ggml_build_backward_expand: Add the Backward Pass to a Computation Graph

In ggmlR: 'GGML' Tensor Operations for Machine Learning

Add the Backward Pass to a Computation Graph

Description

Extends graph in place with the nodes that compute gradients of its output with respect to every tensor marked by ggml_set_param. After computing the graph, read a gradient with ggml_graph_get_grad.

Usage

ggml_build_backward_expand(ctx, graph)

Arguments

ctx	GGML context
graph	Graph from ggml_build_forward_expand_grads

Details

The graph must come from ggml_build_forward_expand_grads: ggml_build_forward_expand produces one without gradient storage, and that is rejected rather than allowed to abort inside ggml.

Value

NULL, invisibly; graph is modified in place

The full sequence

Every step is required; skipping one gives either an error or – for the last – gradients that are silently all zero.

1. ggml_set_param on each tensor to differentiate

2. build the forward ops

3. ggml_set_loss on the output

4. ggml_build_forward_expand_grads

5. ggml_build_backward_expand()

6. ggml_graph_reset – seeds d(loss)/d(loss) = 1

7. ggml_graph_compute

8. ggml_graph_get_grad to read a gradient

Which ops are differentiable

Gradients flow only through operations that implement a backward pass. Most do, but not all – ggml_clamp is a notable exception, and among the state-space ops only ggml_ssm_conv is differentiable.

See Also

ggml_build_forward_expand_grads, ggml_graph_get_grad, ggml_set_param

ggmlR documentation built on Oct. 23, 2026, 1:07 a.m.