# Quantized matmul in CUDA, with a PyTorch interface
Original code from FasterTransformer / TensorRT-LLM:
https://github.com/NVIDIA/TensorRT-LLM/tree/main/cpp/tensorrt_llm/kernels
Adapted to support a different quantization scheme.
Raw data
{
"_id": null,
"home_page": "https://github.com/oadirt/quant-matmul",
"name": "quant-matmul",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.7",
"maintainer_email": null,
"keywords": null,
"author": "Tri Dao",
"author_email": "tri@tridao.me",
"download_url": "https://files.pythonhosted.org/packages/c9/47/44c0eea623bc07148a7b494beaa6662334e99f74d0463c9bf8885798a69f/quant_matmul-1.2.0.tar.gz",
"platform": null,
"description": "# Quantized matmul in CUDA, with a PyTorch interface\nOriginal code from FasterTransformer / TensorRT-LLM:\nhttps://github.com/NVIDIA/TensorRT-LLM/tree/main/cpp/tensorrt_llm/kernels\n\nAdapted to support a different quantization scheme.\n",
"bugtrack_url": null,
"license": null,
"summary": "Quantized MatMul in CUDA with a PyTorch interface",
"version": "1.2.0",
"project_urls": {
"Homepage": "https://github.com/oadirt/quant-matmul"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "c94744c0eea623bc07148a7b494beaa6662334e99f74d0463c9bf8885798a69f",
"md5": "1da26ca040b15ef5d4a5bae69877d10a",
"sha256": "5c5d3dcf2618f97500770a0afcfafab2a13760e7e36543f8cee3b04c7b6d9456"
},
"downloads": -1,
"filename": "quant_matmul-1.2.0.tar.gz",
"has_sig": false,
"md5_digest": "1da26ca040b15ef5d4a5bae69877d10a",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.7",
"size": 11627,
"upload_time": "2024-03-20T03:44:36",
"upload_time_iso_8601": "2024-03-20T03:44:36.854429Z",
"url": "https://files.pythonhosted.org/packages/c9/47/44c0eea623bc07148a7b494beaa6662334e99f74d0463c9bf8885798a69f/quant_matmul-1.2.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-03-20 03:44:36",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "oadirt",
"github_project": "quant-matmul",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "quant-matmul"
}