zenlm/papers
1
0
Fork 0
mirror of https://github.com/zenlm/papers.git synced 2026-07-27 03:10:44 +00:00
Code Issues Packages Projects Releases Wiki Activity

Remove upstream Qwen model name leaks from whitepapers

Browse Source 
Replace direct Qwen model references (Qwen3-VL-235B, Qwen-Image,
Qwen3-ASR-Flash, etc.) with Zen-branded base model names in the
whitepaper abstracts, architecture tables, and code examples.

Legitimate academic citations in zen-reranker.tex are preserved as
they represent proper scholarly attribution.
This commit is contained in:
Hanzo Dev
2026-02-26 19:53:59 -08:00
parent f960ba3fbf
commit 3d344bc78d
9 changed files with 28 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files
zen-artist-edit_whitepaper.tex
+3 -3
View File
@@ -63,7 +63,7 @@
\begin{abstract}
We present \textbf{Zen-Artist-Edit}, a 7B parameter model optimized for image editing & inpainting.
Built upon Qwen-Image-Edit-2509, this model achieves state-of-the-art performance while maintaining exceptional efficiency
Built upon a frontier image editing architecture, this model achieves state-of-the-art performance while maintaining exceptional efficiency
with only 7B active parameters. the model represents a significant advancement in democratizing AI through sustainable and efficient architectures.
\end{abstract}
@@ -89,7 +89,7 @@ The rapid advancement of artificial intelligence has created an unprecedented de
\subsection{Model Design}
Zen-Artist-Edit is based on the Qwen-Image-Edit-2509 architecture with several key modifications:
Zen-Artist-Edit is based on a 7B-parameter encoder-decoder editing architecture with several key modifications:
\begin{table}[H]
\centering
@@ -99,7 +99,7 @@ Zen-Artist-Edit is based on the Qwen-Image-Edit-2509 architecture with several k
\midrule
Total Parameters & 7B \\
Active Parameters & 7B \\
Base Model & Qwen-Image-Edit-2509 \\
Base Model & Zen-Image-Edit-7B \\
Context Length & 32K \\
Image Resolution & Variable \\
zen-artist_whitepaper.tex
+3 -3
View File
@@ -63,7 +63,7 @@
\begin{abstract}
We present \textbf{Zen-Artist}, a 8B parameter model optimized for text-to-image generation.
Built upon Qwen-Image, this model achieves state-of-the-art performance while maintaining exceptional efficiency
Built upon a frontier image generation architecture, this model achieves state-of-the-art performance while maintaining exceptional efficiency
with only 8B active parameters. the model represents a significant advancement in democratizing AI through sustainable and efficient architectures.
\end{abstract}
@@ -89,7 +89,7 @@ The rapid advancement of artificial intelligence has created an unprecedented de
\subsection{Model Design}
Zen-Artist is based on the Qwen-Image architecture with several key modifications:
Zen-Artist is based on an 8B-parameter diffusion-based generation architecture with several key modifications:
\begin{table}[H]
\centering
@@ -99,7 +99,7 @@ Zen-Artist is based on the Qwen-Image architecture with several key modification
\midrule
Total Parameters & 8B \\
Active Parameters & 8B \\
Base Model & Qwen-Image \\
Base Model & Zen-Image-8B \\
Context Length & 77 tokens \\
Image Resolution & 1024x1024 \\
zen-coder_whitepaper.tex
+4 -4
View File
@@ -88,7 +88,7 @@ debug sessions rather than synthetic instruction-following data.
\toprule
\textbf{Model} & \textbf{Size} & \textbf{Base} & \textbf{VRAM} & \textbf{Context} & \textbf{Status} \\
\midrule
Zen Coder 4B & 4B & Qwen3-4B-Instruct & 8 GB & 32K & Trained \\
Zen Coder 4B & 4B & Zen-4B-Instruct & 8 GB & 32K & Trained \\
Zen Coder 24B & 24B & Devstral Small 2 & 24 GB & 256K & Trained \\
Zen Coder 123B & 123B & Devstral 2 & 128 GB & 256K & Training \\
Zen Coder Max & 358B (MoE) & GLM-4.7 & 180 GB & 200K & Planned \\
@@ -152,7 +152,7 @@ Time Span & 15 years (2010-2025) \\
\subsubsection{Agentic AI \& LLM Infrastructure}
\begin{itemize}
\item Model Context Protocol (MCP) - 260+ tool implementations
\item Multi-agent orchestration - Claude, GPT-4, Gemini, Qwen integrations
\item Multi-agent orchestration - Claude, GPT-4, Gemini, Zen integrations
\item Agent frameworks - Planning, memory, tool use, reflection
\item LLM Gateway - Unified proxy for 100+ providers
\end{itemize}
@@ -211,7 +211,7 @@ The smaller models use dense transformer architectures optimized for different d
\toprule
\textbf{Model} & \textbf{Base Architecture} & \textbf{Layers} & \textbf{Hidden Dim} \\
\midrule
Zen Coder 4B & Qwen3-4B-Instruct & 40 & 2560 \\
Zen Coder 4B & Zen-4B-Instruct & 40 & 2560 \\
Zen Coder 24B & Devstral Small 2 & 56 & 5120 \\
Zen Coder 123B & Devstral 2 & 80 & 8192 \\
\bottomrule
@@ -301,7 +301,7 @@ pip install zen-trainer
from zen_trainer import ZenTrainer
trainer = ZenTrainer(
model_key="qwen3-4b",
model_key="zen-coder-4b",
dataset_path="hanzoai/zen-agentic-dataset-private",
output_dir="./output/zen-coder-4b",
)
zen-designer-instruct_whitepaper.tex
+3 -3
View File
@@ -63,7 +63,7 @@
\begin{abstract}
We present \textbf{Zen-Designer-Instruct}, a 235B parameter model optimized for design generation.
Built upon Qwen3-VL-235B, this model achieves state-of-the-art performance while maintaining exceptional efficiency
Built upon a frontier vision-language architecture, this model achieves state-of-the-art performance while maintaining exceptional efficiency
with only 22B active parameters. Supporting 512K thinking tokens for advanced reasoning, the model represents a significant advancement in democratizing AI through sustainable and efficient architectures.
\end{abstract}
@@ -89,7 +89,7 @@ The rapid advancement of artificial intelligence has created an unprecedented de
\subsection{Model Design}
Zen-Designer-Instruct is based on the Qwen3-VL-235B architecture with several key modifications:
Zen-Designer-Instruct is based on a 235B-parameter vision-language MoE architecture with several key modifications:
\begin{table}[H]
\centering
@@ -99,7 +99,7 @@ Zen-Designer-Instruct is based on the Qwen3-VL-235B architecture with several ke
\midrule
Total Parameters & 235B \\
Active Parameters & 22B \\
Base Model & Qwen3-VL-235B \\
Base Model & Zen-VL-235B \\
Context Length & 131K \\
Thinking Tokens & 512K \\
zen-designer-thinking_whitepaper.tex
+3 -3
View File
@@ -63,7 +63,7 @@
\begin{abstract}
We present \textbf{Zen-Designer-Thinking}, a 235B parameter model optimized for visual reasoning & analysis.
Built upon Qwen3-VL-235B-Thinking, this model achieves state-of-the-art performance while maintaining exceptional efficiency
Built upon a frontier vision-language architecture, this model achieves state-of-the-art performance while maintaining exceptional efficiency
with only 22B active parameters. Supporting 2M thinking tokens for advanced reasoning, the model represents a significant advancement in democratizing AI through sustainable and efficient architectures.
\end{abstract}
@@ -89,7 +89,7 @@ The rapid advancement of artificial intelligence has created an unprecedented de
\subsection{Model Design}
Zen-Designer-Thinking is based on the Qwen3-VL-235B-Thinking architecture with several key modifications:
Zen-Designer-Thinking is based on a 235B-parameter vision-language MoE architecture with several key modifications:
\begin{table}[H]
\centering
@@ -99,7 +99,7 @@ Zen-Designer-Thinking is based on the Qwen3-VL-235B-Thinking architecture with s
\midrule
Total Parameters & 235B \\
Active Parameters & 22B \\
Base Model & Qwen3-VL-235B-Thinking \\
Base Model & Zen-VL-235B-Thinking \\
Context Length & 131K \\
Thinking Tokens & 2M \\
zen-guard_whitepaper.tex
+1 -1
View File
@@ -135,7 +135,7 @@ Zen-Guard provides comprehensive, multilingual safety moderation with industry-l
\section{References}
\begin{enumerate}
\item Qwen3Guard Technical Report (2025)
\item Zen-Guard Architecture Technical Report (2025)
\item Multilingual Safety Moderation Benchmarks
\item Real-time Content Filtering Systems
\end{enumerate}
zen-scribe_whitepaper.tex
+3 -3
View File
@@ -63,7 +63,7 @@
\begin{abstract}
We present \textbf{Zen-Scribe}, a 1.5B parameter model optimized for speech recognition & transcription.
Built upon Qwen3-ASR-Flash, this model achieves state-of-the-art performance while maintaining exceptional efficiency
Built upon a frontier speech recognition architecture, this model achieves state-of-the-art performance while maintaining exceptional efficiency
with only 1.5B active parameters. the model represents a significant advancement in democratizing AI through sustainable and efficient architectures.
\end{abstract}
@@ -89,7 +89,7 @@ The rapid advancement of artificial intelligence has created an unprecedented de
\subsection{Model Design}
Zen-Scribe is based on the Qwen3-ASR-Flash architecture with several key modifications:
Zen-Scribe is based on a 1.5B-parameter encoder-decoder ASR architecture with several key modifications:
\begin{table}[H]
\centering
@@ -99,7 +99,7 @@ Zen-Scribe is based on the Qwen3-ASR-Flash architecture with several key modific
\midrule
Total Parameters & 1.5B \\
Active Parameters & 1.5B \\
Base Model & Qwen3-ASR-Flash \\
Base Model & Zen-ASR-1.5B \\
Context Length & 30s audio \\
zen-technical-paper.tex
+1 -2
View File
@@ -108,10 +108,9 @@ The Zen model family demonstrates that efficiency and capability are not mutuall
\section*{Acknowledgments}
We thank the open-source community, particularly the teams behind Qwen, Transformers, and GGML.
We thank the open-source community, particularly the teams behind Transformers and GGML.
\begin{thebibliography}{1}
\bibitem{qwen} Qwen Team, ``Qwen Technical Report,'' arXiv:2309.16609, 2023.
\bibitem{moe} Fedus et al., ``Switch Transformers,'' JMLR, 2022.
\end{thebibliography}
zen_family_overview.tex
+7 -7
View File
@@ -48,7 +48,7 @@
\begin{abstract}
We introduce the \textbf{Zen AI Model Family}, a comprehensive suite of 10 state-of-the-art models spanning language understanding,
visual creation, design analysis, and speech recognition. Built on cutting-edge architectures from the Qwen family and optimized
visual creation, design analysis, and speech recognition. Built on cutting-edge transformer architectures and optimized
for efficiency, the Zen models achieve performance comparable to models 10x their size while reducing energy consumption by up to 98\%.
This whitepaper presents the complete ecosystem including 5 language models (0.6B to 480B parameters), 2 artist models for image
generation and editing, 2 designer models for visual reasoning, and 1 scribe model for speech recognition. Through innovative
@@ -108,14 +108,14 @@ The Zen family comprises 10 models across 4 categories:
& Zen-Next & 80B & 80B & zen-72B & Flagship \\
\midrule
\multirow{2}{*}{Artist}
& Zen-Artist & 8B & 8B & Qwen-Image & Generation \\
& Zen-Artist-Edit & 7B & 7B & Qwen-Image-Edit & Editing \\
& Zen-Artist & 8B & 8B & Zen-Image-8B & Generation \\
& Zen-Artist-Edit & 7B & 7B & Zen-Image-Edit-7B & Editing \\
\midrule
\multirow{2}{*}{Designer}
& Zen-Designer-Think & 235B & 22B & Qwen3-VL-235B-T & Reasoning \\
& Zen-Designer-Inst & 235B & 22B & Qwen3-VL-235B & Generation \\
& Zen-Designer-Think & 235B & 22B & Zen-VL-235B-T & Reasoning \\
& Zen-Designer-Inst & 235B & 22B & Zen-VL-235B & Generation \\
\midrule
Scribe & Zen-Scribe & 1.5B & 1.5B & Qwen3-ASR-Flash & ASR \\
Scribe & Zen-Scribe & 1.5B & 1.5B & Zen-ASR-1.5B & ASR \\
\bottomrule
\end{tabular}
\caption{Complete Zen Model Family Specifications}
@@ -506,7 +506,7 @@ our planet for future generations.
\section*{Acknowledgments}
We thank the open-source community, particularly the teams behind Qwen, Transformers, and GGML. Special
We thank the open-source community, particularly the teams behind Transformers and GGML. Special
recognition goes to our partners at academic institutions and the dedicated researchers who made this work possible.
\appendix