add exception list

Author: cccclai

.github/workflows/_link_check.yml

@@ -77,5 +77,6 @@ jobs:
           echo
           echo "File size lint failed: some files exceed the 1 MB limit."
           echo "If you really need large files, consider using Git LFS or storing them elsewhere."
+          echo "If you really need to get unblocked and check in the file, can add it to the EXCEPTIONS list in scripts/lint_file_size.sh."
           exit 1
         }

scripts/lint_file_size.sh

@@ -11,6 +11,27 @@ status=0
 
 green='\e[1;32m'; red='\e[1;31m'; cyan='\e[1;36m'; reset='\e[0m'
 
+# List of files to skip (relative paths)
+EXCEPTIONS=(
+  "examples/models/llama/params/demo_rand_params.pth"
+  "examples/models/llama/tokenizer/test/resources/test_tiktoken_tokenizer.model"
+  "examples/qualcomm/oss_scripts/llama/artifacts/stories260k_hybrid_llama_qnn.pte"
+  # Following needs to be clean up
+  "examples/mediatek/models/llm_models/weights/Llama-3.2-1B-Instruct/tokenizer.json"
+  "examples/mediatek/models/llm_models/weights/Llama-3.2-3B-Instruct/tokenizer.json"
+  "examples/mediatek/models/llm_models/weights/llama3-8B-instruct/tokenizer.json"
+)
+
+is_exception() {
+  local f=$1
+  for ex in "${EXCEPTIONS[@]}"; do
+    if [[ "$f" == "$ex" ]]; then
+      return 0
+    fi
+  done
+  return 1
+}
+
 if [ $# -eq 2 ]; then
   base=$1
   head=$2
@@ -22,6 +43,10 @@ else
 fi
 
 for file in $files; do
+  if is_exception "$file"; then
+    echo -e "${cyan}SKIP${reset}  $file (in exception list)"
+    continue
+  fi
   if [ -f "$file" ]; then
     # Set size limit depending on extension
     if [[ "$file" =~ \.(png|jpg|jpeg|gif|svg|mp3|mp4)$ ]]; then