diff --git a/README.md b/README.md
index 1e6bf5a..e6b5bc8 100644
--- a/README.md
+++ b/README.md
@@ -77,15 +77,25 @@ HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download --repo-type dataset rajpurk
 gsutil -m rsync -r data/raw_datasets/self_gen gs://ctx-to-lora/data/raw_datasets/self_gen
 
 # downloading from gcp bucket to login node
-gsutil -m rsync -r gs://ctx-to-lora/data/raw_datasets/self_gen data/raw_datasets/self_gen 
+mkdir -p data/raw_datasets/self_gen
+gsutil -m rsync -r gs://ctx-to-lora/data/raw_datasets/self_gen data/raw_datasets/self_gen
 ```
 
-### Data Generation (v2)
+### Upload/download checkpoints
+```bash
+# upload to bucket
+gsutil -m rsync -r train_outputs gs://ctx-to-lora/train_outputs
+
+# download from bucket
+gsutil -m rsync -r gs://ctx-to-lora/train_outputs train_outputs 
+```
+
+<!-- ### Data Generation (v2)
 ***Download a subset of generated Fineweb-QA directly***
 ```bash
 uv run data/download_generated_fineweb_qa.py
 uv run data/download_self_gen_qa.py
-```
+``` -->
 
 Loading self-generated data
 ```python
diff --git a/gcp_bucket_watcher.py b/gcp_bucket_watcher.py
index 835291c..c44a231 100644
--- a/gcp_bucket_watcher.py
+++ b/gcp_bucket_watcher.py
@@ -128,7 +128,12 @@ def sync_cycle(args: argparse.Namespace, client: storage.Client) -> int:
             if should_skip(rel, includes, excludes):
                 continue
             if needs_transfer(meta, remote.get(rel), tolerance=2.0):
-                upload_file(client, args.bucket, args.prefix, meta, args.dest)
+                try:
+                    upload_file(client, args.bucket, args.prefix, meta, args.dest)
+                except FileNotFoundError:
+                    # File disappeared between scan and upload attempt; skip gracefully.
+                    print(f"[skipped missing] {rel}")
+                    continue
                 print(f"[uploaded] {rel} ({meta.size} bytes)")
                 changed += 1
         if changed == 0: