneType arguments with their default values.
This method returns the quantization method used for the model. If the model is not quantizable, it returns
None
.
Serializes this instance to a Python dictionary. Returns:
Dict[str, Any]
: Dictionary of all the attributes that make up this configuration instance.
to_diff_dict
< source >(
)
β
Dict[str, Any]
Returns
Dict[str, Any]
Dictionary of all the attributes that make up this configuration instance,
Removes all attributes from config which correspond to the default config attributes for better readability and serializes to a Python dictionary.
Save this instance to a JSON file.
to_json_string
< source >(
use_diff: bool = True
)
β
str
Serializes this instance to a JSON string.
Quantization with π€ optimum
Please have a look at Optimum documentation to learn more about quantization methods that are supported by optimum
and see if these are applicable for your usecase.