Closed
Bug 2047188
Opened 1 month ago
Closed 1 month ago
[ai-platform] Harden code around model loading
Categories
(Core :: Machine Learning: On Device, enhancement, P2)
Core
Machine Learning: On Device
Tracking
()
RESOLVED
DUPLICATE
of bug 2025207
People
(Reporter: vpollet, Unassigned)
References
(Blocks 1 open bug)
Details
(Whiteboard: [aiplatform])
In current state, models are fetched unsigned and always executed. This can be hardened a lot by signing models and making sure we only load trusted models.
|
Reporter | |
Updated•1 month ago
|
Summary: [llama.cpp] Harden code around model loading → [ai-platform] Harden code around model loading
|
||
Updated•1 month ago
|
Whiteboard: [aiplatform]
|
||
Updated•1 month ago
|
|
|
||
Comment 1•1 month ago
|
||
While the code in mozilla central will likely not be too complicated, we'll need to update Remote Settings as well to store the hash.
|
|
Reporter | |
Updated•1 month ago
|
You need to log in
before you can comment on or make changes to this bug.
Description
•