サイト内検索

AI モデル 2119

• Qwen3.8-27B-BaseQuant-V1↗

  basecompute

  Quantized builds of Qwen/Qwen3.8-27B in five sizes. On held-out text, each one is closer to the BF16 model than the unsloth dynamic quant of the same size. Each GGUF is no larger than its unsloth counterpart. The Small bundle needs baseRT 0.4.0 or later. The XS files are text-only and have no MTP layer, like unsloth's `UD-IQ2_S`; the XS bundle was tested on baseRT 0.4.1. The XL GGUF keeps the MTP layer, like `UD-Q6_K_XL`. baseRT does not load Q8_0, so the XL bundle stores those tensors in baseRT's own 8-bit format at the same 8.5 bits per weight. The bundle carries the MTP head for `--speculate mtp-head`, and was tested with baseRT 0.4.1 and superfluid 0.1.4.

• sopro-v2-turbo↗

  litert-community

  Sample from the converted model (plain fp16 weights + int8 AR, streaming vocoder) cloning a CC0 volunteer voice from kyutai/tts-voices (`voice-donations/0a67`): assets/hero_sample.wav — “The morning light fell across the garden while a small bird sang beside the open window.” Source and license of the reference voice: assets/ATTRIBUTION.md. These classic `.tflite` graphs convert Samuel Vitorino's Sopro v2 turbo, a reference-conditioned text-to-speech model for English, European Portuguese, French and German, to the LiteRT CompiledModel API. The package provides fp32 reference graphs, plain weight-only fp16 graphs and a native-int8 merged AR alternative. It produces 24 kHz audio and includes a streaming vocoder, while resampling, spectral DSP, token sampling and the acoustic solver loop stay on the host. The upstream model and source are Apache-2.0; the included LICENSE is retained verbatim. Conversion used `sopro==2.2.0` and upstream model revision `f747f9edfb7b0233a3b7105af3a75603a7213d26`. The file sizes below are exact bytes. There are fourteen fp32 graph files, fourteen plain wfp16 graph files and one native-int8 merged AR file. The Android variants add eight fp32, seven wfp16 and one int8 graph file (rows marked "Android variant"), plus the app source under `android/`. Separate AR prefill/step files and merged alternatives are included so callers can choose one form without loading both. This table covers the executable artifacts, host assets and conversion sources.

• quality-classifier-deberta↗

  NVIDIA

  This is a text classification model that can enable qualitative data annotation, creation of quality-specific blends and addition of metadata tags. The model classifies documents into one of three classes based on the quality of the document: "High", "Medium", "Low" The model was trained using data annotated by human annotators, who considered quality factors such as content accuracy, clarity, coherence, grammar, depth of information, and overall usefulness of the document. This model is used in the NVIDIA NeMo Curator as part of the qualitative filtering module. The model architecture is Deberta V3 Base Context length is 1024 tokens The training set is 22828 Common Crawl text samples, labeled as "High", "Medium", "Low". Here are some examples: 1. Input:

• Guanaco-13B-Uncensored-GGUF↗

  TheBloke

  GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp. GGUF offers numerous advantages over GGML, such as better tokenisation, and support for special tokens. It is also supports metadata, and is designed to be extensible. The creator of the source model has listed its license as `apache-2.0`, and this quantization has therefore used that same license. As this model is based on Llama 2, it is also subject to the Meta Llama 2 license terms, and the license files for that are additionally included. It should therefore be considered as being claimed to be licensed under both licenses. I contacted Hugging Face for clarification on dual licensing but they do not yet have an official position. Should this change, or should Meta provide any feedback on this situation, I will update this section accordingly.

• muscriptor-large↗

  cocktailpeanut

  **MuScriptor** is an open-weight model for **general-purpose, multi-instrument automatic music transcription (AMT)**: it converts a music recording (any genre, multiple simultaneous instruments) into a stream of notes played. This repository hosts the **large** variant (≈1.3B parameters) — the **flagship, best-quality checkpoint**. For a smaller footprint use `muscriptor-medium` (≈300M, good trade-off) or `muscriptor-small` (≈100M, fastest). `load_model` accepts a size keyword (`"small"`/`"medium"`/`"large"`), a local `.safetensors` path, or an `hf://` / `https://` URL. Weights loaded by size keyword (or any `hf://` URL) are cached in the standard Hugging Face cache (`~/.cache/huggingface/hub`, configurable via `HF_HOME`); weights fetched from a plain `http(s)://` URL are cached under `~/.cache/muscriptor/`. Input audio can be WAV or any format `libsndfile` reads (mp3, flac, ogg, m4a, …); it is resampled to 16 kHz mono internally.

• indic-parler-tts↗

  naklitechie

  It is a fine-tuned version of Indic Parler-TTS Pretrained, trained on a **1,806 hours** multilingual Indic and English dataset. **Indic Parler-TTS Mini** can officially speak in 20 Indic languages, making it comprehensive for regional language technologies, and in English. The **21 languages** supported are: Assamese, Bengali, Bodo, Dogri, English, Gujarati, Hindi, Kannada, Konkani, Maithili, Malayalam, Manipuri, Marathi, Nepali, Odia, Sanskrit, Santali, Sindhi, Tamil, Telugu, and Urdu. Thanks to its **better prompt tokenizer**, it can easily be extended to other languages. This tokenizer has a larger vocabulary and handles byte fallback, which simplifies multilingual training.

• Gmcoder↗

  emperorofrome

  **gmcoder** is a 9B merged coding model developed with support from **Galactic Mandate Linux**. It is intended for code generation, debugging, explanations, and algorithmic problem solving. The reported HumanEval+ Mini result is competitive with the compared 9B coding models. On a separate internal knowledge and reasoning benchmark suite, gmcoder scored 30% higher than Qwen3.5-9B and Ornith-1.5-9B. Each model received one greedy completion per task at temperature 0. The run used EvalPlus 0.4.0.dev2 and all 164 HumanEval tasks. The gmcoder and internal comparator entries are Q8_0. The numbers are pass@1; HumanEval+ requires passing both the original and augmented tests. This is a 164-task, single-sample comparison. Oxcoder's HumanEval/132 response was manually skipped after it stopped progressing; a blank answer was scored as a failure, but the cause of that non-completion was not recorded. The internal comparator returned two empty answers; both counted as failures. These results measure short coding problems, not repository-level or agent performance.

• DeepSeek-V4-Flash-0731-NVFP4↗

  NVIDIA

  The NVIDIA DeepSeek-V4-Flash-0731-NVFP4 model is the quantized version of DeepSeek AI's DeepSeek-V4-Flash-0731 model, an autoregressive Mixture-of-Experts language model that uses an optimized Transformer architecture with hybrid attention (Compressed Sparse Attention and Heavily Compressed Attention) and Manifold-Constrained Hyper-Connections. DeepSeek-V4-Flash-0731 is the official DeepSeek-V4-Flash release and includes DeepSeek's DSpark speculative decoding module in the same checkpoint. For more information, refer to the DeepSeek-V4-Flash-0731 model card. The NVIDIA DeepSeek-V4-Flash-0731-NVFP4 model is quantized with Model Optimizer. This model is not owned or developed by NVIDIA. This model has been developed and built to a third-party’s requirements for this application and use case; see link to Non-NVIDIA (DeepSeek-V4-Flash-0731) Model Card. DeepSeek V4 is well-suited for advanced reasoning, agentic AI applications, tool use scenarios, and complex problem-solving in domains such as mathematics, software engineering, and enterprise AI assistants.

• Minimax-H3-fl2va-ref2va-hybrid-unpruned-int8↗

  sandpies

  --- license: other license_name: minimax license_link: https://huggingface.co/MiniMaxAI/MiniMax-H3 tags: - comfyui - minimax - minimax-h3 - video - audio - diffusion - hybrid - int8 base_model: - Comfy-Org/MiniMax-H3 --- Community weight-selection merge of the **unpruned** MiniMax H3 `int8_convrot` checkpoints from Comfy-Org/MiniMax-H3. Same hybrid idea as smhfacct/Minimax-H3-fl2va-ref2va-hybrid-models, but built from the **full / unpruned** INT8 pair (`time_embedder`) instead of the **pruned** pair (`adaln_t_table`).

• sam-hq-vit-large↗

  syscv-community