papers/gym-training-platform/sections/99-acknowledgments.tex

\section*{Acknowledgments}


Gym builds on the pioneering work of the Hugging Face Transformers team, the original LLaMA Factory contributors, and the broader open-source AI community. We thank the 500+ contributors who have submitted code, documentation, and bug reports. Special thanks to the educational institutions who pilot-tested Gym and provided invaluable feedback. This work is supported by Zoo Labs Foundation Inc, a 501(c)(3) non-profit organization.

\bibliographystyle{plainnat}
\begin{thebibliography}{99}

\bibitem[Brown et~al.(2020)]{brown2020gpt3}
Tom~B. Brown, Benjamin Mann, Nick Ryder, et~al.
\newblock Language models are few-shot learners.
\newblock \emph{NeurIPS}, 2020.

\bibitem[Touvron et~al.(2023)]{touvron2023llama}
Hugo Touvron, Thibaut Lavril, Gautier Izacard, et~al.
\newblock LLaMA: Open and efficient foundation language models.
\newblock \emph{arXiv:2302.13971}, 2023.

\bibitem[Hu et~al.(2021)]{hu2021lora}
Edward~J. Hu, Yelong Shen, Phillip Wallis, et~al.
\newblock LoRA: Low-rank adaptation of large language models.
\newblock \emph{ICLR}, 2022.

\bibitem[Dettmers et~al.(2023)]{dettmers2023qlora}
Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer.
\newblock QLoRA: Efficient finetuning of quantized LLMs.
\newblock \emph{NeurIPS}, 2023.

\bibitem[Liu et~al.(2024)]{liu2024dora}
Shih-Yang Liu, Chien-Yi Wang, Hongxu Yin, et~al.
\newblock DoRA: Weight-decomposed low-rank adaptation.
\newblock \emph{arXiv:2402.09353}, 2024.

\bibitem[Meng et~al.(2024)]{meng2024pissa}
Fanxu Meng, Zhaohui Wang, Muhan Zhang.
\newblock PiSSA: Principal singular values and singular vectors adaptation of large language models.
\newblock \emph{arXiv:2404.02948}, 2024.

\bibitem[Shao et~al.(2024)]{shao2024deepseekmath}
Zhihong Shao, Peiyi Wang, Qihao Zhu, et~al.
\newblock DeepSeekMath: Pushing the limits of mathematical reasoning in open language models.
\newblock \emph{arXiv:2402.03300}, 2024.

\bibitem[Tencent(2025)]{tencent2025trainingfree}
Tencent youtu-agent Team.
\newblock Training-free GRPO: Context-based policy optimization for large language models.
\newblock \emph{arXiv:2510.08191}, 2025.

\bibitem[Liu et~al.(2023)]{liu2023llava}
Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong~Jae Lee.
\newblock Visual instruction tuning.
\newblock \emph{NeurIPS}, 2023.

\bibitem[Bai et~al.(2023)]{bai2023qwenvl}
Jinze Bai, Shuai Bai, Shusheng Yang, et~al.
\newblock Qwen-VL: A versatile vision-language model for understanding, localization, text reading, and beyond.
\newblock \emph{arXiv:2308.12966}, 2023.

\bibitem[Chu et~al.(2024)]{chu2024qwen2audio}
Yunfei Chu, Jin Xu, Qian Yang, et~al.
\newblock Qwen2-Audio: Technical report.
\newblock \emph{arXiv:2407.10759}, 2024.

\bibitem[Qwen(2025)]{qwen2025qwen3}
Qwen Team.
\newblock Qwen3 technical report.
\newblock \emph{Alibaba Cloud}, 2025.

\bibitem[Dao et~al.(2023)]{dao2023flashattention}
Tri Dao, Daniel~Y. Fu, Stefano Ermon, et~al.
\newblock FlashAttention-2: Faster attention with better parallelism and work partitioning.
\newblock \emph{arXiv:2307.08691}, 2023.

\bibitem[Hsu et~al.(2024)]{hsu2024liger}
Pin-Lun Hsu, Yun-Da Tsai, Shou-De Lin.
\newblock Liger Kernel: Efficient Triton kernels for LLM training.
\newblock \emph{GitHub}, 2024.

\bibitem[Unsloth(2024)]{unsloth2024}
Unsloth AI.
\newblock Unsloth: 2x faster, 60\% less memory LLM finetuning.
\newblock \emph{GitHub}, 2024.

\bibitem[Rafailov et~al.(2023)]{rafailov2023dpo}
Rafael Rafailov, Archit Sharma, Eric Mitchell, et~al.
\newblock Direct preference optimization: Your language model is secretly a reward model.
\newblock \emph{NeurIPS}, 2023.

\bibitem[Ethayarajh et~al.(2024)]{ethayarajh2024kto}
Kawin Ethayarajh, Winnie Xu, Niklas Muennighoff, et~al.
\newblock KTO: Model alignment as prospect theoretic optimization.
\newblock \emph{arXiv:2402.01306}, 2024.

\bibitem[Hong et~al.(2024)]{hong2024orpo}
Jiwoo Hong, Noah Lee, James Thorne.
\newblock ORPO: Monolithic preference optimization without reference model.
\newblock \emph{arXiv:2403.07691}, 2024.

\bibitem[Schulman et~al.(2017)]{schulman2017ppo}
John Schulman, Filip Wolski, Prafulla Dhariwal, et~al.
\newblock Proximal policy optimization algorithms.
\newblock \emph{arXiv:1707.06347}, 2017.

\bibitem[Kalajdzievski(2023)]{kalajdzievski2023rslora}
Damjan Kalajdzievski.
\newblock A rank stabilization scaling factor for fine-tuning with LoRA.
\newblock \emph{arXiv:2312.03732}, 2023.

\bibitem[Chen et~al.(2023)]{chen2023longlora}
Yukang Chen, Shengju Qian, Haotian Tang, et~al.
\newblock LongLoRA: Efficient fine-tuning of long-context large language models.
\newblock \emph{arXiv:2309.12307}, 2023.

\bibitem[Frantar et~al.(2023)]{frantar2023gptq}
Elias Frantar, Saleh Ashkboos, Torsten Hoefler, Dan Alistarh.
\newblock GPTQ: Accurate post-training quantization for generative pre-trained transformers.
\newblock \emph{ICLR}, 2023.

\bibitem[Lin et~al.(2023)]{lin2023awq}
Ji Lin, Jiaming Tang, Haotian Tang, et~al.
\newblock AWQ: Activation-aware weight quantization for LLM compression and acceleration.
\newblock \emph{arXiv:2306.00978}, 2023.

\bibitem[Xiao et~al.(2023)]{xiao2023smoothquant}
Guangxuan Xiao, Ji Lin, Mickael Seznec, et~al.
\newblock SmoothQuant: Accurate and efficient post-training quantization for large language models.
\newblock \emph{ICML}, 2023.

\bibitem[Wang et~al.(2024)]{wang2024qwen2vl}
Peng Wang, Shuai Bai, Sinan Tan, et~al.
\newblock Qwen2-VL: Enhancing vision-language model's perception of the world at any resolution.
\newblock \emph{arXiv:2409.12191}, 2024.

\bibitem[Zhao et~al.(2023)]{zhao2023pytorch}
Yanli Zhao, Andrew Gu, Rohan Varma, et~al.
\newblock PyTorch FSDP: Experiences on scaling fully sharded data parallel.
\newblock \emph{arXiv:2304.11277}, 2023.

\bibitem[Rasley et~al.(2020)]{rasley2020deepspeed}
Jeff Rasley, Samyam Rajbhandari, Olatunji Ruwase, Yuxiong He.
\newblock DeepSpeed: System optimizations enable training deep learning models with over 100 billion parameters.
\newblock \emph{KDD}, 2020.

\bibitem[Axolotl(2024)]{axolotl2024}
Axolotl Contributors.
\newblock Axolotl: A streamlined framework for fine-tuning large language models.
\newblock \emph{GitHub}, 2024.

\bibitem[Meta(2024)]{meta2024llamarecipes}
Meta AI.
\newblock llama-recipes: Scripts to fine-tune Meta Llama models.
\newblock \emph{GitHub}, 2024.

\bibitem[von Werra et~al.(2022)]{vonwerra2022trl}
Leandro von Werra, Younes Belkada, Lewis Tunstall, et~al.
\newblock TRL: Transformer reinforcement learning.
\newblock \emph{GitHub}, 2022.

\bibitem[Molino et~al.(2019)]{molino2019ludwig}
Piero Molino, Yaroslav Dudin, Sai~Sumanth Miryala.
\newblock Ludwig: A type-based declarative deep learning toolbox.
\newblock \emph{arXiv:1909.07930}, 2019.

\bibitem[Zoo Labs(2025)]{zoo-per-llm-chains}
Zoo Labs Foundation.
\newblock Per-LLM Chains: Unified Training, Research, and Pooled GPU Hosting on Quasar 3.0.
\newblock \emph{Zoo Whitepaper}, December 2025.

\bibitem[Lux LP-020(2025)]{lp-020-quasar-3}
Lux Foundation.
\newblock LP-020: Quasar Consensus 3.0.
\newblock \emph{Lux Improvement Proposal}, December 2025.

\bibitem[Lux LP-132(2025)]{lp-132-quasargpu}
Lux Foundation.
\newblock LP-132: QuasarGPU Adapter.
\newblock \emph{Lux Improvement Proposal}, December 2025.

\bibitem[Lux LP-133(2025)]{lp-133-app-stack}
Lux Foundation.
\newblock LP-133: Quasar-Native App Stack.
\newblock \emph{Lux Improvement Proposal}, December 2025.

\bibitem[Lux LP-134(2025)]{lp-134-topology}
Lux Foundation.
\newblock LP-134: Lux Chain Topology.
\newblock \emph{Lux Improvement Proposal}, December 2025.

\bibitem[Lux LP-136(2025)]{lp-136-cloud}
Lux Foundation.
\newblock LP-136: Lux Cloud Capacity Federation.
\newblock \emph{Lux Improvement Proposal}, December 2025.

\bibitem[Hanzo AI(2025)]{hanzo-ai-chain}
Hanzo AI.
\newblock The Hanzo AI Chain: Inference Marketplace and Commerce Layer.
\newblock \emph{Hanzo Whitepaper}, December 2025.

\end{thebibliography}