サイト内検索

AI モデル 2119

• odia-distil-gpt2↗

  sam2ai

• Qwen3.8-27B-GGUF-shards↗

  Arki05

  - `source/`: Unsloth's BF16 GGUF (unsloth/Qwen3.8-27B-GGUF), unchanged. - `types/ -hq.gguf`: every linear block tensor at llama.cpp type `T` (IQ1_M … Q8_0, IQ4_NL, MXFP4, NVFP4), rebuilt from the BF16 weights Hessian-quantized. Same type, format and size as llama.cpp's: plain GGUF types, for stock llama.cpp (a GGUF assembled from them holds the plain type `T`); modified versions of the original weights. - `types/ .gguf` (Q2_K … Q8_0): token_embd, output, the multi-token-prediction layer and the linear attention's small `ssm_alpha` / `ssm_beta` tensors at llama.cpp's own quantization. - `manifest.json` (`qcore-store/1`), `skeleton.gguf`, `store.json`, `records/`: what qalloc reads, with every tensor's offset and sha256. The earlier version of this store (every tensor at ~41 types by llama.cpp's and ik_llama.cpp's quantizers) is at revision `0090490`.

• MiniCPM-V-4.6-Thinking↗

  openbmb

  A Pocket-Sized MLLM for Ultra-Efficient Image and Video Understanding on Your Phone * [2026.05.17] ⭐️⭐️⭐️ We release the API service of MiniCPM-V 4.6, with a **public free API key** together! Try it now. **MiniCPM-V 4.6 Thinking** is the long chain-of-thought reasoning variant of MiniCPM-V 4.6. It generates an explicit reasoning trace before producing the final answer, substantially boosting performance on complex multimodal reasoning, math, and OCR-heavy tasks, while keeping the same edge-friendly architecture (SigLIP2-400M vision encoder + Qwen3.5-0.8B LLM) and the mixed 4x/16x visual token compression of MiniCPM-V 4.6.

• Qwen3.8-27B-MLX-4bit↗

  TensorFold

  A native Apple-silicon conversion of Qwen/Qwen3.8-27B , quantized with stock 4-bit affine weights for MLX-VLM and oMLX. This repository contains a stock 4-bit affine MLX conversion of Qwen3.8-27B. The upstream model is a dense, native vision-language model with flexible thinking control and support for text, images, and video. Its tokenizer, processor configuration, chat template, and generation configuration are preserved. The decode figure is the median of three greedy 256-token runs after a 256-token Metal-kernel warm-up. Individual runs measured 39.90, 39.89, and 39.91 tokens/s. This is a practical local reference, not a controlled cross-platform benchmark; prompt length, context growth, sampler settings, memory pressure, thermal state, and runtime versions can materially change performance.

• Qwen3.6-35B-A3B-NVFP4-MTP-GGUF↗

  michaelw9999

  This repo contains two experimental NVFP4 GGUF quantizations of **Qwen3.6-35B-A3B** for `llama.cpp`. This was quantized using my experimental advanced-gguf-quantizer tool. Both models were imatrix calibrated for the first time using a new custom dataset that I am evaluating. All PPL/KLD results were measured against the same BF16 wikitest KLD base, and then compared to the official NVFP4 release by NVIDIA. Further evaluation tests are underway to identify real world performance differences between **TURBO** and **HQ**.

• QwenCleo-ASR-GGUF↗

  mohammedaly22

  **Egyptian Arabic & code-switching speech recognition for audio.cpp: 3–4× faster than PyTorch on the same GPU, no Python** These are standalone GGUF packages of QwenCleo-ASR, a Qwen3-ASR-1.7B fine-tune for Egyptian Arabic and Arabic↔English code-switching (`فبتديك كل ال soft skills بشكل indirect`). On the author's 3,699-utterance Egyptian test set it reaches **19.85% WER, against 41.51% for base Qwen3-ASR-1.7B**. Full results are on the original model card. In audio.cpp it runs through the existing **`qwen3_asr`** family, the same C++ runtime and options as Qwen3-ASR: offline and streaming modes, long-audio chunking, word timestamps with Qwen3 Forced Aligner, and the OpenAI-compatible server.

• Ornith-1.5-35B-A3B-Huihui-Sangreal-MTP-ICE-GGUF↗

  gbuzhf

  GGUF quantizations of huihui-ai's abliterated Ornith-1.5-35B-A3B, made with a Sangreal 36,025-chunk importance matrix, ICE 1.6 per-tensor recipes and ICE 2.0 post-quantization expert repair. Every file keeps the MTP draft layer. The ICE files carry peculiar-ragdoll's v22.5.0 Qwen-Sharp chat template with gbuzhf's arrow4.0 harness operator wrapper. > ⚠️ **Uncensored.** This is an abliterated checkpoint — refusal behaviour has been > suppressed in the weights. Sandbox it at the OS level and control its network and > code-execution access; with no refusal backstop, a prompt injection from a hostile > page or third-party code has nothing to stop it. Score: F16 = 100; 70 % full-distribution fidelity `1/(1+KLD)` + 30 % same-top-1, averaged over the three sets.

• wide_resnet50_2.racm_in1k↗

  timm

  This model features: * ReLU activations * single layer 7x7 convolution with pooling * 1x1 convolution shortcut downsample Recipe details: * RandAugment `RACM` recipe. Inspired by and evolved from EfficientNet RandAugment recipes. Published as `B` recipe in ResNet Strikes Back. * RMSProp (TF 1.0 behaviour) optimizer, EMA weight averaging * Step (exponential decay w/ staircase) LR schedule with warmup - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 68.9 - GMACs: 11.4 - Activations (M): 14.4 - Image size: train = 224 x 224, test = 288 x 288 - **Papers:** - ResNet strikes back: An improved training procedure in timm: https://arxiv.org/abs/2110.00476 - Wide Residual Networks: https://arxiv.org/abs/1605.07146 - Deep Residual Learning for Image Recognition: https://arxiv.org/abs/1512.03385 - **Original:** https://github.com/huggingface/pytorch-image-models

• all-MiniLM-L6-v2↗

  sentence-transformers

  This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. Without sentence-transformers, you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. The project aims to train sentence embedding models on very large sentence level datasets using a self-supervised contrastive learning objective. We used the pretrained `nreimers/MiniLM-L6-H384-uncased` model and fine-tuned in on a 1B sentence pairs dataset. We use a contrastive learning objective: given a sentence from the pair, the model should predict which out of a set of randomly sampled other sentences, was actually paired with it in our dataset.

• granite-speech-5.0-470m-turboctc-nc↗

  ibm-granite

  > **Disclaimer:** This model is intended for research and non-commercial use-only. See Granite Speech 5.0 Turbo CTC if interested in commercial use. **Model Summary:** Granite Speech 5.0 TurboCTC Non-commercial is a compact 470 million parameter English ASR model with very high inference speed that is released for research and noncommercial purposes only. We invite users to refer to granite-speech-5.0-470m-turboctc for other use cases. The model consists of a conformer acoustic encoder with block self-attention, self-conditioning and temporal downsampling with an output layer corresponding to 16,384 BPE units. It was trained on approximately 75,000 hours of English audio from public corpora using Connectionist Temporal Classification (CTC) and inference is done non-autoregressively with greedy decoding.