Llama-3.1-70B-Instruct-FP8-block

Llama-3.1-70B-Instruct-FP8-block

Model Overview

• Model Architecture: LlamaForCausalLM
  • Input: Text
  • Output: Text
• Model Optimizations:
  • Weight quantization: FP8
  • Activation quantization: FP8
• Release Date:
• Version: 1.0
• Model Developers:: Red Hat

Quantized version of meta-llama/Llama-3.1-70B-Instruct.

Model Optimizations

This model was obtained by quantizing the weights and activations of meta-llama/Llama-3.1-70B-Instruct to FP8 data type. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%. Only the weights and activations of the linear operators within transformers blocks of the language model are quantized.

Deployment

Use with vLLM

1. Initialize vLLM server:

vllm serve nm-testing/Llama-3.1-70B-Instruct-FP8-block --tensor_parallel_size 4

1. Send requests to the server:

from openai import OpenAI

# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://<your-server-host>:8000/v1"

client = OpenAI(
    api_key=openai_api_key,
    base_url=openai_api_base,
)

model = "nm-testing/Llama-3.1-70B-Instruct-FP8-block"

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image_url",
                "image_url": {"url": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg"},
            },
            {"type": "text", "text": "Describe this image."},
        ],
    }
]

outputs = client.chat.completions.create(
    model=model,
    messages=messages,
)

generated_text = outputs.choices[0].message.content
print(generated_text)

Creation

This model was quantized using the llm-compressor library as shown below.

<details> <summary>Creation details</summary>

from transformers import AutoProcessor, LlamaForCausalLM

from llmcompressor import oneshot
from llmcompressor.modeling import replace_modules_for_calibration
from llmcompressor.modifiers.quantization import QuantizationModifier

MODEL_ID = "meta-llama/Llama-3.1-70B-Instruct"

# Load model.
model = LlamaForCausalLM.from_pretrained(MODEL_ID, dtype="auto")
processor = AutoProcessor.from_pretrained(MODEL_ID)
model = replace_modules_for_calibration(model)

# Configure the quantization algorithm and scheme.
# In this case, we:
#   * quantize the weights to fp8 with per-block quantization
#   * quantize the activations to fp8 with dynamic token activations
recipe = QuantizationModifier(
    targets="Linear",
    scheme="FP8_BLOCK",
    ignore=["lm_head"],
)

# Apply quantization.
oneshot(model=model, recipe=recipe)

# Save to disk in compressed-tensors format.
SAVE_DIR = MODEL_ID.rstrip("/").split("/")[-1] + "-FP8-block"
model.save_pretrained(SAVE_DIR)
processor.save_pretrained(SAVE_DIR)

</details>

Evaluation

The model was evaluated on the OpenLLMv1 leaderboard task, using lm-evaluation-harness, on reasoning tasks using lighteval. vLLM was used for all evaluations.

<details> <summary>Evaluation details</summary>

lm-evaluation-harness

lm_eval \
  --model vllm \
  --model_args pretrained="nm-testing/Llama-3.1-70B-Instruct-FP8-block",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=4,gpu_memory_utilization=0.8,enable_chunked_prefill=True \
  --tasks openllm \
  --write_out \
  --batch_size auto \
  --output_path output_dir \
  --show_config

lighteval

lighteval_model_arguments.yaml

model_parameters:
  model_name: nm-testing/Llama-3.1-70B-Instruct-FP8-block
  dtype: auto
  gpu_memory_utilization: 0.9
  generation_parameters:
    temperature: 0.6
    min_p: 0.0
    top_p: 0.95
    top_k: 20
    max_new_tokens: 32768

lighteval vllm \
  --model_args lighteval_model_arguments.yaml \
  --tasks lighteval|aime25|0 \

</details>

Accuracy

<table> <thead> <tr> <th>Category</th> <th>Metric</th> <th>meta-llama/Llama-3.1-70B-Instruct</th> <th>nm-testing/Llama-3.1-70B-Instruct-FP8-block</th> <th>Recovery (%)</th> </tr> </thead> <tbody> <!-- OpenLLM Leaderboard V1 --> <tr> <td rowspan="7"><b>OpenLLM V1</b></td> <td>ARC-Challenge (Acc-Norm, 25-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>GSM8K (Strict-Match, 5-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>HellaSwag (Acc-Norm, 10-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>MMLU (Acc, 5-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>TruthfulQA (MC2, 0-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>Winogrande (Acc, 5-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td><b>Average Score</b></td> <td><b>abc</b></td> <td><b>ijk</b></td> <td><b>xyz</b></td> </tr> <!-- OpenLLM Leaderboard V2 --> <tr> <td rowspan="7"><b>OpenLLM V2</b></td> <td>IFEval (Inst Level Strict Acc, 0-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>BBH (Acc-Norm, 3-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>Math-Hard (Exact-Match, 4-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>GPQA (Acc-Norm, 0-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>MUSR (Acc-Norm, 0-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td>MMLU-Pro (Acc, 5-shot)</td> <td>abc</td> <td>ijk</td> <td>xyz</td> </tr> <tr> <td><b>Average Score</b></td> <td><b>abc</b></td> <td><b>ijk</b></td> <td><b>xyz</b></td> </tr> <!-- HumanEval --> <tr> <td rowspan="2"><b>Coding</b></td> <td>HumanEval Pass@1</td> <td>abc</td> <td>ijk</td> <td><b>xyz</b></td> </tr> </tbody> </table>