ResNet34-416x640 (ONNX) – Renesas X5H
⚠️ Task classification needs confirmation. The source checkpoint name (
resnet34_1_3_416_640, i.e. an NCHW input shape of[1, 3, 416, 640]) uses a non-square input resolution. Plain ImageNet-style classification checkpoints are almost always square (e.g. 224×224); a non-square 416×640 input strongly suggests this is a ResNet34 backbone for an object-detection pipeline rather than a standalone classifier. This card defaults the task toobject-detection— confirm the intended task/checkpoint before publishing.
Introduction
This repository hosts a ResNet34 model at a non-square 416×640 input resolution, targeting the Renesas R-Car X5H platform for inference on the NPX6 NPU.
- Model Architecture: ResNet34 (likely used as a detection backbone — see caveat above)
- Source Model: internal export
resnet34_1_3_416_640(no HuggingFace mirror of these weights; seemodel.sourcein.metadata.yaml) - Task: Object Detection (unconfirmed — see caveat above)
- Input Resolution: 416 × 640 (inferred from the checkpoint name)
- Parameters: not published — count them from the ONNX graph (
sum(numpy_helper.to_array(t).size for t in model.graph.initializer))
Deployment Flow
The FP32 ONNX model is auto-cast to INT8 by the Renesas MWMX toolchain at compile time — no separate quantization step is required.
resnet34_1_3_416_640.onnx (FP32)
│
└─▶ MWMX Runtime ──▶ INT8 auto-cast ──▶ NPX6 NPU
Provided Artifacts
| Artifact | Status | Notes |
|---|---|---|
| FP32 (ONNX) | ✅ | fp32/resnet34_1_3_416_640.onnx — FP32 ONNX export |
Performance
Measured on Renesas R-Car X5H via the MWMX runtime (APM50 ship-performance CI pipeline).
Benchmark configuration: Single NPU · Batch size: 1 · Input resolution: 3 × 416 × 640 (inferred)
The 1-AI-core slice failed in the source CI pipeline, so only the 12-core result is available.
| Runtime | Precision | Device | Latency (ms) | Type |
|---|---|---|---|---|
| MWMX Runtime | INT8 (auto) | X5H · 1× NPU · 12 Cores · 850 MHz | 1.370561 | Measured |
Accuracy
TBD — not yet measured/published for this repo.
Runtime Details
MWMX Runtime
- Engine: Renesas MWMX (Middleware MX) native inference runtime
- Input format: FP32 ONNX (compiled by the MWMX toolchain)
- NPU execution precision: INT8 (auto-cast by MWMX toolchain)
- Execution target: NPX6-48K NPU on R-Car X5H
Prerequisites
To run inference on Renesas R-Car X5H, you need:
- Renesas R-Car X5H board with NPX6 NPU
- Renesas MWMX Runtime
- Hugging Face CLI to download the model
Download
hf download Renesas/ResNet34-416x640-ONNX --repo-type=model --include "fp32/*"
Benchmark Methodology
- HIL runs: Hardware-in-the-loop — measured on physical R-Car X5H silicon via the MWMX
runtime (
metawaremx_runtimeCI pipeline, "APM50" ship-performance target) - Precision: FP32 ONNX input; INT8 execution (auto-cast by MWMX)
- Slices: only the 12-AI-core result is available (1-core run failed in the source CI pipeline)