Instructions to use hacnho/tf-savedmodel-assetfiledef-dos-poc with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- TF-Keras
How to use hacnho/tf-savedmodel-assetfiledef-dos-poc with TF-Keras:
# Note: 'keras<3.x' or 'tf_keras' must be installed (legacy) # See https://github.com/keras-team/tf-keras for more details. from huggingface_hub import from_pretrained_keras model = from_pretrained_keras("hacnho/tf-savedmodel-assetfiledef-dos-poc") - Notebooks
- Google Colab
- Kaggle
TensorFlow SavedModel asset_file_def cardinality amplification PoC
This repository hosts a benign security research PoC for a TensorFlow SavedModel load-time denial of service.
Files
saved_model.pbvariables/variables.indexvariables/variables.data-00000-of-00001fingerprint.pb
Reproduction
OPENBLAS_NUM_THREADS=1 python3 - <<'PY'
import tensorflow as tf
m = tf.saved_model.load(".")
print(type(m).__name__)
PY
Observed on the research machine:
saved_model.pbis about70.9 MB- it declares
3,000,000asset_file_defentries tf.saved_model.load()still succeeds but peak RSS reaches roughly1.24 GB
Safety note
This artifact is a non-executable malformed model-file PoC for parser-stress testing only.
- Downloads last month
- 5
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support