embedl_deploy.tensorrt package#

Subpackages:

Module contents:

TensorRT backend — curated pattern lists and convenience API.

Quick start:

import torch
from torchvision.models import resnet50
from embedl_deploy import transform
from embedl_deploy.tensorrt import TENSORRT_PATTERNS

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

Pattern lists#

TENSORRT_CONVERSION_PATTERNS

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

TENSORRT_FUSION_PATTERNS

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

TENSORRT_PATTERNS

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

TENSORRT_QUANTIZED_PATTERNS

Quantized variants (placeholder — not yet implemented).