Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

llama.cpp Prebuilt CUDA Binaries (Kaggle / Ubuntu 24.04)

Prebuilt static binaries (-DBUILD_SHARED_LIBS=OFF) of llama.cpp compiled on Kaggle (NVIDIA T4 environment) with CUDA support.

Build Details

  • Build Date: November 14, 2026
  • llama.cpp version: 1.0.0-dev
  • ggml version: 1.0.0 (commit 32bdc3a89)
  • OS: Ubuntu 24.04 x86_64
  • CUDA Toolkit: 13.0.88
  • CUDA Architecture: sm_75 (75-real β€” NVIDIA T4 / Turing and compatible)
  • Compiler: GCC 13.3.0 (-march=native, OpenMP 4.5)
  • Archive: llama_cuda_bin.zip (~2.25 GB)

Included Binaries

Located inside bin/:

  • llama-server (with embedded Web UI assets b12377)
  • llama-cli / llama
  • llama-bench
  • llama-quantize
  • llama-mtmd-cli
  • llama-embedding, llama-perplexity, llama-tts, llama-gguf-split, and other utilities/tests.

Quick Usage (Kaggle / Colab / Ubuntu 24.04)

# Download and extract
wget https://huggingface.co/datasets/ggggggggg2/llama-cpp-bin-kaggle/resolve/main/llama_cuda_bin.zip
unzip llama_cuda_bin.zip
chmod +x bin/*

# Run server example
./bin/llama-server -m your-model.gguf --port 8080 -ngl 99

Build Command Used

git clone https://github.com/ggml-org/llama.cpp.git
cmake -B build -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=OFF
cmake --build build --config Release -j 4

Notes

  • Compiled specifically for sm_75 (Kaggle T4). It will not run on older GPUs (e.g., P100 / sm_60).
  • Requires compatible NVIDIA CUDA runtime libraries and drivers on the host machine.
  • Source code and license belong to ggml-org/llama.cpp (MIT License).
Downloads last month
45