FHE Op-graph design
The ONNX import section gave an overview of the conversion of a generic ONNX graph to an FHE-compatible Concrete ML op-graph. This section describes the implementation of operations in the Concrete ML op-graph and the way floating point can be used in some parts of the op-graphs through table lookup operations.
Float vs. quantized operations
Concrete, the underlying implementation of TFHE that powers Concrete ML, enables two types of operations on integers:
arithmetic operations: the addition of two encrypted values and multiplication of encrypted values with clear scalars. These are used, for example, in dot-products, matrix multiplication (linear layers), and convolution.
table lookup operations (TLU): using an encrypted value as an index, return the value of a lookup table at that index. This is implemented using Programmable Bootstrapping. This operation is used to perform any non-linear computation such as activation functions, quantization, and normalization.
Since machine learning models use floating point inputs and weights, they first need to be converted to integers using quantization.
Alternatively, it is possible to use a table lookup to avoid the quantization of the entire graph, by converting floating-point ONNX subgraphs into lambdas and computing their corresponding lookup tables to be evaluated directly in FHE. This operator-fusion technique only requires the input and output of the lambdas to be integers.
For example, in the following graph there is a single input, which must be an encrypted integer tensor. The following series of univariate functions is then fed into a matrix multiplication (MatMul) and fused into a single table lookup with integer inputs and outputs.
![](https://docs.zama.ai/~gitbook/image?url=https%3A%2F%2F2241057372-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252FTeFWUaIKrKiFj6iYvz8W%252Fuploads%252Fgit-blob-7c87d4e6489ff3b4d9268b1c962b12604382c66d%252Fimage_3.png%3Falt%3Dmedia&width=768&dpr=4&quality=100&sign=df67873e&sv=1)
ONNX operations
Concrete ML implements ONNX operations using Concrete, which can handle floating point operations, as long as they can be fused to an integer lookup table. The ONNX operations implementations are based on the QuantizedOp
class.
There are two modes of creation of a single table lookup for a chain of ONNX operations:
float mode: when the operation can be fused
mixed float/integer: when the ONNX operation needs to perform arithmetic operations
Thus, QuantizedOp
instances may need to quantize their inputs or the result of their computation, depending on their position in the graph.
The QuantizedOp
class provides a generic implementation of an ONNX operation, including the quantization of inputs and outputs, with the computation implemented in NumPy in ops_impl.py
. It is possible to picture the architecture of the QuantizedOp
as the following structure:
![](https://docs.zama.ai/~gitbook/image?url=https%3A%2F%2F2241057372-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252FTeFWUaIKrKiFj6iYvz8W%252Fuploads%252Fgit-blob-9bbcc71145bb785207d883be40d67e2be889daec%252Fimage_4.png%3Falt%3Dmedia&width=768&dpr=4&quality=100&sign=5247d79&sv=1)
This figure shows that the QuantizedOp
has a body that implements the computation of the operation, following the ONNX spec. The operation's body can take either integer or float inputs and can output float or integer values. Two quantizers are attached to the operation: one that takes float inputs and produces integer inputs and one that does the same for the output.
Operations that can fuse to a TLU
Depending on the position of the op in the graph and its inputs, the QuantizedOp
can be fully fused to a TLU.
![](https://docs.zama.ai/~gitbook/image?url=https%3A%2F%2F2241057372-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252FTeFWUaIKrKiFj6iYvz8W%252Fuploads%252Fgit-blob-fa2d7589017b8d5b5d04cdccda73816713e02fad%252Fimage_8.png%3Falt%3Dmedia&width=768&dpr=4&quality=100&sign=8db355e&sv=1)
Many ONNX ops are trivially univariate, as they multiply variable inputs with constants or apply univariate functions such as ReLU, Sigmoid, etc. This includes operations between the input and the MatMul in the graph above (subtraction, comparison, multiplication, etc. between inputs and constants).
Operations that work on integers
Operations, such as matrix multiplication of encrypted inputs with a constant matrix or convolution with constant weights, require that the encrypted inputs be integers. In this case, the input quantizer of the QuantizedOp
is applied. These types of operations are implemented with a class that derives from QuantizedOp
and implements q_impl
, such as QuantizedGemm
and QuantizedConv
.
![](https://docs.zama.ai/~gitbook/image?url=https%3A%2F%2F2241057372-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252FTeFWUaIKrKiFj6iYvz8W%252Fuploads%252Fgit-blob-4f63c6ab3123af6b9ec0b98faef7515d1e35a010%252Fimage_5.png%3Falt%3Dmedia&width=768&dpr=4&quality=100&sign=70f8f650&sv=1)
Operations that produce graph outputs
Finally, some operations produce graph outputs, which must be integers. These operations need to quantize their outputs as follows:
![](https://docs.zama.ai/~gitbook/image?url=https%3A%2F%2F2241057372-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252FTeFWUaIKrKiFj6iYvz8W%252Fuploads%252Fgit-blob-8f99d2af86758bfa01ec4cb383335365f8823b19%252Fimage_1.png%3Falt%3Dmedia&width=768&dpr=4&quality=100&sign=a30d87bf&sv=1)
![](https://docs.zama.ai/~gitbook/image?url=https%3A%2F%2F2241057372-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252FTeFWUaIKrKiFj6iYvz8W%252Fuploads%252Fgit-blob-344310db8c6b80184ba6a043672f91ea108aa009%252Fimage_9.png%3Falt%3Dmedia&width=768&dpr=4&quality=100&sign=99865f0a&sv=1)
The diagram above shows that both float ops and integer ops need to quantize their outputs to integers when placed at the end of the graph.
Putting it all together
To chain the operation types described above following the ONNX graph, Concrete ML constructs a function that calls the q_impl
of the QuantizedOp
instances in the graph in sequence, and uses Concrete to trace the execution and compile to FHE. Thus, in this chain of function calls, all groups of that instruction that operate in floating point will be fused to TLUs. In FHE, this lookup table is computed with a PBS.
![](https://docs.zama.ai/~gitbook/image?url=https%3A%2F%2F2241057372-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252FTeFWUaIKrKiFj6iYvz8W%252Fuploads%252Fgit-blob-94bdf2baa7e789a165cc1d53b5388c2437f178f6%252Fimage_6.png%3Falt%3Dmedia&width=768&dpr=4&quality=100&sign=c2f8fc88&sv=1)
The red contours show the groups of elementary Concrete instructions that will be converted to TLUs.
Note that the input is slightly different from the QuantizedOp
. Since the encrypted function takes integers as inputs, the input needs to be de-quantized first.
Implementing a QuantizedOp
QuantizedOp
QuantizedOp
is the base class for all ONNX-quantized operators. It abstracts away many things to allow easy implementation of new quantized ops.
Determining if the operation can be fused
The QuantizedOp
class exposes a function can_fuse
that:
helps to determine the type of implementation that will be traced.
determines whether operations further in the graph, that depend on the results of this operation, can fuse.
In most cases, ONNX ops have a single variable input and one or more constant inputs.
When the op implements element-wise operations between the inputs and constants (addition, subtract, multiplication, etc), the operation can be fused to a TLU. Thus, by default in QuantizedOp
, the can_fuse
function returns True
.
When the op implements operations that mix the various scalars in the input encrypted tensor, the operation cannot fuse, as table lookups are univariate. Thus, operations such as QuantizedGemm
and QuantizedConv
return False
in can_fuse
.
Some operations may be found in both settings above. A mechanism is implemented in Concrete ML to determine if the inputs of a QuantizedOp
are produced by a unique integer tensor. Therefore, the can_fuse
function of some QuantizedOp
types (addition, subtraction) will allow fusion to take place if both operands are produced by a unique integer tensor:
Case 1: A floating point version of the op is sufficient
You can check ops_impl.py
to see how some operations are implemented in NumPy. The declaration convention for these operations is as follows:
The required inputs should be positional arguments only before the
/
, which marks the limit of the positional arguments.The optional inputs should be positional or keyword arguments between the
/
and*
, which marks the limits of positional or keyword arguments.The operator attributes should be keyword arguments only after the
*
.
The proper use of positional/keyword arguments is required to allow the QuantizedOp
class to properly populate metadata automatically. It uses Python inspect modules and stores relevant information for each argument related to its positional/keyword status. This allows using the Concrete implementation as specifications for QuantizedOp
, which removes some data duplication and generates a single source of truth for QuantizedOp
and ONNX-NumPy implementations.
In that case (unless the quantized implementation requires special handling like QuantizedGemm
), you can just set _impl_for_op_named
to the name of the ONNX op for which the quantized class is implemented (this uses the mapping ONNX_OPS_TO_NUMPY_IMPL
in onnx_utils.py
to get the correct implementation).
Case 2: An integer implementation of the op is necessary
Providing an integer implementation requires sub-classing QuantizedOp
to create a new operation. This sub-class must override q_impl
in order to provide an integer implementation. QuantizedGemm
is an example of such a case where quantized matrix multiplication requires proper handling of scales and zero points. The q_impl
of that class reflects this.
In the body of q_impl
, you can use the _prepare_inputs_with_constants
function in order to obtain quantized integer values:
Here, prepared_inputs
will contain one or more QuantizedArray
, of which the qvalues
are the quantized integers.
Once the required integer processing code is implemented, the output of the q_impl
function must be implemented as a single QuantizedArray
. Most commonly, this is built using the de-quantized results of the processing done in q_impl
.
Case 3: Both a floating point and an integer implementation are necessary
In this case, in q_impl
you can check whether the current operation can be fused by calling self.can_fuse()
. You can then have both a floating-point and an integer implementation. The traced execution path will depend on can_fuse()
:
Last updated