Search FoxAI
AI Models 2119
- gemma-3-12b-it-qat-GGUF↗
unsloth
> [!Note] > This repository corresponds to the 12B **instruction-tuned** version of the Gemma 3 model using Quantization Aware Training (QAT). > > **The checkpoint in this repository is unquantized, please make sure to quantize with Q4_0 with your favorite tool** > > Thanks to QAT, the model is able to preserve similar quality as `bfloat16` while significantly reducing the memory requirements > to load the model. * [Gemma 3 Technical Report][g3-tech-report] * [Responsible Generative AI Toolkit][rai-toolkit] * [Gemma on Kaggle][kaggle-gemma] * [Gemma on Vertex Model Garden][vertex-mg-gemma3] Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. Gemma 3 models are multimodal, handling text and image input and generating text output, with open weights for both pre-trained variants and instruction-tuned variants. Gemma 3 has a large, 128K context window, multilingual support in over 140 languages, and is available in more sizes than previous versions. Gemma 3 models are well-suited for a variety of text generation and image understanding tasks, including question answering, summarization, and reasoning. Their relatively small size makes it possible to deploy them in environments with limited resources such as laptops, desktops or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone.
- medsam-vit-base↗
wanglab
The **Segment Anything Model (SAM)** produces high-quality object masks from input prompts such as points or boxes, and it can be used to generate masks for all objects in an image. It has been trained on a dataset of 11 million images and 1.1 billion masks, and has strong zero-shot performance on a variety of segmentation tasks. The abstract of the paper states: > We introduce the Segment Anything (SA) project: a new task, model, and dataset for image segmentation. Using our efficient model in a data collection loop, we built the largest segmentation dataset to date (by far), with over 1 billion masks on 11M licensed and privacy respecting images. The model is designed and trained to be promptable, so it can transfer zero-shot to new image distributions and tasks. We evaluate its capabilities on numerous tasks and find that its zero-shot performance is impressive -- often competitive with or even superior to prior fully supervised results. We are releasing the Segment Anything Model (SAM) and corresponding dataset (SA-1B) of 1B masks and 11M images at https://segment-anything.com to foster research into foundation models for computer vision. **Disclaimer**: Content from **this** model card has been written by the Hugging Face team, and parts of it were copy pasted from the original SAM model card.
- Qwen3.8-27B-heretic-SmoothQuant-W8A8-INT8↗
akumaburn
> [!CAUTION] > **Research artifact. Safety alignment has been deliberately removed.** This > model will attempt to comply with harmful, dangerous, illegal, and unethical > requests that the base model refuses, with no content moderation. See > **Safety**. **Provided for research purposes only, with no warranty and no > liability accepted by the author — see the disclaimer at the bottom.** An INT8 **W8A8** quantization of **Qwen3.8-27B-heretic** (the abliterated BF16 build of Qwen/Qwen3.8-27B), for fast serving on Ampere-class (SM 8.0) GPUs where native FP8 is unavailable. The MTP head is unquantized but is **not** byte-identical to the BF16 build's: it carries the same residual-stream rotation and RMSNorm fold as the rest of the model, so the two heads are not interchangeable. Its RMSNorm weights are stored as zeros, which is the identity scale under Qwen3.5's `(1 + weight)` RMSNorm.
- wav2vec2-large-xlsr-kn↗
amoghsgopadi
Fine-tuned facebook/wav2vec2-large-xlsr-53 on Kannada using the OpenSLR SLR79 dataset. When using this model, make sure that your speech input is sampled at 16kHz. The model can be used directly (without a language model) as follows, assuming you have a dataset with Kannada `sentence` and `path` fields: The colab notebook used for training can be found here.
- noobai-XL-1.1↗
Laxhar
It utilizes the latest full Danbooru and e621 datasets for training, with native tags caption. - CFG: 5 ~ 6 - Steps: 25 ~ 30 - Sampling Method:Euler a - Resolution: Total area around 1024x1024. Best to choose from: 768x1344, **832x1216**, 896x1152, 1024x1024, 1152x896, 1216x832, 1344x768 - Data normalization based on various sources and ratings. - Application of time-based decay coefficients according to date recency. - Ranking of images within the entire dataset based on this processing.
- stable-diffusion-2-1-base↗
Manojb
This model card focuses on the model associated with the Stable Diffusion v2-1-base model. This `stable-diffusion-2-1-base` model fine-tunes stable-diffusion-2-base (`512-base-ema.ckpt`) with 220k extra steps taken, with `punsafe=0.98` on the same dataset. - Use it with the `stablediffusion` repository: download the `v2-1_512-ema-pruned.ckpt` here. - Use it with 🧨 `diffusers`
- stable-diffusion-1.5↗
pt-sk
Stable Diffusion is a latent text-to-image diffusion model capable of generating photo-realistic images given any text input. For more information about how Stable Diffusion functions, please have a look at 🤗's Stable Diffusion blog. The **Stable-Diffusion-v1-5** checkpoint was initialized with the weights of the Stable-Diffusion-v1-2 checkpoint and subsequently fine-tuned on 595k steps at resolution 512x512 on "laion-aesthetics v2 5+" and 10% dropping of the text-conditioning to improve classifier-free guidance sampling. 1. Download the weights - v1-5-pruned-emaonly.ckpt - 4.27GB, ema-only weight. uses less VRAM - suitable for inference - v1-5-pruned.ckpt - 7.7GB, ema+non-ema weights. uses more VRAM - suitable for fine-tuning
- Llama-OuteTTS-1.0-1B↗
unsloth
> [!IMPORTANT] > **Important Sampling Considerations** > > When using OuteTTS version 1.0, it is crucial to use the settings specified in the Sampling Configuration section. > > The **repetition penalty implementation** is particularly important - this model requires penalization applied to a **64-token recent window**, > rather than across the entire context window. Penalizing the entire context will cause the model to produce **broken or low-quality output**. > > Currently, **llama.cpp** delivers the most reliable and consistent output quality by default. > Both **llama.cpp** and **EXL2** support this windowed sampling approach, while **Transformers** doesn't. > > To address this limitation, I've implemented a **windowed repetition penalty** for the **Hugging Face Transformers** backend in the **OuteTTS** library, > which significantly improves output quality and resolves sampling issues, providing comparable results to llama.cpp. This update brings significant improvements in speech synthesis and voice cloning—delivering a more powerful, accurate, and user-friendly experience in a compact size. The model is designed to be used with a speaker reference. Without one, it generates random vocal characteristics, often leading to lower-quality outputs. The model inherits the referenced speaker's emotion, style, and accent. When transcribing to other languages with the same speaker, you may observe the model retaining the original accent.
- colqwen2.5-v0.2↗
vidore
ColQwen is a model based on a novel model architecture and training strategy based on Vision Language Models (VLMs) to efficiently index documents from their visual features. It is a Qwen2.5-VL-3B extension that generates ColBERT- style multi-vector representations of text and images. It was introduced in the paper ColPali: Efficient Document Retrieval with Vision Language Models and first released in this repository This model takes dynamic image resolutions in input and does not resize them, changing their aspect ratio as in ColPali. Maximal resolution is set so that 768 image patches are created at most. Experiments show clear improvements with larger amounts of image patches, at the cost of memory requirements. Our training dataset of 127,460 query-page pairs is comprised of train sets of openly available academic datasets (63%) and a synthetic dataset made up of pages from web-crawled PDF documents and augmented with VLM-generated (Claude-3 Sonnet) pseudo-questions (37%). Our training set is fully English by design, enabling us to study zero-shot generalization to non-English languages. We explicitly verify no multi-page PDF document is used both *ViDoRe* and in the train set to prevent evaluation contamination. A validation set is created with 2% of the samples to tune hyperparameters.
- rudialogpt3_medium_based_on_gpt2_2ch↗
BlackSamorez