Update runners/KAGGLE_BUILD_SARUS_SFT_V2_DATASET.txt with user enhancements

runners/KAGGLE_BUILD_SARUS_SFT_V2_DATASET.txt

@@ -1,26 +1,27 @@
-# ==============================================================================
-# 🚀 ViuAI Sarus-500M — SFT v2 Pure Chat Dataset Builder (Kaggle 1-Click)
-# ==============================================================================
-# Run this entire block in a single Kaggle Notebook cell.
-# It streams and builds ~200,000 pure conversational samples (0% math, 0% code)
-# and uploads directly to Hugging Face!
-# ==============================================================================
-
-# 1. Install required packages
-!pip install -q tokenizers datasets huggingface_hub numpy
-
-# 2. Fetch the builder script from Hugging Face
-import os, shutil
-from huggingface_hub import hf_hub_download
-
-HF_TOKEN = os.environ.get("HF_TOKEN") or ("".join(["hf_", "ssyCVhuny", "XxjGdqKp", "VLPpkmWK", "FrrMOIFbg"]))
-os.environ["HF_TOKEN"] = HF_TOKEN
-REPO_ID = "ViuAI/ViuAI-500M"
-
-os.makedirs("code", exist_ok=True)
-p = hf_hub_download(repo_id=REPO_ID, filename="code/build_sft_v2_chat_dataset.py", token=HF_TOKEN, force_download=True)
-shutil.copy(p, "code/build_sft_v2_chat_dataset.py")
-print("✅ build_sft_v2_chat_dataset.py ready!")
-
-# 3. Launch Builder with LIVE Real-Time Terminal Output (-u unbuffered)
-!python -u code/build_sft_v2_chat_dataset.py --phase all --upload_hf
+# ==============================================================================
+# 🚀 ViuAI Sarus-500M — SFT v2 Pure Chat Dataset Builder (Kaggle 1-Click)
+# ==============================================================================
+# Run this entire block in a single Kaggle Notebook cell.
+# It streams and builds ~200,000 pure conversational samples (0% math, 0% code)
+# and uploads directly to Hugging Face!
+# ==============================================================================
+
+# 1. Install required packages
+!pip install -q tokenizers datasets huggingface_hub numpy
+
+# 2. Fetch the builder script from Hugging Face
+import os, shutil
+from huggingface_hub import hf_hub_download
+
+HF_TOKEN = os.environ.get("HF_TOKEN")
+if HF_TOKEN:
+    os.environ["HF_TOKEN"] = HF_TOKEN
+REPO_ID = "ViuAI/ViuAI-500M"
+
+os.makedirs("code", exist_ok=True)
+p = hf_hub_download(repo_id=REPO_ID, filename="code/build_sft_v2_chat_dataset.py", token=HF_TOKEN, force_download=True)
+shutil.copy(p, "code/build_sft_v2_chat_dataset.py")
+print("✅ build_sft_v2_chat_dataset.py ready!")
+
+# 3. Launch Builder with LIVE Real-Time Terminal Output (-u unbuffered)
+!python -u code/build_sft_v2_chat_dataset.py --phase all --upload_hf