Search FoxAI
AI Models 2119
- Kimi-K3-UD-Q4_K_XL-layers↗
meshllm
This package is derived from unsloth/Kimi-K3-GGUF and keeps the original GGUF distribution split into per-layer artifacts for distributed inference. For upstream architecture details, chat template guidance, sampling recommendations, license terms, and benchmark notes, see the source model card: unsloth/Kimi-K3-GGUF. Generated by the Mesh LLM HF Jobs splitter from `mesh-llm` ref `f932c4d1dc12b3e3a670d5f470cedd5cdcc5db39`. Each artifact is checksummed as it is written, uploaded to this repository, and removed from the job workspace before the next artifact is produced.
- Qwen3.8-Flash-Next-DFlash2↗
incoai
This repository contains the DFlash 2 draft model for `Qwen/Qwen3.8-Flash-Next`. It is not a standalone language model: it runs inside a speculative decoding server and drafts tokens for the target model to verify. DFlash 2 is a block-diffusion drafter for speculative decoding. It predicts a whole block of tokens in a single pass and keeps the top candidates at every position. A lightweight selector then traces one coherent path through them. Two-tap dynamic convolutions in the backbone keep the draft from decaying toward the end of the block. Decoding is lossless: greedy output matches the target model exactly, and sampling preserves its distribution. Serve with SGLang. Flash-Next needs a small target-feature capture patch, included in this repository and validated at TP1, until support lands upstream:
- functiongemma-270m-it-GGUF↗
unsloth
> [!Note] > FunctionGemma is intended to be fine-tuned for your specific function-calling task, including multi-turn use cases. FunctionGemma is a lightweight, open model from Google, built as a foundation for creating your own specialized function calling models. FunctionGemma is not intended for use as a direct dialogue model, and is designed to be highly performant after further fine-tuning, as is typical of models this size. Built on the Gemma 3 270M model and with the same research and technology used to create the Gemini models, FunctionGemma has been trained specifically for function calling. The model has the same architecture as Gemma 3, but uses a different chat format. The model is well suited for text-only function calling. The uniquely small size makes it possible to deploy 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. Furthermore, akin to the base Gemma 270M, the model has been optimized to be extremely versatile, performant on a variety of hardware in single turn scenarios, but should be finetuned on single turn or multiturn task specific data to achieve best accuracy in specific domains. To demonstrate how specializing the 270M parameter model can achieve high performance on specific agentic workflows, we have highlighted two use cases in the Google AI Edge Gallery app. The following is a code example of how to use FunctionGemma to generate a function call from a JSON definition using the Hugging Face Transformers library.
- Wulver↗
Vaelico
**Full fine-tune of Krea 2 Raw (12.8B DiT)** — anthro & furry, with strong anime / kemono range, true multi-character composition and **1,113 artist styles** you can call by name. > In Shetland folklore, the Wulver is a wolf-headed being that was never cruel — > it fished the lochs and left its catch on the windowsills of those in need. - **Much longer training.** 13 full epochs over the curated corpus at 512 px (384,672 samples per epoch, about 156k steps at batch 32), the 13th with the learning rate annealed to zero. Then an artist-focused pass at 512 px and a final curated pass at **1024 px**. - **Artist tokens work.** 1,113 `@artist` tokens, listed in ARTISTS.md. In v0.1 the `@` prefix was ignored. - **Turbo from the official Krea 2 Turbo LoRA**, merged into the Turbo files at strength 1.0 (8 steps, CFG 1). The Non-Turbo files ship too, so you can pick your own strength or skip it.
- Phi-4-reasoning↗
Microsoft
> [!IMPORTANT] > To fully take advantage of the model's capabilities, inference must use `temperature=0.8`, `top_k=50`, `top_p=0.95`, and `do_sample=True`. For more complex queries, set `max_new_tokens=32768` to allow for longer chain-of-thought (CoT). Given the nature of the training data, **always use** ChatML template with the **following system prompt** for inference: *Phi-4-reasoning is also supported out-of-the-box by Ollama, llama.cpp, and any Phi-4 compatible framework.*
- Qwen3.8-27B-Samantha-GGUF↗
mradermacher
- encodec_48khz↗
facebook
This model card provides details and information about EnCodec, a state-of-the-art real-time audio codec developed by Meta AI. EnCodec is a high-fidelity audio codec leveraging neural networks. It introduces a streaming encoder-decoder architecture with quantized latent space, trained in an end-to-end fashion. The model simplifies and speeds up training using a single multiscale spectrogram adversary that efficiently reduces artifacts and produces high-quality samples. It also includes a novel loss balancer mechanism that stabilizes training by decoupling the choice of hyperparameters from the typical scale of the loss. Additionally, lightweight Transformer models are used to further compress the obtained representation while maintaining real-time performance. EnCodec can be used directly as an audio codec for real-time compression and decompression of audio signals. It provides high-quality audio compression and efficient decoding. The model was trained on various bandwiths, which can be specified when encoding (compressing) and decoding (decompressing). Two different setup exist for EnCodec:
- genderage2↗
ivensamdh
- Qwen2.5-14B-Instruct-1M-GGUF↗
Mungert
Our latest quantization method introduces **precision-adaptive quantization** for ultra-low-bit models (1-2 bit), with benchmark-proven improvements on **Llama-3-8B**. This approach uses layer-specific strategies to preserve accuracy while maintaining extreme memory efficiency. All tests conducted on **Llama-3-8B-Instruct** using: - Standard perplexity evaluation pipeline - 2048-token context window - Same prompt set across all quantizations **Key**: - PPL = Perplexity (lower is better) - Δ PPL = Percentage change from standard to DynamicGate - Speed = Inference time (CPU avx2, 2048 token context) - Size differences reflect mixed quantization overhead
- Qwen3.8-27B-DFlash2-W4A16↗
syvai
incoai/Qwen3.8-27B-DFlash2 — the DFlash2 block drafter for Qwen3.8-27B (5 Qwen3-style layers, 1.92B parameters, 3.85 GB in bf16) — requantized to **W4A16 compressed-tensors** (pack-quantized, group 128, symmetric, Marlin) with **GPTQ**, so it fits next to a 27B target on a 24 GB card: **1.19 GB**. Built by and for syv-ai/qwen38-27b-rtx3090 (Qwen3.8-27B on one RTX 3090 with vLLM): `single-user/start_qwen.sh` with `SPEC=dflash2`. Measured on the W4A16 target (RTX 3090, 8 realistic chat prompts × 1,024 tokens, vLLM 0.27.1 with the repo's patches): same acceptance as the bf16 drafter at greedy (3.34-3.65 vs 3.54 tokens per step), about 5% lower at the model's default sampling; per decode step it reads 2.7 GB less, which is what makes it a net win on this card. Details and the pipeline (`drafter/capture_dflash2.py`, `drafter/quant_dflash2.py`) in the repo's `drafter/README.md`.