You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I’m developing an SDK that uses the TensorFlow.js WASM backend. However, since this is an SDK, I want to avoid asking users to add wasm-unsafe-eval to their Content Security Policy (CSP).
Is there any way to use the WASM backend without requiring wasm-unsafe-eval?
Thanks in advance for your help!
The text was updated successfully, but these errors were encountered:
A Content Security Policy(CSP) issue might arise if you attempt to call external APIs from within a SDK, and these calls are restricted by the SDK. The WASM compiler may try to call dynamic compiler APIs, which can be restricted because of CSP.
Based on my understanding, there's currently no way to use the WASM backend without updating the CSP. This is because it ultimately involves calling dynamic compiler APIs.
Hi,
I’m developing an SDK that uses the TensorFlow.js WASM backend. However, since this is an SDK, I want to avoid asking users to add wasm-unsafe-eval to their Content Security Policy (CSP).
Is there any way to use the WASM backend without requiring wasm-unsafe-eval?
Thanks in advance for your help!
The text was updated successfully, but these errors were encountered: