Open-Weight Models

Model · Image and text to text

DeepSeek-OCR-2

DeepSeek

Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.12.9 + CUDA11.8： Refer to GitHub for guidance on model inference acceleration and PDF processing, etc. We would like to thank DeepSeek-OCR, Vary, GOT-OCR2.0, MinerU, PaddleOCR for their valuable models and ideas. We also appreciate the benchmark OmniDocBench. author={Wei, Haoran and Sun, Yaofeng and Li, Yukun}, year={2025} title={DeepSeek-OCR 2: Visual Causal Flow}, author={Wei, Haoran and Sun, Yaofeng and Li, Yukun}, year={2026}

Uncensored version of Qwen/Qwen3.6-35B-A3B with refusal behavior removed via abliteration (norm-preserving orthogonalization). Zero refusals on harmful prompts. No false refusals on harmless prompts. Abliteration identifies the "refusal direction" in the model's residual stream — the linear direction that activates when the model decides to refuse — and surgically removes it from all output projection weights using norm-preserving orthogonalization. 1. Collect residual stream activations (last token position) for 512 harmful + 512 harmless prompts across all 40 layers 2. Compute mean difference vector per layer → this is the "refusal direction" candidate 3. Score layers by…

Model · Text generation

DeepSeek-R1-0528-Qwen3-8B

DeepSeek

The DeepSeek R1 model has undergone a minor version upgrade, with the current version being DeepSeek-R1-0528. In the latest update, DeepSeek R1 has significantly improved its depth of reasoning and inference capabilities by leveraging increased computational resources and introducing algorithmic optimization mechanisms during post-training. The model has demonstrated outstanding performance across various benchmark evaluations, including mathematics, programming, and general logic. Its overall performance is now approaching that of leading models, such as O3 and Gemini 2.5 Pro. Compared to the previous version, the upgraded model shows significant improvements in handling complex reasoning…

However, we observe that the speed and accuracy of YOLOs are negatively affected by the NMS. Recently, end-to-end Transformer-based detectors (DETRs) have provided an alternative to eliminating NMS. Nevertheless, the high computational cost limits their practicality and hinders them from fully exploiting the advantage of excluding NMS. In this paper, we propose the Real-Time DEtection TRansformer (RT-DETR), the first real-time end-to-end object detector to our best knowledge that addresses the above dilemma. We build RT-DETR in two steps, drawing on the advanced DETR: first we focus on maintaining accuracy while improving speed, followed by maintaining speed while improving accuracy.…

Liang Wang, Nan Yang, Xiaolong Huang, Binxing Jiao, Linjun Yang, Daxin Jiang, Rangan Majumder, Furu Wei, arXiv 2022 This model has 12 layers and the embedding size is 768. Below is an example to encode queries and passages from the MS-MARCO passage ranking dataset. Please refer to our paper at https://arxiv.org/pdf/2212.03533.pdf. Check out unilm/e5 to reproduce evaluation results on the BEIR and MTEB benchmark. Below is an example for usage with sentencetransformers. Package requirements pip install sentencetransformers~=2.2.2 1. Do I need to add the prefix "query: " and "passage: " to input texts? Yes, this is how the model is trained, otherwise you will see a performance degradation.…

Model · Fill mask

umberto-commoncrawl-cased-v1

Musixmatch

UmBERTo is a Roberta-based Language Model trained on large Italian Corpora and uses two innovative approaches: SentencePiece and Whole Word Masking. Now available at github.com/huggingface/transformers Marco Lodola, Monument to Umberto Eco, Alessandria 2019 UmBERTo-Commoncrawl-Cased utilizes the Italian subcorpus of OSCAR as training set of the language model. We used deduplicated version of the Italian corpus that consists in 70 GB of plain text data, 210M sentences with 11B words where the sentences have been filtered and shuffled at line level in order to be used for NLP research. This model was trained with SentencePiece and Whole Word Masking. These results refers to…

Model

jina-bert-flash-implementation

Jina AI

To run the model on GPU, you need to install Flash Attention. You may either install from pypi (which may not work with fused-dense), or from source. To install from source, clone the GitHub repository: The code provided here should work with commit 43950dd. Change to the cloned repo and install: This will compile the flash-attention kernel, which will take some time. If you would like to use fused MLPs (e.g. to use activation checkpointing), you may install fused-dense also from source: The config adds some new parameters: - useflashattn: If True, always use flash attention. If None, use flash attention when GPU is available. If False, never use flash attention (works on CPU).…

Model · Text generation

DeepSeek-Coder-V2-Lite-Instruct

DeepSeek

We present DeepSeek-Coder-V2, an open-source Mixture-of-Experts (MoE) code language model that achieves performance comparable to GPT4-Turbo in code-specific tasks. Specifically, DeepSeek-Coder-V2 is further pre-trained from an intermediate checkpoint of DeepSeek-V2 with additional 6 trillion tokens. Through this continued pre-training, DeepSeek-Coder-V2 substantially enhances the coding and mathematical reasoning capabilities of DeepSeek-V2, while maintaining comparable performance in general language tasks. Compared to DeepSeek-Coder-33B, DeepSeek-Coder-V2 demonstrates significant advancements in various aspects of code-related tasks, as well as reasoning and general capabilities.…

Model · Image and text to text

XYZAILab_XYZ-Aquila-mini-GGUF

Bartowski

Using llama.cpp release b10142 for quantization. All quants made using imatrix option with dataset from here Run them in your choice of tools: Note: if it's a newly supported model, you may need to wait for an update from the developers. Some of these quants (Q3KXL, Q4KL etc) are the standard quantization method with the embeddings and output weights quantized to Q80 instead of what they would normally default to. First, make sure you have huggingface-cli installed: Then, you can target the specific file you want: If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run: You can either specify a new local-dir…

Model · Text generation

Qwen2.5-1.5B

Qwen

Qwen2.5 is the latest series of Qwen large language models. For Qwen2.5, we release a number of base language models and instruction-tuned language models ranging from 0.5 to 72 billion parameters. Qwen2.5 brings the following improvements upon Qwen2: - Significantly more knowledge and has greatly improved capabilities in coding and mathematics, thanks to our specialized expert models in these domains. - Significant improvements in instruction following, generating long texts (over 8K tokens), understanding structured data (e.g, tables), and generating structured outputs especially JSON. More resilient to the diversity of system prompts, enhancing role-play implementation and…

Model

vae-text-encorder-for-flux-klein-9b

Comfy Org

Repackaged model files for ComfyUI. Place the files in the following folders

Model · Sentence similarity

snowflake-arctic-embed-l-v2.0

Snowflake

12/11/2024: Release of Technical Report - 12/04/2024: Release of snowflake-arctic-embed-l-v2.0 and snowflake-arctic-embed-m-v2.0 our newest models with multilingual workloads in mind. Snowflake arctic-embed-l-v2.0 is the newest addition to the suite of embedding models Snowflake has released optimizing for retrieval performance and inference efficiency. Arctic Embed 2.0 introduces a new standard for multilingual embedding models, combining high-quality multilingual text retrieval without sacrificing performance in English. Released under the permissive Apache 2.0 license, Arctic Embed 2.0 is ideal for applications that demand reliable, enterprise-grade multilingual search and retrieval at…

Model · Text generation

Qwen2.5-Coder-7B-Instruct-AWQ

Qwen

Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen). As of now, Qwen2.5-Coder has covered six mainstream model sizes, 0.5, 1.5, 3, 7, 14, 32 billion parameters, to meet the needs of different developers. Qwen2.5-Coder brings the following improvements upon CodeQwen1.5: - Significantly improvements in code generation, code reasoning and code fixing. Base on the strong Qwen2.5, we scale up the training tokens into 5.5 trillion including source code, text-code grounding, Synthetic data, etc. Qwen2.5-Coder-32B has become the current state-of-the-art open-source codeLLM, with its coding abilities matching those of GPT-4o. - A more…

Model · Text generation

Qwen3-0.6B-Base

Qwen

Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Building upon extensive advancements in training data, model architecture, and optimization techniques, Qwen3 delivers the following key improvements over the previously released Qwen2.5: Qwen3-0.6B-Base has the following features: For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our blog, GitHub, and Documentation. The code of Qwen3 has been in the latest Hugging Face transformers and we advise you to use the latest version of transformers. With transformers<4.51.0, you will…

Model · Text to image

Juggernaut-XL-v9

RunDiffusion

The SDXL ecosystem is the single most mature corner of open image generation, and v9 is its most refined photorealism checkpoint. Choose Juggernaut XL v9 when you want: - Photorealism that holds up under scrutiny — skin texture, micro-contrast, and natural lighting that translates from concept to print. - Reasonable hardware — runs comfortably on 8 GB of VRAM, unlike newer DiT-based models that demand 16+ GB. - The full SDXL toolbox — drop-in compatibility with the thousands of SDXL ControlNets, IP-Adapter variants, AnimateDiff, regional prompting tools, and LoRAs already in your workflow. - Battle-tested reliability — 26+ months in production, used in agencies, studios, and shipping…

Model · Text classification

deberta-v3-base-prompt-injection-v2

Protect AI

This model is a fine-tuned version of microsoft/deberta-v3-base specifically developed to detect and classify prompt injection attacks which can manipulate language models into producing unintended outputs. Prompt injection attacks manipulate language models by inserting or altering prompts to trigger harmful or unintended responses. The deberta-v3-base-prompt-injection-v2 model is designed to enhance security in language model applications by detecting these malicious interventions. This model classifies inputs into benign (0) and injection-detected (1). deberta-v3-base-prompt-injection-v2 is highly accurate in identifying prompt injections in English. It does not detect jailbreak attacks…

Model · Image and text to text

Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF

David Belton

in both 8 bit and 4 bit. This repo contains both "regular" and "MTP" Neo MAX Imatrix GGUF quants. Many other additional quant types avail too. 3rd parties confirm this model's performance in the "community tab". 40B versions: Eleanor-DECKARD and Grand Intelligence - FF711-717 || Qwen 3.8 27B Cold Fusion (1/2 to 1/10 thinking size, more brainpower): COLD FUSION Meet the newest, strongest and fastest Qwen 3.8: The TURBO Fable 738-882 The strongest, smartest open source multi-stage model fine tune for consumer hardware ever and BUILT on consumer hardware via Unsloth. The first model of this size/type to breach "700" ARC-C in both 8 bit and 4 bit; hench the "711" in the name. This model (both 4…

Model · Image text to video

MiniMax-H3-GGUF

Unsloth AI

Instructions further below. GGUF for MiniMax-H3, compatible on most platforms including stablediffusion.cpp and Unsloth. You can run MiniMax-H3 via Unsloth: https://github.com/unslothai/unsloth/ GGUF quantizations of MiniMaxAI/MiniMax-H3 MiniMax H3 is an omni-modal generative system that produces video with native stereo audio, up to 15 seconds at 24 FPS with 32 kHz stereo audio. Both halves of the runtime are in this repo: the denoisers and the Qwen3-VL text encoder they need. H3 ships two denoisers, and which one you load decides what the model can be given: - fl2vapruned, the H3-Base first-and-last-frame variant. Text, plus zero, one or two frames. - ref2vapruned, the reference variant.…

Model · Any to any

gemma-4-12B-it-qat-w4a16-ct

Google

Gemma is a family of open models built by Google DeepMind. Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output. This release includes open-weights models in both pre-trained and instruction-tuned variants. Gemma 4 features a context window of up to 256K tokens and maintains multilingual support in over 140 languages. Featuring both Dense and Mixture-of-Experts (MoE) architectures, Gemma 4 is well-suited for tasks like text generation, coding, and reasoning. The models are available in five distinct sizes: E2B, E4B, 12B, 26B A4B, and 31B. Their diverse sizes make them deployable in environments ranging from…

Model · Translation

opus-mt-fr-en

Helsinki-NLP Research Group

source languages: fr; target languages: en; OPUS readme: fr-en; dataset: opus; model: transformer-align; pre-processing: normalization + SentencePiece.

Model · Text generation

GLM-5.2-NVFP4

NVIDIA

The NVIDIA GLM-5.2 NVFP4 model is the quantized version of ZAI’s GLM-5.2 model, which is an auto-regressive language model that uses an optimized transformer architecture. GLM-5.2 is a Mixture-of-Experts (MoE) model for reasoning and coding that uses sparse attention (with an IndexShare indexer) to support a long context. For more information, please check here. The NVIDIA GLM-5.2 NVFP4 model is quantized with Model Optimizer. This model is ready for commercial or non-commercial use. GOVERNING TERMS: Use of the model is governed by the MIT License, same as the base model. Global Developers looking to take off-the-shelf, pre-quantized models for deployment in AI Agent systems, chatbots, RAG…

Model · Zero-shot classification

mDeBERTa-v3-base-xnli-multilingual-nli-2mil7

Moritz Borrett-Laurer (formerly Laurer)

This multilingual model can perform natural language inference (NLI) on 100 languages and is therefore also suitable for multilingual zero-shot classification. The underlying mDeBERTa-v3-base model was pre-trained by Microsoft on the CC100 multilingual dataset with 100 languages. The model was then fine-tuned on the XNLI dataset and on the multilingual-NLI-26lang-2mil7 dataset. Both datasets contain more than 2.7 million hypothesis-premise pairs in 27 languages spoken by more than 4 billion people. As of December 2021, mDeBERTa-v3-base is the best performing multilingual base-sized transformer model introduced by Microsoft in this paper. This model was trained on the…

Model · Image and text to text

Cosmos-Reason2-2B

NVIDIA

NVIDIA Cosmos Reason 2 is an open, customizable, 2B-parameter reasoning vision language model (VLM) for physical AI and robotics that enables robots and vision AI agents to reason like humans, using prior knowledge, physics understanding and common sense to understand and act in the real world. This model understands space, time, and fundamental physics, and can serve as a planning model to reason what steps an embodied agent might take next. New features with Cosmos Reason 2: Enhanced physical AI reasoning with improved spatio-temporal understanding and timestamp precision. Supports object detection with 2D/3D point localization and bounding box coordinates with reasoning explanations and…

Model · Image and text to text

Qwen3.8-27B-NVFP4-RTX5090

Gittensor Model Hub

Runs on SparkInfer, SGLang and vLLM unmodified — configs for all three are below. Serving many users at once? See concurrency. SparkInfer × this NVFP4 build × the DSpark v2 drafter — an engine, a checkpoint, and a speculative drafter optimized against each other, compounding to 4.3×. The drafter never changes what the model says: the target verifies every drafted token. GeForce RTX 5090–specific NVFP4 checkpoint of Qwen/Qwen3.8-27B, quantized with NVIDIA Model Optimizer. Serves the full native 262,144-token context on 32 GB. With the DSpark v2 drafter: 264.8 tok/s overall — up to 420 on code — on SparkInfer (its bench harness; the HTTP server is autoregressive-only today) and 161.7 tok/s on…