サイト内検索

AI モデル 2119

• minimax_h3_pdmd_2nfe_comfyui↗

  Iwannapose

  ComfyUI-format conversion of the PDMD 2-NFE LoRA — the 2-step (2 NFE) student distilled from MiniMax-H3-33B with Projected Distribution Matching Distillation, rank 128, covering attention projections and both feed-forward layers of all 50 transformer blocks + 2 token-refiner blocks. (Companion repo for the 4-NFE variant: `Iwannapose/minimax_h3_pdmd_4nfe_comfyui`.) The two are numerically identical (the source fp32 file is an exact cast of the bf16 one); the bf16 file is the one to use. Both are larger than the source because the fused `qkv_proj` stores the block-diagonal zero padding explicitly. Source keys are Diffusers PEFT names (`transformer.transformer_blocks.N.attn.to_q.lora_A.weight` …); target keys follow the ComfyUI H3 layout (same scheme as the proven `minimax_h3_fl2v_turbo_8step_v1.0_768p_comfyui_bf16.safetensors`):

• Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NEO-CODER-MAX-MTP-GGUF↗

  Sameer3336s

• Wan2.1-VACE-1.3B-GGUF↗

  samuelchristlie

  Wan2.1 is an open-source suite of video foundation models, compatible with consumer-grade GPUs, that excels in various video generation tasks like text-to-video, image-to-video, and video editing, even supporting visual text generation. This model is a derivative work of the original model licensed under the Apache 2.0 License, and is therefore distributed under the terms of the same license. Thanks to Patrick Gillespie for creating the ASCII text art tool used in this project https://patorjk.com/software/taag/

• Ternary-Bonsai-2-27B-PTQ1_0-MTP-GGUF↗

  sudoingx

  One line installs it on the card you already own, RTX 20, 30, 40 and 50, and AMD Radeon in beta, with 8 GB or more: it picks the build and the model file for your VRAM, checks every SHA256 and starts it, with the chat page at `http://127.0.0.1:8899` and an OpenAI-compatible API at `/v1` (model `bonsai-2-27b`). Windows, in PowerShell (measured on native Windows on an RTX 3060 12GB and an RTX 5070 Ti 16GB; RTX 20 in beta): AMD Radeon (beta): the same two lines, on PrismML's own Vulkan build, no ROCm to install. 8 GB cards get up to 64K, 12 GB about 195K, 16 GB and up the full 262K with vision, all from the files in this repo; the cards and the numbers are in Radeon (beta).

• gemma-4-26B-A4B-it-qat-AWQ-INT4↗

  cyankiwi

  Hugging Face | GitHub | Launch Blog | Documentation | Technical Report > [!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.

• pythia-1b-deduped↗

  EleutherAI

  The *Pythia Scaling Suite* is a collection of models developed to facilitate interpretability research (see paper). It contains two sets of eight models of sizes 70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two models: one trained on the Pile, and one trained on the Pile after the dataset has been globally deduplicated. All 8 model sizes are trained on the exact same data, in the exact same order. We also provide 154 intermediate checkpoints per model, hosted on Hugging Face as branches. The Pythia model suite was designed to promote scientific research on large language models, especially interpretability research. Despite not centering downstream performance as a design goal, we find the models match or exceed the performance of similar and same-sized models, such as those in the OPT and GPT-Neo suites. Previously, we released an early version of the Pythia suite to the public. However, we decided to retrain the model suite to address a few hyperparameter discrepancies. This model card lists the changes ; see appendix B in the Pythia paper for further discussion. We found no difference in benchmark performance between the two Pythia versions. The old models are still available, but we suggest the retrained suite if you are just starting to use Pythia. **This is the current release.**

• SmolLM-135M↗

  HuggingFaceTB

  1. Model Summary 2. Limitations 3. Training 4. License 5. Citation SmolLM is a series of state-of-the-art small language models available in three sizes: 135M, 360M, and 1.7B parameters. These models are built on Cosmo-Corpus, a meticulously curated high-quality training dataset. Cosmo-Corpus includes Cosmopedia v2 (28B tokens of synthetic textbooks and stories generated by Mixtral), Python-Edu (4B tokens of educational Python samples from The Stack), and FineWeb-Edu (220B tokens of deduplicated educational web samples from FineWeb). SmolLM models have shown promising results when compared to other models in their size categories across various benchmarks testing common sense reasoning and world knowledge. For detailed information on training, benchmarks and performance, please refer to our full blog post. While SmolLM models have been trained on a diverse dataset including educational content and synthetic texts, they have limitations. The models primarily understand and generate content in English. They can produce text on a variety of topics, but the generated content may not always be factually accurate, logically consistent, or free from biases present in the training data. These models should be used as assistive tools rather than definitive sources of information. Users should always verify important information and critically evaluate any generated content. For a more comprehensive discussion of the models' capabilities and limitations, please refer to our full blog post..

• HY-MT1.5-1.8B-GGUF↗

  腾讯

  🤗  Hugging Face   |   🕹️  Demo      🤖  ModelScope   |   Hunyuan Translation Model Version 1.5 includes a 1.8B translation model, HY-MT1.5-1.8B, and a 7B translation model, HY-MT1.5-7B. Both models focus on supporting mutual translation across 33 languages and incorporating 5 ethnic and dialect variations. Among them, HY-MT1.5-7B is an upgraded version of our WMT25 championship model, optimized for explanatory translation and mixed-language scenarios, with newly added support for terminology intervention, contextual translation, and formatted translation. Despite having less than one-third the parameters of HY-MT1.5-7B, HY-MT1.5-1.8B delivers translation performance comparable to its larger counterpart, achieving both high speed and high quality. After quantization, the 1.8B model can be deployed on edge devices and support real-time translation scenarios, making it widely applicable. - HY-MT1.5-1.8B achieves the industry-leading performance among models of the same size, surpassing most commercial translation APIs. - HY-MT1.5-1.8B supports deployment on edge devices and real-time translation scenarios, offering broad applicability. - HY-MT1.5-7B, compared to its September open-source version, has been optimized for annotated and mixed-language scenarios. - Both models support terminology intervention, contextual translation, and formatted translation.

• gemma-3-12b-it-qat-q4_0-unquantized↗

  Lightricks

  > [!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.

• zerank-2-reranker-vllm↗

  polaria-tech

  **ZeroEntropy's zerank-2 reranker, ready for `vllm serve`.** Same weights, same scores, about twice as fast as the official path. Serving zerank-2 on vLLM is deceptively easy to get wrong: the server starts, every request returns 200, and the scores are quietly off. We converted the model, measured it against the official implementation, benchmarked it on an H100 and wrote down every trap we hit, so you can just serve it. > [!TIP] > **Also available in FP8:** `polaria-tech/zerank-2-reranker-vllm-fp8`. > Same results (NDCG@10 within ±0.004 on 4 MTEB reranking tasks), ~1.2x faster on H100 and half the > size (4.2 GB). Needs a GPU with native FP8 (Ada, Hopper or newer).