Llm Compiler 7b Ftd

🚀 Meta Large Language Model Compiler (LLM Compiler)

Meta Large Language Model Compiler (LLM Compiler) is a state - of - the - art LLM that builds upon Code Llama. It offers improved performance for code optimization and compiler reasoning, and is free for both research and commercial use.

🚀 Quick Start

To access this model, you need to share contact information with Meta. The information you provide will be collected, stored, processed and shared in accordance with the Meta Privacy Policy.

The LLM Compiler models are available in two sizes (7B and 13B parameters) and two flavors: the foundational LLM Compiler and the fine - tuned LLM Compiler FTD. You can download them from the following links:

• [llm - compiler - 7b](https://huggingface.co/facebook/llm - compiler - 7b)
• [llm - compiler - 7b - ftd](https://huggingface.co/facebook/llm - compiler - 7b - ftd)
• [llm - compiler - 13b](https://huggingface.co/facebook/llm - compiler - 13b)
• [llm - compiler - 13b - ftd](https://huggingface.co/facebook/llm - compiler - 13b - ftd)

When using the LLM Compiler models, users must abide by the LLM Compiler License Agreement, which incorporates the [Acceptable Use Policy](https://llama.meta.com/llama3/use - policy) for Llama Materials.

✨ Features

• Based on Code Llama: LLM Compiler builds upon Code Llama, with enhanced performance for code optimization and compiler reasoning.
• Two Flavors:
  • LLM Compiler: Pretrained on over 500B tokens of LLVM - IR, x86_84, ARM, and CUDA assembly codes, trained to predict the effect of LLVM optimizations.
  • LLM Compiler FTD: Further fine - tuned to predict the best optimizations for code in LLVM assembly to reduce code size and to disassemble assembly code to LLVM - IR.
• Strong Understanding of Compiler Optimizations: Demonstrates far stronger understanding of compiler optimizations than existing publicly available LLMs, perfectly emulating the compiler 20% of the time.
• State - of - the - Art Results: LLM Compiler FTD sets state - of - the - art results on the tasks of optimization for code size and disassembly.

📚 Documentation

How LLM Compiler works

LLM Compiler is a specialization of Code Llama, a cutting - edge tool for code optimization using deep learning. It has been pre - trained on a large amount of LLVM assembly (IR), x86_64, ARM, and CUDA assembly codes.

It can predict the change in code size and the output code after applying a sequence of optimization passes for opt (the LLVM optimizer) to a piece of LLVM assembly. It has a deep understanding of the optimizing compiler and can often replicate its output.

In addition to the core functionality, LLM Compiler has been fine - tuned for two specific downstream tasks:

1. Predicting Best Optimization Passes: Given a piece of LLVM assembly code, it predicts the best optimization passes for opt to minimize code size.
2. Generating LLVM IR: It can generate LLVM IR from a piece of x86_64 or ARM assembly code.

The models are available in 7B and 13B parameter sizes, trained with a context window of 16,000 tokens. The 7B model is suitable for low - latency tasks, while the 13B model returns the best results.

LLM Compiler performance

• Emulating Compiler Transformations: We tested the model's ability to emulate compiler optimizations by giving it unoptimized intermediate representation and a list of optimizations. The LLM Compiler can generate character - by - character matches of expected assembly in 20% of the cases, while Code Llama struggles due to limited IR knowledge. | Model | Size | Accuracy at emulating compiler optimizations | | ---- | ---- | ---- | | Code Llama | 7B | 1.2% | | Code Llama | 13B | 0.8% | | LLM Compiler | 7B | 16% | | LLM Compiler | 13B | 20% |

• Optimizing for Code Size: We evaluated the model's ability to optimize IR for code size. Only LLM Compiler FTD models provide an improvement over -Oz. The 13B parameter model outperforms the 7B model, generating smaller object files in 61% of cases. | Model | Size | Code Size Improvement | Round trip BLEU | | ---- | ---- | ---- | ---- | | GPT - 4 Turbo | | -0.01% | 0.43 | | Code Llama Inst | 7B | -0.49% | 0.48 | | Code Llama Inst | 13B | -0.42% | 0.62 | | LLM Compiler FTD | 7B | 4.77% | 0.95 | | LLM Compiler FTD | 13B | 4.88% | 0.96 |

Releasing LLM Compiler

Meta believes that AI models, especially those for coding, benefit from an open strategy. By making LLM Compiler available, the community can explore its potential, identify problems, and fix vulnerabilities.

Responsible use

Our research paper provides in - depth information about the development process, benchmarking methods, and model limitations. Developers are advised to use compiler - specific evaluation benchmarks, rigorously test suggested compiler optimizations, and confirm the accuracy when the model decompiles assembly code.

The future of generative AI for optimisation

LLM Compiler is designed to support compiler researchers and engineers. We hope it will inspire others to create new innovative tools for research and commercial products.

📄 License

The LLM Compiler is released under the LLM Compiler License Agreement, which incorporates the [Acceptable Use Policy](https://llama.meta.com/llama3/use - policy) for Llama Materials.

The full license agreement details are as follows:

Meta Large Language Model Compiler (LLM Compiler) LICENSE AGREEMENT

Version Release Date: 27th June 2024

“Agreement” means the terms and conditions for use, reproduction, distribution and modification of the LLM Compiler Materials set forth herein.

“Documentation” means the specifications, manuals and documentation accompanying the LLM Compiler distributed by Meta at:

• [https://huggingface.co/facebook/llm - compiler - 7b](https://huggingface.co/facebook/llm - compiler - 7b)
• [https://huggingface.co/facebook/llm - compiler - 7b - ftd](https://huggingface.co/facebook/llm - compiler - 7b - ftd)
• [https://huggingface.co/facebook/llm - compiler - 13b](https://huggingface.co/facebook/llm - compiler - 13b)
• [https://huggingface.co/facebook/llm - compiler - 13b - ftd](https://huggingface.co/facebook/llm - compiler - 13b - ftd)

“Licensee” or “you” means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity’s behalf), of the age required under applicable laws, rules or regulations to provide legal consent and that has legal authority to bind your employer or such other person or entity if you are entering in this Agreement on their behalf.

“Meta Large Language Model Compiler” and “LLM Compiler” mean the foundational large language models and software and algorithms, including machine - learning model code, trained model weights, inference - enabling code, training - enabling code, fine - tuning enabling code and other elements of the foregoing distributed by Meta at:

• [https://huggingface.co/facebook/llm - compiler - 7b](https://huggingface.co/facebook/llm - compiler - 7b)
• [https://huggingface.co/facebook/llm - compiler - 7b - ftd](https://huggingface.co/facebook/llm - compiler - 7b - ftd)
• [https://huggingface.co/facebook/llm - compiler - 13b](https://huggingface.co/facebook/llm - compiler - 13b)
• [https://huggingface.co/facebook/llm - compiler - 13b - ftd](https://huggingface.co/facebook/llm - compiler - 13b - ftd)

“LLM Compiler Materials” means, collectively, Meta’s proprietary LLM Compiler and Documentation (and any portion thereof) made available under this Agreement.

“Meta” or “we” means Meta Platforms Ireland Limited (if you are located in or, if you are an entity, your principal place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland).

By clicking “I Accept” below or by using or distributing any portion or element of the LLM Compiler Materials, you agree to be bound by this Agreement.

1. License Rights and Redistribution a. Grant of Rights: You are granted a non - exclusive, worldwide, non - transferable and royalty - free limited license under Meta’s intellectual property or other rights owned by Meta embodied in the LLM Compiler Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the LLM Compiler Materials. b. Redistribution and Use i. If you distribute or make available the LLM Compiler Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such LLM Compiler Materials; and (B) prominently display “Built with LLM Compiler” on a related website, user interface, blogpost, about page, or product documentation. If you use the LLM Compiler Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include “LLM Compiler” at the beginning of any such AI model name. ii. If you receive LLM Compiler Materials, or any derivative works thereof, from a Licensee as part of an integrated end user product, then Section 2 of this Agreement will not apply to you. iii. You must retain in all copies of the LLM Compiler Materials that you distribute the following attribution notice within a “Notice” text file distributed as a part of such copies: “LLM Compiler is licensed under the LLM Compiler License, Copyright © Meta Platforms, Inc. All Rights Reserved.” iv. Your use of the LLM Compiler Materials must comply with applicable laws and regulations (including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for Llama Materials (available at https://llama.meta.com/llama3/use - policy), which is hereby incorporated by reference into this Agreement. v. You will not use the LLM Compiler Materials or any output or results of the LLM Compiler Materials to improve any other large language model.
2. Additional Commercial Terms: If, on the LLM Compiler release date, the monthly active users of the products or services made available by or for Licensee, or Licensee’s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.
3. Disclaimer of Warranty: UNLESS REQUIRED BY APPLICABLE LAW, THE LLM COMPILER MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN “AS IS” BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE, NON - INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLM COMPILER MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLM COMPILER MATERIALS AND ANY OUTPUT AND RESULTS.
4. Limitation of Liability: IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.
5. Intellectual Property a. No trademark licenses are granted under this Agreement, and in connection with the LLM Compiler Materials, neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates, except as required for reasonable and customary use in describing and redistributing the LLM Compiler Materials or as set forth in this Section 5(a). Meta hereby grants you a license to use LLM Compiler (the “Mark”) solely as required to comply with the last sentence of Section 1.b.i. You will comply with Meta’s brand guidelines (currently accessible at [https://about.meta.com/brand/resources/meta/company - brand/)](https://about.meta.com/brand/resources/meta/company - brand/). All goodwill arising out of your use of the Mark will inure to the benefit of Meta. b. Subject to Meta’s ownership of LLM Compiler Materials and derivatives made by or for Meta, with respect to any derivative works and modifications of the LLM Compiler Materials that are made by you, as between you and Meta, you are and will be the owner of such derivative works and modifications. c. If you institute litigation or other proceedings against Meta or any entity (including a cross - claim or counterclaim in a lawsuit) alleging that the LLM Compiler Materials or LLM Compiler outputs or results, or any portion of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third party arising out of or related to your use or distribution of the LLM Compiler Materials.
6. Term and Termination: The term of this Agreement will commence upon your acceptance of this Agreement or access to the LLM Compiler Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this Agreement. Upon termination of this Agreement, you shall delete and cease use of the LLM Compiler Materials. Sections 3, 4 and 7 shall survive the termination of this Agreement.
7. Governing Law and Jurisdiction: This Agreement will be governed and construed under the laws of the State of California without regard to choice of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of any dispute arising out of this Agreement.