1 files changed, 34 insertions, 0 deletions
diff --git a/nixpkgs/pkgs/development/libraries/sentencepiece/default.nix b/nixpkgs/pkgs/development/libraries/sentencepiece/default.nix
new file mode 100644
index 000000000000..8ab0f8ef69a6
--- /dev/null
+++ b/nixpkgs/pkgs/development/libraries/sentencepiece/default.nix
@@ -0,0 +1,34 @@
+{ lib
+, fetchFromGitHub
+, stdenv
+, cmake
+, gperftools
+
+, withGPerfTools ? true
+}:
+
+stdenv.mkDerivation rec {
+  pname = "sentencepiece";
+  version = "0.1.95";
+
+  src = fetchFromGitHub {
+    owner = "google";
+    repo = pname;
+    rev = "v${version}";
+    sha256 = "0mv7vgsvd7hjssidxy7fjfmwqy68vjcia8pajji11q2fkfp3cg67";
+  };
+
+  nativeBuildInputs = [ cmake ];
+
+  buildInputs = lib.optionals withGPerfTools [ gperftools ];
+
+  outputs = [ "bin" "dev" "out" ];
+
+  meta = with lib; {
+    homepage = "https://github.com/google/sentencepiece";
+    description = "Unsupervised text tokenizer for Neural Network-based text generation";
+    license = licenses.asl20;
+    platforms = platforms.unix;
+    maintainers = with maintainers; [ danieldk pashashocky ];
+  };
+}