전체 검색

AI 모델 2119

• gemma-4-31B-it-qat-AWQ-INT4↗

  cyankiwi

  > [!Note] > This model card is for the new versions of the Gemma 4 family optimized with Quantization-Aware Training (QAT), which allows preserving similar quality to bfloat16 while dramatically reducing the memory requirements to load the model. > Four versions of the QAT checkpoints are available: > * **Unquantized QAT checkpoints** (Q4_0): Half-precision weights extracted from the QAT pipeline, ideal for custom downstream compilation and research. Available for Gemma 4 E2B, E4B, 12B, 26B A4B, and 31B, and their drafter models. > * **GGUF** (Q4_0): Ready-to-deploy formats for broad ecosystem compatibility. Available for Gemma 4 E2B, E4B, 12B, 26B A4B, and 31B. > * **Mobile-optimized** (wNa8o8): A custom schema engineered explicitly for mobile hardware efficiency. It features targeted 2-bit decoding layers, optimized KV caches, and static activations to maximize VRAM savings. Available for Gemma 4 E2B and E4B. > * **Compressed Tensors** (w4a16): QAT checkpoints serialized in the compressed-tensors format for native, optimized inference with vLLM. Available for Gemma 4 E2B, E4B, 12B, and 31B. > * **Assistant Compatibility**: When using multi-token prediction (speculative decoding) with an assistant model alongside a QAT target model, the assistant model must also be a QAT checkpoint with the same precision to ensure compatibility. 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 high-end phones to laptops and servers, democratizing access to state-of-the-art AI.

• Florence-2-large-ft↗

  florence-community

  > [!NOTE] > This is the repository for official transformers converted checkpoint of Microsoft's Florence model. **This is a continued pretrained version of Florence-2-large model with 4k context length, only 0.1B samples are used for continue pretraining, thus it might not be trained well. In addition, OCR task has been updated with line separator ('\n'). COCO OD AP 39.8** Florence-2 is an advanced vision foundation model that uses a prompt-based approach to handle a wide range of vision and vision-language tasks. Florence-2 can interpret simple text prompts to perform tasks like captioning, object detection, and segmentation. It leverages our FLD-5B dataset, containing 5.4 billion annotations across 126 million images, to master multi-task learning. The model's sequence-to-sequence architecture enables it to excel in both zero-shot and fine-tuned settings, proving to be a competitive vision foundation model.

• tcd-segformer-mit-b0↗

  restor

  This is a semantic segmentation model that can delineate tree cover in high resolution (10 cm/px) aerial images. This model card is mostly the same for all similar models uploaded to Hugging Face. The model name refers to the specific architecture variant (e.g. nvidia-mit-b0 to nvidia-mit-b5) but the broad details for training and evaluation are identical. This semantic segmentation model was trained on global aerial imagery and is able to accurately delineate tree cover in similar images. The model does not detect individual trees, but provides a per-pixel classification of tree/no-tree.

• vocos-encodec-24khz↗

  charactr

  Audio samples | Paper [[abs]](https://arxiv.org/abs/2306.00814) [[pdf]](https://arxiv.org/pdf/2306.00814.pdf) Vocos is a fast neural vocoder designed to synthesize audio waveforms from acoustic features. Trained using a Generative Adversarial Network (GAN) objective, Vocos can generate waveforms in a single forward pass. Unlike other typical GAN-based vocoders, Vocos does not model audio samples in the time domain. Instead, it generates spectral coefficients, facilitating rapid audio reconstruction through inverse Fourier transform. Additionally, you need to provide a `bandwidth_id` which corresponds to the embedding for bandwidth from the list: `[1.5, 3.0, 6.0, 12.0]`.

• wavlm-base↗

  Microsoft

  The base model pretrained on 16kHz sampled speech audio. When using the model, make sure that your speech input is also sampled at 16kHz. **Note**: This model does not have a tokenizer as it was pretrained on audio alone. In order to use this model **speech recognition**, a tokenizer should be created and the model should be fine-tuned on labeled text data. Check out this blog for more in-detail explanation of how to fine-tune the model. The model was pre-trained on 960h of Librispeech.

• Clef-Flash-GGUF↗

  Pitaya1219

  Low-bit quantizations (Q4_K_M, IQ3_XXS, IQ2_M, IQ2_XS) of Cloudflare/clef-flash in the native `clef` GGUF architecture. Serve with `llama-server` b11415 or later (`/v1/systemone`, llama.cpp#29831). Run through `llama-server` b11415 (`/v1/systemone`, Vulkan, AMD Radeon 8060S), one request at a time, with the same 600 items for every model: 200 each from SST-2 (validation), AG News (test) and BoolQ (validation), sampled with a fixed seed. Agreement and KL are against `Clef-Flash-Q8_0.gguf` from ggml-org, not against bf16. Livesport's per-tensor layout applied step by step to IQ2_M (all with the head at Q8_0), same 600 items, against the same Q8_0 baseline.

• sam3↗

  jetjodh

  SAM 3 is a unified foundation model for promptable segmentation in images and videos. It can detect, segment, and track objects using text or visual prompts such as points, boxes, and masks. Compared to its predecessor SAM 2, SAM 3 introduces the ability to exhaustively segment all instances of an open-vocabulary concept specified by a short text phrase or exemplars. Unlike prior work, SAM 3 can handle a vastly larger set of open-vocabulary prompts. It achieves 75-80% of human performance on our new SA-CO benchmark which contains 270K unique concepts, over 50 times more than existing benchmarks. SAM3 performs Promptable Concept Segmentation (PCS) on images, taking text and/or image exemplars as prompts and returning segmentation masks for **all matching object instances** in the image. SAM3 Video performs Promptable Concept Segmentation (PCS) on videos, taking text as prompts and detecting and tracking **all matching object instances** across video frames.

• hyenadna-medium-160k-seqlen-hf↗

  LongSafari

  Welcome! HyenaDNA is a long-range genomic foundation model pretrained on context lengths of up to **1 million tokens** at **single nucleotide resolution**. You'll see models of different sizes and sequence lengths. There are also original weights-only versions of each model in the LongSafari organization, which are designed to be loaded with the original github repo. These models have identical outputs to the models in the collection above, just different interfaces. In this brief code sample we demonstrate fine-tuning HyenaDNA on a sequence classification task. This sample uses the `medium` checkpoint, with a maximum sequence length of 160k nucleotides. Note that training will fail if you use a sequence length longer than the maximum supported length for your chosen checkpoint.

• Samanthav3-MIXTRALDOLPHIN-LORA-i1-GGUF↗

  mradermacher

• Qwen2.5-VL-3B-Instruct-unsloth-bnb-4bit↗

  unsloth

  In the past five months since Qwen2-VL’s release, numerous developers have built new models on the Qwen2-VL vision-language models, providing us with valuable feedback. During this period, we focused on building more useful vision-language models. Today, we are excited to introduce the latest addition to the Qwen family: Qwen2.5-VL. We extend dynamic resolution to the temporal dimension by adopting dynamic FPS sampling, enabling the model to comprehend videos at various sampling rates. Accordingly, we update mRoPE in the time dimension with IDs and absolute time alignment, enabling the model to learn temporal sequence and speed, and ultimately acquire the ability to pinpoint specific moments. We enhance both training and inference speeds by strategically implementing window attention into the ViT. The ViT architecture is further optimized with SwiGLU and RMSNorm, aligning it with the structure of the Qwen2.5 LLM.