HF-Quantization/Llama-3.3-70B-Instruct-TORCHAO-W4

meta-llama/Llama-3.3-70B-Instruct (Quantized)

Description

This model is a quantized version of the original model meta-llama/Llama-3.3-70B-Instruct. It was quantized using TorchAo.

Quantization Details

• Quantization Parameters: TorchAoConfig("int4_weight_only", group_size=128)

Usage

You can use this model in your applications by loading it directly from the Hugging Face Hub.

In order to run the inference with Llama-3.3-70B-Instruct-TORCHAO-W4, torch andtorchao need to be installed as:

pip install torch torchao --upgrade

Then, preferably the latest version of transformers need to be installed, as:

pip install transformers[accelerate] --upgrade

To run the inference the model can be instantiated as any other causal language modeling model via AutoModelForCausalLM and run the inference normally.

from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Llama-3.3-70B-Instruct-TORCHAO-W4")