visual.build_tensor_mermaid_chart¶

lucid.visual.build_tensor_mermaid_chart(tensor: Tensor, horizontal: bool = False, title: str | None = None, start_id: int | None = None, end_semicolons: bool = True, copy_to_clipboard: bool = False, use_class_defs: bool = True, op_fill: str = 'lightgreen', param_fill: str = 'plum', result_fill: str = 'lightcoral', leaf_fill: str = 'lightgray', grad_fill: str = 'lightblue', start_fill: str = 'gold', stroke_color: str = '#666', stroke_width_px: int = 1) → str¶

Generates a Mermaid flowchart diagram from a computation graph rooted at a Tensor. This is useful for inspecting how tensors and operations connect during forward computation and for embedding lightweight graphs in docs or notebooks.

Basic Example¶

import lucid
import lucid.nn.functional as F
from lucid.visual import build_tensor_mermaid_chart

x = lucid.random.rand(1, 3, 8, 8, requires_grad=True)
w = lucid.random.randn(4, 3, 3, 3, requires_grad=True)
b = lucid.random.randn(4, requires_grad=True)

out = F.conv2d(x, w, b, stride=1, padding=1)
chart = build_tensor_mermaid_chart(out, horizontal=True, copy_to_clipboard=True)
print(chart)

        flowchart LR;
t_4418730848["(1, 3, 8, 8)"];
t_4418525520["(4, 3, 3, 3)"];
op_4419339664(("conv_nd_kernel"));
t_4299709040["(1, 4, 8, 8)"];
t_4418526480["(4,)"];
op_4419340000(("_reshape_immediate"));
t_4418992160["(1, 4, 1, 1)"];
op_4419340336(("add"));
t_4418559376["(1, 4, 8, 8)"];
op_4419339664 --> t_4299709040;
t_4418730848 --> op_4419339664;
t_4418525520 --> op_4419339664;
op_4419340000 --> t_4418992160;
t_4418526480 --> op_4419340000;
op_4419340336 --> t_4418559376;
t_4299709040 --> op_4419340336;
t_4418992160 --> op_4419340336;
classDef op fill:lightgreen,stroke:#666,stroke-width:1px;
classDef param fill:plum,stroke:#666,stroke-width:1px;
classDef result fill:lightcoral,stroke:#666,stroke-width:1px;
classDef leaf fill:lightgray,stroke:#666,stroke-width:1px;
classDef grad fill:lightblue,stroke:#666,stroke-width:1px;
classDef start fill:gold,stroke:#666,stroke-width:1px;
class t_4418730848 grad;
class t_4418525520 grad;
class op_4419339664 op;
class t_4299709040 grad;
class t_4418526480 grad;
class op_4419340000 op;
class t_4418992160 grad;
class op_4419340336 op;
class t_4418559376 result;

Key Parameters¶

tensor: The root output tensor from which to traverse the computation graph.
horizontal: If True, the graph is left-to-right; default is top-down.
title: Optional Mermaid comment title (rendering depends on viewer).
start_id: Highlights a tensor node by ID (useful for pinpointing a specific tensor).
copy_to_clipboard: If True, copies the Mermaid text to the system clipboard.
use_class_defs and color args: Control node color coding for ops, params, outputs, leaves, grads, and start node.