Search FoxAI

AI Models 2119

• Qwen3.5-122B-A10B-abliterix-NVFP4↗

  bjk110

  NVFP4 (W4A4) quantization of `wangzhang/Qwen3.5-122B-A10B-abliterix` — the 122B/A10B Abliterix-abliterated (uncensored) Qwen3.5 MoE — packaged for vLLM serving on NVIDIA Blackwell hardware (DGX Spark GB10 / SM121). The Abliterix-trained uncensored behavior (0.5% refusal rate, KL divergence 0.0115 vs the Qwen3.5-122B-A10B baseline) is preserved while dropping weight memory from BF16 (~230 GB) to NVFP4 (~70 GB) — fitting comfortably on two DGX Spark nodes (2 × 119 GiB unified memory) with headroom for KV-cache, with the same compressed-tensors runtime path as the FP8 sibling. Compared to the FP8 W8A8 variant (bjk110/Qwen3.5-122B-A10B-abliterix-FP8, 116 GB), this NVFP4 build cuts another ~40% off-disk and unlocks the **FlashInfer-CUTLASS NVFP4 MoE backend** for higher throughput on Blackwell.

• gliclass-base-v3.0↗

  knowledgator

  This is an efficient zero-shot classifier inspired by GLiNER work. It demonstrates the same performance as a cross-encoder while being more compute-efficient because classification is done at a single forward path. It can be used for `topic classification`, `sentiment analysis`, and as a reranker in `RAG` pipelines. The model was trained on logical tasks to induce reasoning. LoRa adapters were used to fine-tune the model without destroying the previous knowledge.

• stable-diffusion-xl-base-1.0↗

  Stability AI

  SDXL consists of an ensemble of experts pipeline for latent diffusion: In a first step, the base model is used to generate (noisy) latents, which are then further processed with a refinement model (available here: https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/) specialized for the final denoising steps. Note that the base model can be used as a standalone module. Alternatively, we can use a two-stage pipeline as follows: First, the base model is used to generate latents of the desired output size. In the second step, we use a specialized high-resolution model and apply a technique called SDEdit (https://arxiv.org/abs/2108.01073, also known as "img2img") to the latents generated in the first step, using the same prompt. This technique is slightly slower than the first one, as it requires more function evaluations. For research purposes, we recommend our `generative-models` Github repository (https://github.com/Stability-AI/generative-models), which implements the most popular diffusion frameworks (both training and inference) and for which new functionalities like distillation will be added over time. Clipdrop provides free SDXL inference.

• flux-3-image-klein↗

  unsloth

  > Ungated mirror of black-forest-labs/flux-3-image-klein, republished by Unsloth so it > can be downloaded without a Hub gate. The weights are unmodified. See NOTICE.md and LICENSE.md > in this repo for the terms, which are the upstream terms and are unchanged. FLUX 3 Image Klein is a 5.6B open weights model for text-to-image generation and image editing. It takes a prompt and up to 10 reference images and returns one image. Text-to-image, single-reference edits and multi-reference compositions run through the same model, with Qwen3-VL-4B-Instruct as the text encoder. This repository holds the distilled checkpoint for generation and image editing. For fine-tuning, LoRA training and research, use FLUX 3 Image Klein Base.

• WizardMath-7B-V1.1↗

  WizardLMTeam

  Before model training, we carefully and rigorously checked all the training data, and used multiple deduplication methods to verify and prevent data leakage on GSM8k and MATH test set. Please use **the same systems prompts strictly** with us, and we do not guarantee the accuracy of the **quantified versions**.

• bert-base-greek-uncased-v1↗

  nlpaueb

  * The Greek part of Wikipedia, * The Greek part of European Parliament Proceedings Parallel Corpus, and * The Greek part of OSCAR, a cleansed version of Common Crawl. * The entire corpus of Greek legislation, as published by the National Publication Office, * The entire corpus of EU legislation (Greek translation), as published in Eur-Lex. * We trained BERT using the official code provided in Google BERT's GitHub repository (https://github.com/google-research/bert).* We then used Hugging Face's Transformers conversion script to convert the TF checkpoint and vocabulary in the desired format in order to be able to load the model in two lines of code for both PyTorch and TF2 users. * We released a model similar to the English `bert-base-uncased` model (12-layer, 768-hidden, 12-heads, 110M parameters). * We chose to follow the same training set-up: 1 million training steps with batches of 256 sequences of length 512 with an initial learning rate 1e-4. * We were able to use a single Google Cloud TPU v3-8 provided for free from TensorFlow Research Cloud (TFRC), while also utilizing GCP research credits. Huge thanks to both Google programs for supporting us!

• MobileCLIP2-S2-OpenCLIP↗

  timm

  These weights and model card are adapted from the original Apple model at https://huggingface.co/apple/MobileCLIP2-S2. This version uses canonical OpenCLIP configs and weight naming. MobileCLIP2 was introduced in MobileCLIP2: Improving Multi-Modal Reinforced Training (TMLR August 2025 Featured ), by Fartash Faghri, Pavan Kumar Anasosalu Vasu, Cem Koc, Vaishaal Shankar, Alexander T Toshev, Oncel Tuzel, Hadi Pouransari. * `MobileCLIP2-S4` matches the accuracy of SigLIP-SO400M/14 with 2x fewer parameters and surpasses DFN ViT-L/14 at 2.5x lower latency measured on iPhone12 Pro Max. * `MobileCLIP-S3/S4` are our new architectures trained on MobileCLIP’s training dataset, DataCompDR-1B (dashed lines). * Our smallest variant `MobileCLIP-S0` obtains similar zero-shot performance as OpenAI's ViT-B/16 model while being 4.8x faster and 2.8x smaller. * `MobileCLIP-S2` obtains better avg zero-shot performance than SigLIP's ViT-B/16 model while being 2.3x faster and 2.1x smaller, and trained with 3x less seen samples. * `MobileCLIP-B (LT)` attains zero-shot ImageNet performance of **77.2%** which is significantly better than recent works like DFN and SigLIP with similar architectures or even OpenAI's ViT-L/14@336.

• MiniCPM-V-4.7-35B-A3B-Uncensored-GGUF↗

  AtomicChat

  One refusal direction projected out of OpenBMB's original weights, shipped two ways: three baked GGUF builds, and a 106 MB LoRA adapter for the MiniCPM-V 4.7 quants you already have. The measurements behind this card are public. On held-out prompts, refusals fall from 96.3% to 0% in English, and from 13 of 30 to none with thinking on. MMLU moves by 0.25 points, inside the noise. You can now run the Uncensored MiniCPM-V 4.7 in Atomic Chat , locally on your own machine. The 20.1 GB build matches the original's token choice 94.5% of the time; our quant of the unmodified model at the same size does 94.9%. Already running our MiniCPM-V 4.7 quants ? Add the adapter with --lora-scaled instead of downloading a new model. *28 Oct 2026: each build is now a single file instead of 11 to 16 shards of 2 GB. The weights are unchanged (tensor-for-tensor identical).*

• deepseek-vl-1.3b-chat↗

  deepseek-community

  Deepseek-VL was introduced by the DeepSeek AI team. It is a vision-language model (VLM) designed to process both text and images for generating contextually relevant responses. The model leverages LLaMA as its text encoder, while SigLip is used for encoding images. > We present DeepSeek-VL, an open-source Vision-Language (VL) Model designed for real-world vision and language understanding applications. Our approach is structured around three key dimensions: We strive to ensure our data is diverse, scalable, and extensively covers real-world scenarios including web screenshots, PDFs, OCR, charts, and knowledge-based content, aiming for a comprehensive representation of practical contexts. Further, we create a use case taxonomy from real user scenarios and construct an instruction tuning dataset accordingly. The fine-tuning with this dataset substantially improves the model's user experience in practical applications. Considering efficiency and the demands of most real-world scenarios, DeepSeek-VL incorporates a hybrid vision encoder that efficiently processes high-resolution images (1024 x 1024), while maintaining a relatively low computational overhead. This design choice ensures the model's ability to capture critical semantic and detailed information across various visual tasks. We posit that a proficient Vision-Language Model should, foremost, possess strong language abilities. To ensure the preservation of LLM capabilities during pretraining, we investigate an effective VL pretraining strategy by integrating LLM training from the beginning and carefully managing the competitive dynamics observed between vision and language modalities. The DeepSeek-VL family (both 1.3B and 7B models) showcases superior user experiences as a vision-language chatbot in real-world applications, achieving state-of-the-art or competitive performance across a wide range of visual-language benchmarks at the same model size while maintaining robust performance on language-centric benchmarks. We have made both 1.3B and 7B models publicly accessible to foster innovations based on this foundation model. Quantization reduces the memory burden of large models by representing the weights i

• Spark-X2.5-4B-abliterated-FIT-GGUF↗

  SC117

  FIT-GGUF v0.2.0 GATE-VERIFIED TIERS 4 TIERS · 2.26–3.15 GiB ABLITERATED (T615) MEASURED KL + SAME-TOP APACHE-2.0 Spark-X2.5-4B-abliterated · FIT-GGUF Four fidelity tiers of a 1M-context hybrid-attention 4B — every shipped file re-verified against its own BF16, byte-exact and reproducible. 🧭 About FIT-GGUF — the tool behind these files Every file in this repository was planned, executed and verified by FIT-GGUF , an open-source, deterministic tensor-level planning layer on top of llama.cpp quantization. Standard GGUF quantization asks you to pick one of a handful of presets; FIT-GGUF instead asks what quality do you want , then finds and verifies the smallest GGUF that demonstrably meets it. Traditional GGUF gives you presets. FIT gives you a fidelity contract: macro KL ≤ tier anchor ∧ same-top ≥ model-calibrated floor. Deterministic size prediction & byte-exact delivery ✅ Validated (G2 gate, delta = 0) Universally optimal tensor allocation ⚠️ Not established — FIT claims verified fidelity contracts, not a universal quality optimum The method, the preregistered research record and the fit CLI are open source: github.com/Scorp1o117/FIT-GGUF ⚠️ Safety notice / 安全提示 The source model is an abliterated, refusal-removed model with no meaningful built-in guardrails, and may comply with harmful, illegal or unsafe requests. Use it only where you can provide appropriate moderation, access control and legal review. Do not deploy it to end users without your own safety layer. 源模型经过拒答方向消融，不具备可靠的内置安全护栏；请仅在合法、受控、具备审核与滥用防护的环境中使用，使用者自行承担部署责任。