Llama 4 Maverick 17B 128E Instruct FP8

🚀 Llama 4 Model

Llama 4 is a collection of natively multimodal AI models that offer text and multimodal experiences. Leveraging a mixture-of-experts architecture, these models deliver industry-leading performance in text and image understanding.

🚀 Quick Start

Please ensure you have transformers v4.51.0 installed. You can upgrade it using the following command:

pip install -U transformers

Here is a basic usage example:

from transformers import AutoTokenizer, Llama4ForConditionalGeneration
import torch

model_id = "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8"

tokenizer = AutoTokenizer.from_pretrained(model_id)

messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt", return_dict=True)

model = Llama4ForConditionalGeneration.from_pretrained(
    model_id,
    tp_plan="auto",
    torch_dtype="auto",
)

outputs = model.generate(**inputs.to(model.device), max_new_tokens=100)
outputs = tokenizer.batch_decode(outputs[:, inputs["input_ids"].shape[-1]:])
print(outputs[0])

✨ Features

• Multimodal Capabilities: Llama 4 models support both text and image processing, enabling a wide range of applications such as visual recognition, image reasoning, and captioning.
• Mixture-of-Experts Architecture: This architecture allows the models to achieve industry-leading performance in text and image understanding.
• Multiple Language Support: The models are designed for commercial and research use in multiple languages, including Arabic, English, French, German, Hindi, Indonesian, Italian, Portuguese, Spanish, Tagalog, Thai, and Vietnamese.

📦 Installation

To use the Llama 4 models, you need to install the transformers library. You can install it using pip:

pip install -U transformers

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, Llama4ForConditionalGeneration
import torch

model_id = "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8"

tokenizer = AutoTokenizer.from_pretrained(model_id)

messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt", return_dict=True)

model = Llama4ForConditionalGeneration.from_pretrained(
    model_id,
    tp_plan="auto",
    torch_dtype="auto",
)

outputs = model.generate(**inputs.to(model.device), max_new_tokens=100)
outputs = tokenizer.batch_decode(outputs[:, inputs["input_ids"].shape[-1]:])
print(outputs[0])

📚 Documentation

Model Information

The Llama 4 collection of models are natively multimodal AI models. We are launching two efficient models in the Llama 4 series:

• Llama 4 Scout: A 17 billion parameter model with 16 experts.
• Llama 4 Maverick: A 17 billion parameter model with 128 experts.

Model developer: Meta

Model Architecture: The Llama 4 models are auto-regressive language models that use a mixture-of-experts (MoE) architecture and incorporate early fusion for native multimodality.

Intended Use

Intended Use Cases: Llama 4 is intended for commercial and research use in multiple languages. Instruction tuned models are intended for assistant-like chat and visual reasoning tasks, whereas pretrained models can be adapted for natural language generation. For vision, Llama 4 models are also optimized for visual recognition, image reasoning, captioning, and answering general questions about an image. The Llama 4 Community License allows for these use cases.

Out-of-scope:

• Use in any manner that violates applicable laws or regulations (including trade compliance laws).
• Use in any other way that is prohibited by the Acceptable Use Policy and Llama 4 Community License.
• Use in languages or capabilities beyond those explicitly referenced as supported in this model card.

⚠️ Important Note

1. Llama 4 has been trained on a broader collection of languages than the 12 supported languages (pre-training includes 200 total languages). Developers may fine-tune Llama 4 models for languages beyond the 12 supported languages provided they comply with the Llama 4 Community License and the Acceptable Use Policy. Developers are responsible for ensuring that their use of Llama 4 in additional languages is done in a safe and responsible manner.
2. Llama 4 has been tested for image understanding up to 5 input images. If leveraging additional image understanding capabilities beyond this, Developers are responsible for ensuring that their deployments are mitigated for risks and should perform additional testing and tuning tailored to their specific applications.

Hardware and Software

Training Factors: We used custom training libraries, Meta's custom built GPU clusters, and production infrastructure for pretraining. Fine-tuning, quantization, annotation, and evaluation were also performed on production infrastructure.

Training Energy Use: Model pre-training utilized a cumulative of 7.38M GPU hours of computation on H100-80GB (TDP of 700W) type hardware. The methodology used to determine training energy use and greenhouse gas emissions can be found here. Since Meta is openly releasing these models, the training energy use and greenhouse gas emissions will not be incurred by others.

Benchmarks

In this section, we report the results for Llama 4 relative to our previous models. All reported evaluations and testing were conducted on bf16 models.

Pre-trained models

Instruction tuned models

^reported numbers for MMMU Pro is the average of Standard and Vision tasks

Quantization

The Llama 4 Scout model is released as BF16 weights, but can fit within a single H100 GPU with on-the-fly int4 quantization. The Llama 4 Maverick model is released as both BF16 and FP8 quantized weights. The FP8 quantized weights fit on a single H100 DGX host while still maintaining quality. We provide code for on-the-fly int4 quantization which minimizes performance degradation as well.

Safeguards

As part of our release approach, we followed a three-pronged strategy to manage risks:

• Enable developers to deploy helpful, safe and flexible experiences for their target audience and for the use cases supported by Llama.
• Protect developers against adversarial users aiming to exploit Llama capabilities to potentially cause harm.
• Provide protections for the community to help prevent the misuse of our models.

📄 License

A custom commercial license, the Llama 4 Community License Agreement, is available at: https://github.com/meta-llama/llama-models/blob/main/models/llama4/LICENSE

⚠️ Important Note

Please read the LLAMA 4 COMMUNITY LICENSE AGREEMENT carefully before using the Llama 4 models.