Document verified local weight cleanup

chapter-04-efficient-and-scalable-vision-transformers/EXPERIMENT_REPORT.md

@@ -16,8 +16,8 @@ deployment cost. Just as importantly, the benchmark showed why parameter count,
 operations, latency, and throughput must be measured separately.
 
 > **Experiment status:** complete. All 13 training runs reached 20 epochs, all 24 inference cases
-> completed, the report artifacts were generated, and all 13 validation-selected checkpoints were
-> downloaded and verified locally.
+> completed, the report artifacts were generated, and all 13 validation-selected model states were
+> published to Hugging Face and verified before their local weight copies were removed.
 
 ---
 
@@ -525,8 +525,8 @@ The completed local result bundle is
 | [`model_profiles.csv`](results/model_profiles.csv) | Parameters, MACs, and unsupported tracing operators |
 | [`inference_benchmark.csv`](results/inference_benchmark.csv) | All 24 timing and memory cases |
 | [`report.md`](results/report.md) | Automatically generated compact report |
-| [`checkpoints`](checkpoints) | Exported model-only Safetensors files |
-| [`SHA256SUMS`](checkpoints/SHA256SUMS) | Integrity manifest for all local checkpoints |
+| [Hugging Face checkpoints](https://huggingface.co/Mayank022/Transformer-Vision-From-Image-Video-to-World-Models-and-Robotics-Book/tree/main/chapter-04-efficient-and-scalable-vision-transformers/checkpoints) | Thirteen public model-only Safetensors files |
+| [Hugging Face manifest](https://huggingface.co/Mayank022/Transformer-Vision-From-Image-Video-to-World-Models-and-Robotics-Book/blob/main/chapter-04-efficient-and-scalable-vision-transformers/manifest.json) | Model paths, metrics, source hashes, and exported hashes |
 
 ### Hugging Face distribution
 
@@ -543,14 +543,19 @@ Apache-2.0 and the Swin-T source reports MIT. The published inference files cont
 scheduler, gradient-scaler, or random-number-generator state.
 
 The Hugging Face checkpoint directory contains 13 model-only
-`model.safetensors` files totaling approximately 1.06 GiB. The larger source
-`best.pt` files remain in controlled local and Modal storage because they also contain optimizer,
-scheduler, scaler, and random-state data used for exact training resume.
+`model.safetensors` files totaling approximately 1.06 GiB. Their local source and
+temporary export copies were removed after public hash verification. Complete resumable `best.pt`
+and `latest.pt` artifacts remain on the Modal volume.
 
-To verify every downloaded model file:
+To download and verify every published model file:
 
 ```bash
-cd checkpoints
+hf download \
+  Mayank022/Transformer-Vision-From-Image-Video-to-World-Models-and-Robotics-Book \
+  --include "chapter-04-efficient-and-scalable-vision-transformers/checkpoints/*" \
+  --local-dir transformer-vision-book
+
+cd transformer-vision-book/chapter-04-efficient-and-scalable-vision-transformers/checkpoints
 shasum -a 256 -c SHA256SUMS
 ```
 
@@ -607,18 +612,23 @@ Ruff and mypy completed cleanly.
 
 ## 14. Loading a downloaded checkpoint
 
-The Hugging Face files contain model parameters only and use Safetensors. This example restores the
-validation-selected ViT on CPU after downloading the chapter folder:
+The Hugging Face files contain model parameters only and use Safetensors. This example downloads and
+restores the validation-selected ViT on CPU:
 
 ```python
-from pathlib import Path
-
+from huggingface_hub import hf_hub_download
 from safetensors.torch import load_file
 import timm
 
-weights_path = Path(
+repo_id = (
+    "Mayank022/"
+    "Transformer-Vision-From-Image-Video-to-World-Models-and-Robotics-Book"
+)
+filename = (
+    "chapter-04-efficient-and-scalable-vision-transformers/"
     "checkpoints/vit-s16/standard/seed-42/model.safetensors"
 )
+weights_path = hf_hub_download(repo_id=repo_id, filename=filename)
 state_dict = load_file(weights_path, device="cpu")
 
 model = timm.create_model(