Merge pull request #116287 from anna328p/gpt2tc

gpt2tc: init at 2020-12-30

pkgs/tools/text/gpt2tc/0001-add-python-shebang.patch

@@ -0,0 +1,10 @@
+diff --git a/gpt2convert.py b/gpt2convert.py
+index 34ca909..6e6cac5 100644
+--- a/gpt2convert.py
++++ b/gpt2convert.py
+@@ -1,3 +1,5 @@
++#!/usr/bin/env python3
++
+ import sys
+ import tensorflow as tf
+ import numpy as np

pkgs/tools/text/gpt2tc/0002-fix-download-url.patch

@@ -0,0 +1,11 @@
+diff --git a/download_model.sh b/download_model.sh
+index 9cb401f..ad1dc62 100755
+--- a/download_model.sh
++++ b/download_model.sh
+@@ -13,5 +13,5 @@ mkdir -p models/$model
+ for filename in checkpoint encoder.json hparams.json model.ckpt.data-00000-of-00001 model.ckpt.index model.ckpt.meta vocab.bpe; do
+   fetch=$model/$filename
+   echo "Fetching $fetch"
+-  curl --output models/$fetch https://storage.googleapis.com/gpt-2/models/$fetch
++  curl --output models/$fetch https://openaipublic.blob.core.windows.net/gpt-2/models/$fetch
+ done

pkgs/tools/text/gpt2tc/default.nix

@@ -0,0 +1,47 @@
+{ lib, stdenv, fetchurl, autoPatchelfHook, python3 }:
+
+stdenv.mkDerivation rec {
+  pname = "gpt2tc";
+  version = "2020-12-30";
+
+  src = fetchurl {
+    url = "https://bellard.org/nncp/gpt2tc-${version}.tar.gz";
+    hash = "sha256-KFcl7E6iGx50JaJI1jwzKAdkrkbNngPbrEA/ZSyG+uY=";
+  };
+
+  patches = [
+    # Add a shebang to the python script so that nix detects it as such and
+    # wraps it properly. Otherwise, it runs in shell and freezes the system.
+    ./0001-add-python-shebang.patch
+
+    # Update the source URL for the models because the old one is down.
+    ./0002-fix-download-url.patch
+  ];
+
+  nativeBuildInputs = [ autoPatchelfHook ];
+
+  buildInputs = [
+    (python3.withPackages (p: with p; [ numpy tensorflow ]))
+  ];
+
+  dontBuild = true;
+
+  installPhase = ''
+    runHook preInstall
+
+    install -D -m755 -t $out/bin gpt2tc
+    install -T -m755 download_model.sh $out/bin/gpt2-download-model
+    install -T -m755 gpt2convert.py $out/bin/gpt2-convert
+    install -D -m644 -t $out/share/gpt2tc readme.txt gpt2vocab.txt Changelog
+
+    runHook postInstall
+  '';
+
+  meta = with lib; {
+    description = "Text completion and compression using GPT-2";
+    homepage = "https://bellard.org/nncp/gpt2tc.html";
+    license = licenses.unfree;
+    platforms = [ "x86_64-linux" ];
+    maintainers = with maintainers; [ anna328p ];
+  };
+}

pkgs/top-level/all-packages.nix

@@ -5059,6 +5059,8 @@ in
 
   gparted = callPackage ../tools/misc/gparted { };
 
+  gpt2tc = callPackage ../tools/text/gpt2tc { };
+
   ldmtool = callPackage ../tools/misc/ldmtool { };
 
   gpodder = callPackage ../applications/audio/gpodder { };