全站搜索
AI 模型 85021
- SGLang-EAGLE3-Qwen3-30B-A3B-Instruct-2507-SpecForge-Nex↗
lmsys
**SpecBundle** is an open-source initiative, jointly driven by the community and industry, to democratize speculative decoding by providing high-performance speculative decoding draft weights for mainstream open-source models. This checkpoint was trained by the **Nex-AGI** Team and released as the **phase 1 of SpecBundle release**. We regenerated the responses in the mlabonne/open-perfectblend and trained the model on 1.4M data samples for 2 epochs. This checkpoint was trained using the **SpecForge** framework. We sincerely appreciate the collective efforts from both the developers in the open-source community and our industrial partners, especially Ant Group AQ Team, Meituan, Nex-AGI (Qiji Zhifeng), EigenAI for their invaluable contributions to the release of SpecBundle Phase 1.
- Lythri-7B-A4B-GGUF↗
Lythri
This repository contains GGUF quantizations of Lythri-7B-A4B for use with llama.cpp and compatible apps. For model details, evaluation results and limitations, see the original model card. > **Note:** Lythri models are highly sensitive to quantization. We recommend using the highest-precision quantization your hardware allows. Lower-bit quantizations still work but are not recommended. To use a specific quantization, append its type to the repository name, for example `Lythri/Lythri-7B-A4B-GGUF:Q8_0`.
- LeoScorpius-7B-Chat-DPO-Mistral-7B-Instruct-v0.2-slerp-GGUF↗
MaziyarPanahi
- Llama-3.1-Nemotron-70B-Instruct-HF-GGUF↗
MaziyarPanahi
- Phi-3-Context-Obedient-RAG-GGUF↗
MaziyarPanahi
- qwen-image-2-1-controlnet-union-mflux-q4↗
mflux-community
**MFlux** runs the latest state-of-the-art generative image models locally on your Mac in native MLX. MFlux is open source and free. > [!NOTE] > Community conversion by MFlux-Community -- not an official release from the original model's authors. The original model's license applies unchanged, see License. Standard MFlux quantization uses MLX affine quantization (4-bit) across the text encoder, transformer and VAE. The quantization level is stored in the weights, so no `--quantize` flag is needed.
- qwen-image-2-1-turbo-mflux-q6↗
mflux-community
- beit-large-patch16-384↗
Microsoft
- LLM2CLIP-EVA02-B-16↗
Microsoft
- MiniMax-M1-40k↗
MiniMaxAI
We introduce MiniMax-M1, the world's first open-weight, large-scale hybrid-attention reasoning model. MiniMax-M1 is powered by a hybrid Mixture-of-Experts (MoE) architecture combined with a lightning attention mechanism. The model is developed based on our previous MiniMax-Text-01 model, which contains a total of 456 billion parameters with 45.9 billion parameters activated per token. Consistent with MiniMax-Text-01, the M1 model natively supports a context length of 1 million tokens, 8x the context size of DeepSeek R1. Furthermore, the lightning attention mechanism in MiniMax-M1 enables efficient scaling of test-time compute – For example, compared to DeepSeek R1, M1 consumes 25% of the FLOPs at a generation length of 100K tokens. These properties make M1 particularly suitable for complex tasks that require processing long inputs and thinking extensively. MiniMax-M1 is trained using large-scale reinforcement learning (RL) on diverse problems ranging from traditional mathematical reasoning to sandbox-based, real-world software engineering environments. We develop an efficient RL scaling framework for M1 highlighting two perspectives: (1) We propose CISPO, a novel algorithm that clips importance sampling weights instead of token updates, which outperforms other competitive RL variants; (2) Our hybrid-attention design naturally enhances the efficiency of RL, where we address unique challenges when scaling RL with the hybrid architecture. We train two versions of MiniMax-M1 models with 40K and 80K thinking budgets respectively. Experiments on standard benchmarks show that our models outperform other strong open-weight models such as the original DeepSeek-R1 and Qwen3-235B, particularly on complex software engineering, tool using, and long context tasks. With efficient scaling of test-time compute, MiniMax-M1 serves as a strong foundation for next-generation language model agents to reason and tackle real-world challenges. Benchmark performance comparison of leading commercial and open-weight models across competition-level mathematics, coding, software engineering, agentic tool use, and long-context understanding tasks. We use the MiniMax-M1-80k model here for MiniMax-M1. We report results derived from the Agentless scaffold. Departing from the original pipeline, our methodology employs a two-stage localization process (without any embedding-based retrieval mec
