quantization/versions.json

{
  "nvidia-modelopt": "0.47.0rc0",
  "torch": "2.13.0",
  "transformers": "5.14.1",
  "datasets": "5.0.1",
  "accelerate": "1.14.0",
  "safetensors": "0.8.0",
  "flash-linear-attention": "0.5.2"
}