trained embeddings added

IITH-Compilers · Feb 13, 2025 · 05fb1a8 · 05fb1a8
1 parent ca5f884
commit 05fb1a8
Show file tree

Hide file tree

Showing 151 changed files with 8,055 additions and 7,458 deletions.
diff --git a/seed_embeddings/OpenKE/config/Trainer.py b/seed_embeddings/OpenKE/config/Trainer.py
@@ -57,6 +57,7 @@ def __init__(
         self.analogies = analogy.AnalogyScorer(analogy_file=analogy_file)
 
     def load_entity_names(self, index_dir):
+        print("Loading entity names...", index_dir)
         with open(os.path.join(index_dir, "entity2id.txt")) as fEntity:
             content = fEntity.read()
 

diff --git a/seed_embeddings/OpenKE/generate_embedding.py b/seed_embeddings/OpenKE/generate_embedding.py
@@ -95,6 +95,7 @@ def train(arg_conf):
         alpha=0.00729,
         opt_method="Adam",
         checkpoint_dir=checkpoint_dir,
+        index_dir=arg_conf.index_dir,
     )
     trainer.run(
         link_prediction=False,
@@ -137,10 +138,10 @@ def findRep(src, dest, index_dir):
         metavar="DIRECTORY",
         help="Location of the directory entity2id.txt, train2id.txt and relation2id.txt",
         required=False,
-        default="../seed_embeddings/preprocessed/",
+        default="../preprocessed/",
     )
     parser.add_argument(
-        "--epoch", dest="epoch", help="Epochs", required=False, type=int, default=1000
+        "--epoch", dest="epoch", help="Epochs", required=False, type=int, default=20
     )
 
     parser.add_argument(
@@ -188,7 +189,7 @@ def findRep(src, dest, index_dir):
         ),
     )
 
-    findRep(outfilejson, seedfile, arg_conf.index_dir)
+    findRep(outfile, seedfile, arg_conf.index_dir)
 
     print("Training finished...")
     print("seed file : ", seedfile)
diff --git a/seed_embeddings/OpenKE/generate_embedding_ray.py b/seed_embeddings/OpenKE/generate_embedding_ray.py
@@ -261,6 +261,8 @@ def reformat_embeddings(input_str):
         metric = "loss"
         mode = "min"
 
+    print(metric, mode)
+
     scheduler = ASHAScheduler(
         time_attr="training_iteration",
         max_t=arg_conf.epoch,
@@ -270,11 +272,12 @@ def reformat_embeddings(input_str):
         mode=mode,
     )
     optuna = OptunaSearch(metric="loss", mode="min")
+    print("GPU usage ", arg_conf.use_gpu)
 
     if arg_conf.use_gpu:
         train_with_resources = tune.with_resources(
             tune.with_parameters(train, args=arg_conf),
-            resources={"cpu": 8, "gpu": 0.15},
+            resources={"cpu": 2, "gpu": 0.11},
         )
     else:
         train_with_resources = tune.with_resources(
@@ -286,9 +289,9 @@ def reformat_embeddings(input_str):
         param_space=search_space,
         tune_config=TuneConfig(
             search_alg=optuna,
-            max_concurrent_trials=12,
+            max_concurrent_trials=4,
             scheduler=scheduler,
-            num_samples=128,
+            num_samples=16,
         ),
         run_config=RunConfig(
             storage_path=arg_conf.storage_path,
@@ -336,6 +339,7 @@ def reformat_embeddings(input_str):
     # Construct the output file name using the best hyperparameters
     outfile = os.path.join(
         index_dir,
+        "embeddings/",
         "seedEmbedding_{}_{}Dim_{}Alpha_{}batchsize_{}margin.ckpt".format(
             metric,
             dim,
@@ -357,6 +361,7 @@ def reformat_embeddings(input_str):
 
         embeddings_path = os.path.join(
             index_dir,
+            "embeddings/",
             "seedEmbedding_{}_{}Dim_{}Alpha_{}batchsize_{}margin.ckpt".format(
                 metric,
                 dim,

diff --git a/seed_embeddings/OpenKE/launch_training.sh b/seed_embeddings/OpenKE/launch_training.sh
@@ -0,0 +1,4 @@
+#! /bin/bash
+
+python3 generate_embedding_ray.py --is_analogy True --use_gpu True --dim 75 --index_dir "/home/intern23002/iitH/ir2vec/IR2Vec/seed_embeddings/preprocessed" \
+	--analogy_file "/home/intern23002/iitH/ir2vec/IR2Vec/seed_embeddings/OpenKE/analogies.txt"
diff --git a/seed_embeddings/triplets.sh b/seed_embeddings/triplets.sh
@@ -34,7 +34,7 @@ if [ -z $4 ]; then
 	exit
 fi
 
-LLVM_BUILD=$5
+LLVM_BUILD=/usr
 
 if [ -z $LLVM_BUILD ]; then
 	echo "5st arg should have a valid Build path"
@@ -70,13 +70,13 @@ while read p; do
 		fi
 		USED_OPT[$a]=$opt
 		DEBUG echo "opt from $opt"
-		${LLVM_BUILD}/bin/opt-19 -S -$opt $p -o $tmpfile
+		${LLVM_BUILD}/bin/opt-20 -S -$opt $p -o $tmpfile
 		$COLLECT_BUILD/bin/ir2vec -collectIR -o $4 $tmpfile &>/dev/null
 		let "a++"
 		rm "$tmpfile"
 	done &
 	if [ $counter == 100 ]; then
-		sleep 20
+		sleep 3
 		counter=0
 	fi
 

diff --git a/src/test-suite/CMakeLists.txt b/src/test-suite/CMakeLists.txt
@@ -70,11 +70,11 @@ endif()
 
 # sanity checks and lit configs
 configure_file(sanity_check.sh.cmake sanity_check.sh @ONLY)
-file(COPY PE-benchmarks-llfiles-llvm19 DESTINATION ./)
+file(COPY PE-benchmarks-llfiles-llvm20 DESTINATION ./)
 file(COPY sqlite3.ll DESTINATION ./)
 file(COPY oracle DESTINATION ./)
 file(COPY ../../vocabulary DESTINATION ./)
-file(COPY index-llvm19.files DESTINATION ./)
+file(COPY index-llvm20.files DESTINATION ./)
 
 
 configure_file(lit.site.cfg.py.in lit.site.cfg.py @ONLY)

diff --git a/...he_closest_pair_from_two_sorted_arrays.ll → ...he_closest_pair_from_two_sorted_arrays.ll b/...he_closest_pair_from_two_sorted_arrays.ll → ...he_closest_pair_from_two_sorted_arrays.ll
@@ -1,7 +1,7 @@
 ; ModuleID = 'PE-benchmarks/Find_the_closest_pair_from_two_sorted_arrays.cpp'
 source_filename = "PE-benchmarks/Find_the_closest_pair_from_two_sorted_arrays.cpp"
 target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
-target triple = "x86_64-unknown-linux-gnu"
+target triple = "x86_64-pc-linux-gnu"
 
 module asm ".globl _ZSt21ios_base_library_initv"
 
@@ -195,6 +195,6 @@ attributes #4 = { nocallback nofree nounwind willreturn memory(argmem: readwrite
 !2 = !{i32 7, !"PIE Level", i32 2}
 !3 = !{i32 7, !"uwtable", i32 2}
 !4 = !{i32 7, !"frame-pointer", i32 2}
-!5 = !{!"clang version 19.1.7 (https://github.com/llvm/llvm-project.git cd708029e0b2869e80abe31ddb175f7c35361f90)"}
+!5 = !{!"Ubuntu clang version 20.1.0 (++20250204023448+d185bd94ff77-1~exp1~20250204023606.14)"}
 !6 = distinct !{!6, !7}
 !7 = !{!"llvm.loop.mustprogress"}
diff --git a/...rks-llfiles-llvm19/Iterative_QuickSort.ll → ...rks-llfiles-llvm20/Iterative_QuickSort.ll b/...rks-llfiles-llvm19/Iterative_QuickSort.ll → ...rks-llfiles-llvm20/Iterative_QuickSort.ll
@@ -1,7 +1,7 @@
 ; ModuleID = 'PE-benchmarks/Iterative_QuickSort.cpp'
 source_filename = "PE-benchmarks/Iterative_QuickSort.cpp"
 target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
-target triple = "x86_64-unknown-linux-gnu"
+target triple = "x86_64-pc-linux-gnu"
 
 module asm ".globl _ZSt21ios_base_library_initv"
 
@@ -319,7 +319,7 @@ attributes #5 = { nocallback nofree nounwind willreturn memory(argmem: readwrite
 !2 = !{i32 7, !"PIE Level", i32 2}
 !3 = !{i32 7, !"uwtable", i32 2}
 !4 = !{i32 7, !"frame-pointer", i32 2}
-!5 = !{!"clang version 19.1.7 (https://github.com/llvm/llvm-project.git cd708029e0b2869e80abe31ddb175f7c35361f90)"}
+!5 = !{!"Ubuntu clang version 20.1.0 (++20250204023448+d185bd94ff77-1~exp1~20250204023606.14)"}
 !6 = distinct !{!6, !7}
 !7 = !{!"llvm.loop.mustprogress"}
 !8 = distinct !{!8, !7}