xmadai/Llama-3.1-405B-Instruct-xMADai-4bit

This model has been xMADified!

This repository contains meta-llama/Meta-Llama-3.1-405B-Instruct quantized, using xMAD.ai proprietary technology, from 16-bit floats to 4-bit integers.

How to Run Model

Loading the model checkpoint of this xMADified model requires < 200 GiB of VRAM.

Hence it can be efficiently run on 1 node of 8 x V100-32GB GPUs, or 3 x A100-80GB GPUs.

Package prerequisites: Run the following commands to install the required packages.

pip install -q --upgrade transformers accelerate optimum
pip install -q --no-build-isolation auto-gptq

Sample Inference Code

from transformers import AutoTokenizer
from auto_gptq import AutoGPTQForCausalLM

model_id = "xmadai/Llama-3.1-405B-Instruct-xMADai-4bit"
prompt = [
  {"role": "system", "content": "You are a helpful assistant, that responds as a pirate."},
  {"role": "user", "content": "What's Deep Learning?"},
]

tokenizer = AutoTokenizer.from_pretrained(model_id)

inputs = tokenizer.apply_chat_template(
  prompt,
  tokenize=True,
  add_generation_prompt=True,
  return_tensors="pt",
  return_dict=True,
).to("cuda")

model = AutoGPTQForCausalLM.from_quantized(
    model_id,
    device_map='auto',
    trust_remote_code=True,
)

outputs = model.generate(**inputs, do_sample=True, max_new_tokens=256)
print(tokenizer.batch_decode(outputs, skip_special_tokens=True))

Model Quality

We report the zero-shot accuracy of this xMADified model on popular benchmarks below. The results are obtained using lm-evaluation-harness.

Model	Arc Challenge	Arc Easy	LAMBADA OpenAI	LAMBADA Standard	MMLU Humanities	MMLU STEM	WinoGrande
xMADified Llama-3.1-405B-Instruct	64.76	88.26	77.08	73.32	82.74	82.18	81.22

Other xMADified models and their GPU memory requirements are listed below.

For additional xMADified models, access to fine-tuning, and general questions, please contact us at [email protected] and join our waiting list.