Model Signing

This project demonstrates how to protect the integrity of a model by signing it with Sigstore, a tool for making code signatures transparent without requiring management of cryptographic key material.

When users download a given version of a signed model they can check that the signature comes from a known or trusted identity and thus that the model hasn't been tampered with after training.

Signing events are recorded to Sigstore's append-only transparency log. Transparency logs make signing events discoverable: Model verifiers can validate that the models they are looking at exist in the transparency log by checking a proof of inclusion (which is handled by the model signing library). Furthermore, model signers that monitor the log can check for any unexpected signing events.

Model signers should monitor for occurences of their signing identity in the log. Sigstore is actively developing a log monitor that runs on GitHub Actions.

Model Signing CLI

The sign.py and verify.py scripts aim to provide the necessary functionality to sign and verify ML models. For signing and verification the following methods are supported:

Bring your own key pair
Bring your own PKI
Skip signing (only hash and create a bundle)

The signing part creates a sigstore bundle protobuf that is stored as in JSON format. The bundle contains the verification material necessary to check the payload and a payload as a DSSE envelope. Further the DSSE envelope contains an in-toto statment and the signature over that statement. The signature format and how the the signature is computed can be seen here.

Finally, the statement itself contains subjects which are a list of (file path, digest) pairs a predicate type set to model_signing/v1/modeland a dictionary f predicates. The idea is to use the predicates to store (and therefor sign) model card information in the future.

The verification part reads the sigstore bundle file and firstly verifies that the signature is valid and secondly compute the model's file hashes again to compare against the signed ones.

Note: The signature is stored as ./model.sig by default and can be adjusted by setting the --sig_out flag.

Usage

There are two scripts one can be used to create and sign a bundle and the other to verify a bundle. Furthermore, the functionality can be used directly from other Python tools. The sign.py and verify.py scripts can be used as canonical how-to examples.

The easiest way to use the scripts directly is from a virtual environment:

$ python3 -m venv .venv
$ source .venv/bin/activate
(.venv) $ pip install -r install/requirements.in

Sign

(.venv) $ python3 sign.py --model_path ${MODEL_PATH} --sig_out ${OUTPUT_PATH} --method {private-key, pki} {additional parameters depending on method}

Verify

(.venv) $ python3 verify.py --model_path ${MODEL_PATH} --method {private-key, pki} {additional parameters depending on method}

Examples

Bring Your Own Key

$ MODEL_PATH='/path/to/your/model'
$ openssl ecparam -name secp256k1 -genkey -noout -out ec-secp256k1-priv-key.pem
$ openssl ec -in ec-secp256k1-priv-key.pem -pubout > ec-secp256k1-pub-key.pem
$ source .venv/bin/activate
# SIGN
(.venv) $ python3 sign_model.py --model_path ${MODEL_PATH} --method private-key --private-key ec-secp256k1-priv-key.pem
...
#VERIFY
(.venv) $ python3 verify_model.py --model_path ${MODEL_PATH} --method private-key --public-key ec-secp256k1-pub-key.pem
...

Bring your own PKI

In order to sign a model with your own PKI you need to create the following information:

- The signing certificate
- The elliptic curve private key matching the signing certificate's public key
- Optionally, the certificate chain used for verification.

$ MODEL_PATH='/path/to/your/model'
$ CERT_CHAIN='/path/to/cert_chain'
$ SIGNING_CERT='/path/to/signing_certificate'
$ PRIVATE_KEY='/path/to/private_key'
# SIGN
(.venv) $ python3 sign_model.py --model_path ${MODEL_PATH} \
    --method pki \
    --private-key ${PRIVATE_KEY} \
    --signing_cert ${SIGNING_CERT} \
    [--cert_chain ${CERT_CHAIN}]
...
#VERIFY
$ ROOT_CERTS='/path/to/root/certs'
(.venv) $ python3 verify_model.py --model_path ${MODEL_PATH} \
     --method pki \
     --root_certs ${ROOT_CERTS}
...

Sigstore ID providers

For developers signing models, there are three identity providers that can be used at the moment:

Google's provider is https://accounts.google.com.
GitHub's provider is https://github.com/login/oauth.
Microsoft's provider is https://login.microsoftonline.com.

For automated signing using a workload identity, the following platforms are currently supported, shown with their expected identities:

GitHub Actions (https://github.com/octo-org/octo-automation/.github/workflows/oidc.yml@refs/heads/main)
GitLab CI (https://gitlab.com/my-group/my-project//path/to/.gitlab-ci.yml@refs/heads/main)
Google Cloud Platform (SERVICE_ACCOUNT_NAME@PROJECT_ID.iam.gserviceaccount.com)
Buildkite CI (https://buildkite.com/ORGANIZATION_SLUG/PIPELINE_SLUG)

Supported Models

The library supports multiple models, from multiple training frameworks and model hubs.

For example, to sign and verify a Bertseq2seq model, trained with TensorFlow, stored in TFHub, run the following commands:

model_path=bertseq2seq
wget "https://tfhub.dev/google/bertseq2seq/bert24_en_de/1?tf-hub-format=compressed" -O "${model_path}".tgz
mkdir -p "${model_path}"
cd "${model_path}" && tar xvzf ../"${model_path}".tgz && rm ../"${model_path}".tgz && cd -
python3 main.py sign --path "${model_path}"
python3 main.py verify --path "${model_path}" \
    --identity-provider https://accounts.google.com \
    --identity myemail@gmail.com

For models stored in Hugging Face we need the large file support from git, which can be obtained via

sudo apt install git-lfs
git lfs install

After this, we can sign and verify a Bert base model:

model_name=bert-base-uncased
model_path="${model_name}"
git clone --depth=1 "https://huggingface.co/${model_name}" && rm -rf "${model_name}"/.git
python3 main.py sign --path "${model_path}"
python3 main.py verify --path "${model_path}" \
    --identity-provider https://accounts.google.com \
    --identity myemail@gmail.com

Similarly, we can sign and verify a Falcon model:

model_name=tiiuae/falcon-7b
model_path=$(echo "${model_name}" | cut -d/ -f2)
git clone --depth=1 "https://huggingface.co/${model_name}" && rm -rf "${model_name}"/.git
python3 main.py sign --path "${model_path}"
python3 main.py verify --path "${model_path}" \
    --identity-provider https://accounts.google.com \
    --identity myemail@gmail.com

We can also support models from the PyTorch Hub:

model_name=hustvl/YOLOP
model_path=$(echo "${model_name}" | cut -d/ -f2)
wget "https://github.com/${model_name}/archive/main.zip" -O "${model_path}".zip
mkdir -p "${model_path}"
cd "${model_path}" && unzip ../"${model_path}".zip && rm ../"${model_path}".zip && shopt -s dotglob && mv YOLOP-main/* . && shopt -u dotglob && rmdir YOLOP-main/ && cd -
python3 main.py sign --path "${model_path}"
python3 main.py verify --path "${model_path}" \
    --identity-provider https://accounts.google.com \
    --identity myemail@gmail.com

We also support ONNX models, for example Roberta:

model_name=roberta-base-11
model_path="${model_name}.onnx"
wget "https://github.com/onnx/models/raw/main/text/machine_comprehension/roberta/model/${model_name}.onnx"
python3 main.py sign --path "${model_path}"
python3 main.py verify --path "${model_path}" \
    --identity-provider https://accounts.google.com \
    --identity myemail@gmail.com

Benchmarking

Install as per Usage section. Ensure you have enough disk space:

if passing 3rd script argument as true: at least 50GB
otherwise: at least 100GB

To run the benchmarks:

git clone git@github.com:sigstore/model-transparency.git
cd model-transparency/model_signing
bash benchmarks/run.sh https://accounts.google.com myemail@gmail.com [true]

A single run was performed.

Hashes used:

H1: Hashing using a tree representation of the directory.
H2: Hashing using a list representation of the directory. (Implementation is parallized with shards of 1GB sizes across vCPUs).

Machine M1: Debian 6.3.11 x86_64 GNU/Linux, 200GB RAM, 48 vCPUs, 512KB cache, AMD EPYC 7B12:

Hash	Model	Size	Sign Time	Verify Time
H1	roberta-base-11	8K	0.8s	0.6s
H1	hustvl/YOLOP	215M	1.2s	0.8s
H1	bertseq2seq	2.8G	4.6s	4.4s
H1	bert-base-uncased	3.3G	5s	4.7s
H1	tiiuae/falcon-7b	14GB	12.2s	11.8s
H2	roberta-base-11	8K	1s	0.6s
H2	hustvl/YOLOP	215M	1s	1s
H2	bertseq2seq	2.8G	1.9s	1.4s
H2	bert-base-uncased	3.3G	1.6s	1.1s
H2	tiiuae/falcon-7b	14GB	2.1s	1.8s

Machine M2: Debian 5.10.1 x86_64 GNU/Linux, 4GB RAM, 2 vCPUs, 56320 KB, Intel(R) Xeon(R) CPU @ 2.20GHz:

Hash	Model	Size	Sign Time	Verify Time
H1	roberta-base-11	8K	1.1s	0.7s
H1	hustvl/YOLOP	215M	1.9s	1.7s
H1	bertseq2seq	2.8G	18s	23.2s
H1	bert-base-uncased	3.3G	23.4s	18.9s
H1	tiiuae/falcon-7b	14GB	2m4s	2m2s
H2	roberta-base-11	8K	1.1s	0.8s
H2	hustvl/YOLOP	215M	1.9s	1.6s
H2	bertseq2seq	2.8G	13.8s	25.9s
H2	bert-base-uncased	3.3G	22.7s	23.3s
H2	tiiuae/falcon-7b	14GB	2m.1s	2m3s

Model Signing Format

For a diagram showing the model signing format as well as an explanation of the layers, see the model signing format document.

Development steps

Linting

model_signing is automatically linted and formatted with a collection of tools:

flake8
pytype

You can run the type checker locally by installing the dev dependencies:

python3 -m venv dev_env
source dev_env/bin/activate
os=Linux # Supported: Linux, Darwin.
python3 -m pip install --require-hashes -r "install/requirements_dev_${os}".txt

Then point pytype at the desired module or package:

pytype --keep-going model_signing/hashing

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

README.model_signing.md

README.model_signing.md

Model Signing

Model Signing CLI

Usage

Sign

Verify

Examples

Bring Your Own Key

Bring your own PKI

Sigstore ID providers

Supported Models

Benchmarking

Model Signing Format

Development steps

Linting

Files

README.model_signing.md

Latest commit

History

README.model_signing.md

File metadata and controls

Model Signing

Model Signing CLI

Usage

Sign

Verify

Examples

Bring Your Own Key

Bring your own PKI

Sigstore ID providers

Supported Models

Benchmarking

Model Signing Format

Development steps

Linting