NOTICE

This product includes software from the following projects:

1. FlashAttention (https://github.com/Dao-AILab/flash-attention)
Licensed under BSD 3-Clause License
Copyright (c) 2022, the respective contributors, as shown by the AUTHORS file.

The FlashAttention implementation provides fast and memory-efficient exact attention with IO-awareness. 
The original work was published in the following papers:
- "FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness"
  Authors: Tri Dao, Daniel Y. Fu, Stefano Ermon, Atri Rudra, Christopher Ré
  https://arxiv.org/abs/2205.14135

- "FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning"
  Author: Tri Dao
  https://arxiv.org/abs/2307.08691

2. AWQ (https://github.com/mit-han-lab/llm-awq)
Licensed under MIT License
Copyright (c) 2023 MIT HAN Lab

AWQ provides activation-aware weight quantization for LLM compression and acceleration.
The original work was published in the following paper:
- "AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration"
  Authors: Ji Lin, Jiaming Tang, Haotian Tang, Shang Yang, Xingyu Dang, Song Han
  https://arxiv.org/abs/2306.00978

3. ExLlamaV2 (https://github.com/turboderp/exllamav2)
Licensed under MIT License
Copyright (c) 2023 Thomas Randall

ExLlamaV2 is an inference library for running local LLMs on modern consumer GPUs.

For the complete license texts, please refer to:
- FlashAttention: https://github.com/Dao-AILab/flash-attention/blob/main/LICENSE
- AWQ: https://github.com/mit-han-lab/llm-awq/blob/main/LICENSE
- ExLlamaV2: https://github.com/turboderp/exllamav2/blob/master/LICENSE

Users of this software should note that the BSD 3-Clause License and MIT License require preservation 
of copyright notices and acknowledgment of the use of the original software in documentation and/or 
other materials provided with the distribution.