Transformers documentation

Efficient Inference on a Single GPU

Hugging Face's logo
Join the Hugging Face community

and get access to the augmented documentation experience

to get started

Efficient Inference on a Single GPU

This document will be completed soon with information on how to infer on a single GPU. In the meantime you can check out the guide for training on a single GPU and the guide for inference on CPUs.

bitsandbytes integration for Int8 mixed-precision matrix decomposition

Note that this feature is also totally applicable in a multi GPU setup as well.

From the paper LLM.int8() : 8-bit Matrix Multiplication for Transformers at Scale, we support HuggingFace integration for all models in the Hub with a few lines of code. The method reduce nn.Linear size by 2 for float16 and bfloat16 weights and by 4 for float32 weights, with close to no impact to the quality by operating on the outliers in half-precision.

HFxbitsandbytes.png

Int8 mixed-precision matrix decomposition works by separating a matrix multiplication into two streams: (1) a systematic feature outlier stream matrix multiplied in fp16 (0.01%), (2) a regular stream of int8 matrix multiplication (99.9%). With this method, int8 inference with no predictive degradation is possible for very large models. For more details regarding the method, check out the paper or our blogpost about the integration.

MixedInt8.gif

Note, that you would require a GPU to run mixed-8bit models as the kernels have been compiled for GPUs only. Make sure that you have enough GPU memory to store the quarter (or half if your model weights are in half precision) of the model before using this feature. Below are some notes to help you use this module, or follow the demos on Google colab.

Requirements

  • Make sure you run that on NVIDIA GPUs that support 8-bit tensor cores (Turing, Ampere or newer architectures - e.g. T4, RTX20s RTX30s, A40-A100).
  • Install the correct version of bitsandbytes by running: pip install bitsandbytes>=0.31.5
  • Install accelerate pip install accelerate>=0.12.0

Running mixed-int8 models - single GPU setup

After installing the required libraries, the way to load your mixed 8-bit model is as follows:

model_name = "bigscience/bloom-2b5"
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", load_in_8bit=True)

Running mixed-int8 models - multi GPU setup

The way to load your mixed 8-bit model in multiple GPUs is as follows (same command as single GPU setup):

model_name = "bigscience/bloom-2b5"
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", load_in_8bit=True)

But you can control the GPU RAM you want to allocate on each GPU using accelerate. Use the max_memory argument as follows:

max_memory_mapping = {0: "1GB", 1: "2GB"}
model_name = "bigscience/bloom-3b"
model_8bit = AutoModelForCausalLM.from_pretrained(
    model_name, device_map="auto", load_in_8bit=True, max_memory=max_memory_mapping
)

In this example, the first GPU will use 1GB of memory and the second 2GB.

Colab demos

With this method you can infer on models that were not possible to infer on a Google Colab before. Check out the demo for running T5-11b (42GB in fp32)! Using 8-bit quantization on Google Colab:

Open In Colab: T5-11b demo

Or this demo for BLOOM-3B:

Open In Colab: BLOOM-3b demo