全站搜索
AI 模型 84760
- qwen3-next-80b-a3b-thinking↗
阿里巴巴 · Qwen
Over the past few months, we have observed increasingly clear trends toward scaling both total parameters and context lengths in the pursuit of more powerful and agentic artificial intelligence (AI). We are excited to share our latest advancements in addressing these demands, centered on improving scaling efficiency through innovative model architecture. We call this next-generation foundation models **Qwen3-Next**. **Qwen3-Next-80B-A3B** is the first installment in the Qwen3-Next series and features the following key enchancements: - **Hybrid Attention**: Replaces standard attention with the combination of **Gated DeltaNet** and **Gated Attention**, enabling efficient context modeling for ultra-long context length. - **High-Sparsity Mixture-of-Experts (MoE)**: Achieves an extreme low activation ratio in MoE layers, drastically reducing FLOPs per token while preserving model capacity. - **Stability Optimizations**: Includes techniques such as **zero-centered and weight-decayed layernorm**, and other stabilizing enhancements for robust pre-training and post-training. - **Multi-Token Prediction (MTP)**: Boosts pretraining model performance and accelerates inference. We are seeing strong performance in terms of both parameter efficiency and inference speed for Qwen3-Next-80B-A3B: - Qwen3-Next-80B-A3B-Base outperforms Qwen3-32B-Base on downstream tasks with 10% of the total training cost and with 10 times inference throughput for context over 32K tokens. - Leveraging GSPO, we have addressed the stability and efficiency challenges posed by the hybrid attention mechanism combined with a high-sparsity MoE architecture in RL training. Qwen3-Next-80B-A3B-Thinking demonstrates outstanding performance on complex reasoning tasks, not only **surpassing Qwen3-30B-A3B-Thinking-2507 and Qwen3-32B-Thinking**, but also **outperforming the proprietary model Gemini-2.5-Flash-Thinking** across multiple benchmarks.
- DeepSeek-V4.1-Flash-Quark-MXFP4↗
amd
- MaskGCT↗
amphion
- Race-Classification-FairFace-YOLOv8↗
Anzhc
This model is based on FairFace 0.25 padding variant dataset composed by Microsoft researchers, aiming to reduce bias by better balancing classes in dataset. > **Karkkainen, Kimmo, and Joo, Jungseock.** > *FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age for Bias Measurement and Mitigation.* > Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 1548–1558. > FairFace Dataset on GitHub This YOLOv8 training is meant only for race classification. I wanted a really really fast model for tagging, and this is likely what it's useful for! ~~I will provide a pipeline for running it on your datasets in future.~~ I've made simple scripts for you to use on your data. By default it will output .txt files(or append to existing), so modify for your specific needs: https://github.com/Anzhc/Simple-Utility-Scripts-for-YOLO/tree/main
- APUS-Reflex-35B-A3B-GGUF↗
apus-ailab
> **Branding note:** This model family is now **APUS Reflex**. Some weight filenames, script/module names, served-model aliases, and the separate evaluation dataset ID retain their historical identifiers for download and runtime compatibility; these do not indicate different model versions. GGUF conversions of APUS-Reflex-35B-A3B (revision `494738529137`) for **Ollama, llama.cpp and LM Studio** on Linux, Windows and macOS (Metal). APUS Reflex is a **decision model**: each request supplies a state, an instruction and 2–16 candidates; the model answers with one candidate label (A–P) and the application reads the distribution over those labels. It is not a chat model.
- WeSpeaker-ResNet34-LM-CoreML↗
aufklarer
- google_gemma-3-270m-it-GGUF↗
bartowski
- openai_gpt-oss-20b-GGUF↗
bartowski
All quants keep the feed forward networks at mxfp4 for optimal performance, which does mean the size differences are negligible unfortunately, but being provided just because. No chat template specified so default is used. This may be incorrect, check original model card for details. The reason is, the FFN (feed forward networks) of gpt-oss do not behave nicely when quantized to anything other than MXFP4, so they are kept at that level for everything.
- beetle-bilingual-l2-50-simultaneous-b2-fineweb-2b-jpn-eng↗
Beetle-FineWeb-24B-4
- Bespoke-Nimble-9B-v2↗
bespokelabs
