Saved toolsSubmit a tool
FOXAI SEARCH

Search FoxAI

AI Models 2119

  • Qwen3.8-27B-GPTQ-INT4-FP8KV↗

    abhishekchohan

    A 4-bit weight-only **GPTQ** quantization of **Qwen/Qwen3.8-27B** with an **fp8-e4m3 KV cache calibrated alongside the weights**. INT4 weights shrink the footprint ≈2.8× and fp8 KV halves the cache, so the model serves its native **262,144-token** context on a **single 48 GB GPU**. fp8 KV also makes YaRN extension to **524,288 tokens** fit the same card, and up to 1M tokens on 96 GB (retrieval beyond 262k unvalidated). Identical harness and hardware for base and quantized: lm-evaluation-harness with the vLLM 0.27.1 backend on a single NVIDIA RTX PRO 6000 (Blackwell, 96 GB), greedy decoding, seed 1234. The base runs with bf16 KV and the quantized model with fp8-e4m3 KV, each at its recommended serving configuration, so every Δ measures the full int4 + fp8-KV stack rather than weights alone. Generative tasks run with the model's native chat template, thinking enabled, `max_gen_toks=8192`; loglikelihood tasks run untemplated. `minerva_math500` reports `math_verify`, since strict `exact_match` measures answer format, not content, on thinking-model outputs. The BF16 base scores identically (9/9), showing no retrieval degradation through the int4 + fp8-KV stack at the context length it was calibrated at.

  • vit_base_patch16_224.sam_in1k↗

    timm

  • Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO-GGUF↗

    jegreene

    This is a **partially decensored** build of UkisAI's Swift 1.5 Qwen3.8-27B, made with Heretic. It is packaged as a single-file llama.cpp GGUF and quantized the same way as UkisAI's own GSQ-RCO `-mtp` releases, using the same per-tensor allocations and the same imatrix. Three sizes are provided: IQ3_S, IQ3_XXS and IQ2_S. It includes the **MTP head**, for speculative decoding, and a **vision projector** (mmproj) built from Swift's own vision tower. > **Not an official UkisAI release.** UkisAI and Alibaba Cloud do not endorse it. > "Swift" is used only to describe where the model comes from.

  • Xiaomi-OCR-0↗

    SeerRay-Lab

    Xiaomi-OCR-0 A unified 0.8B model for document parsing and OCR-related understanding. English · 简体中文 Xiaomi-OCR-0 is a unified 0.8B OCR-specialized vision-language model for both **document parsing** and **OCR-centric understanding**. Starting from Qwen3.5-0.8B-Base, Xiaomi-OCR-0 is trained on an approximately 170M-sample OCR-centric corpus with a progressive training recipe:

  • CommunityForensics-DeepfakeDet-ViT↗

    buildborderless

    Vision Transformer (ViT-Small) trained on 2.7M samples across 4,803 generators for detecting AI-generated images. Presented in Community Forensics: Using Thousands of Generators to Train Fake Image Detectors (CVPR 2025). **Uploaded for community validation as part of OpenSight** — An upcoming open-source framework for adaptive deepfake detection. **Project OpenSight HF Spaces coming soon with an eval playground and eventually a leaderboard. Preview:**

  • minimax-h3-loras↗

    Hearmeman

    **Free ComfyUI workflows, Runpod templates, and guides:** 👉 hearmemanai.com A mirror of the MiniMax-H3 LoRAs published on CivitAI. CivitAI-matched files have identical SHA-256 checksums (listed below). CivitAI carries the sample videos, the version history and the comment threads; this repository exists so the weights can be pulled with a plain `resolve` URL, without an account. One **anatomy** adapter teaches H3 what a body part looks like. One **action** adapter teaches it what a body does. They are single-concept and they stack, so put them under whatever character or scene LoRA you are already running.

  • Drama-3-GGUF↗

    Jmica

    Experimental native audio.cpp conversions of Fish Audio Drama 3, prepared from checkpoint revision `dd64a85cd82e9521f7202610ede97b367fc83946`. All packages use the same trained model architecture and checkpoint. **Standard GGML K-quants with mixed precision.** The packages use Q2_K, Q3_K_S and Q4_K_S file profiles with pure expert banks and explicit precision overrides. Routed expert tensors use Q2_K, Q3_K and Q4_K storage respectively. `backbone-8bit` means Q8_0 language-backbone matrices; `backbone-4bit` means Q4_K matrices with a Q4_0 text embedding. The table below lists the precision overrides. Audio-tower matrices and waveform codec weights remain FP16; audio-tower normalization and other retained FP32 vectors are unchanged. The Q3_K_S and Q4_K_S file profiles use pure expert banks: all 120 routed expert projection banks have uniform Q3_K or Q4_K storage. Their storage-type selection corresponds to a standard `--pure` K profile with explicit precision overrides for the dense backbone, text embedding and audio components. Architecture-specific higher-bit promotion heuristics are not applied; an unmodified default llama.cpp preset can choose different types for some tensors. See the standard converter's pure and per-tensor options. The native audio.cpp converter takes `--expert-type q3_k` or `--expert-type q4_k`. The table above lists each component's actual storage type.

  • PIGuard↗

    leolee99

    - Website: https://injecguard.github.io/ - Paper: https://aclanthology.org/2025.acl-long.1468.pdf - Code Repo: https://github.com/leolee99/PIGuard Due to some licensing issues, the model name has been changed from **InjecGuard** to **PIGuard**. We apologize for any inconvenience this may have caused. Prompt injection attacks pose a critical threat to large language models (LLMs), enabling goal hijacking and data leakage. Prompt guard models, though effective in defense, suffer from over-defense—falsely flagging benign inputs as malicious due to trigger word bias. To address this issue, we introduce ***NotInject***, an evaluation dataset that systematically measures over-defense across various prompt guard models. NotInject contains 339 benign samples enriched with trigger words common in prompt injection attacks, enabling fine-grained evaluation. Our results show that state-of-the-art models suffer from over-defense issues, with accuracy dropping close to random guessing levels (60\%). To mitigate this, we propose ***PIGuard***, a novel prompt guard model that incorporates a new training strategy, *Mitigating Over-defense for Free* (MOF), which significantly reduces the bias on trigger words. InjecGuard demonstrates state-of-the-art performance on diverse benchmarks including NotInject, surpassing the existing best model by 30.8\%, offering a robust and open-source solution for detecting prompt injection attacks.

  • Homo-GE2PE-Persian-HF↗

    mehdi-hf

    **This is not a new model.** It is a repackaging of MahtaFetrat/Homo-GE2PE-Persian by Elnaz Rahmati et al. — the weights are byte-identical. All credit for the model belongs to the original authors. The `-HF` suffix follows the `Llama-2-7b-hf` convention: same weights, packaged in transformers-native format rather than the original release format. It is not a new version of the model. What changed is only the packaging: the original ships zipped weights plus a vendored copy of Parsivar and a custom `GE2PE.py` wrapper. This repo is a plain `from_pretrained()` model with no extra dependencies, so it can be dropped into an inference pipeline directly.

  • SambaLingo-Hungarian-Base-i1-GGUF↗

    mradermacher