embedl_deploy.lattice package

Subpackages:

• embedl_deploy.lattice.modules package
• embedl_deploy.lattice.patterns package

Module contents:

Lattice backend — curated pattern lists and convenience API.

Quick start:

import torch
from torchvision.models import resnet50
from embedl_deploy import transform
from embedl_deploy.lattice import LATTICE_PATTERNS

model = resnet50(weights=None).eval()
deployed = transform(model, patterns=LATTICE_PATTERNS).model

Pattern lists

LATTICE_CONVERSION_PATTERNS

Structural conversions applied before fusion (e.g. Flatten→Linear → Conv1×1→Flatten).

LATTICE_FUSION_PATTERNS

Fusion-only patterns (Conv→BN→ReLU, Stem, residual, etc.).

LATTICE_PATTERNS

Union of conversions + fusions (the default for most users).

LATTICE_QUANTIZED_PATTERNS

Quantized variants (placeholder — not yet implemented).