rtmo

3 contributors

History: 43 commits

Luigi

Deal with multiple cameras

98d66d4 5 months ago

video
Add a sample video 8 months ago
.gitattributes

1.66 kB

Add a sample video 8 months ago
README.md

1.47 kB

Update README.md 7 months ago
constant_folding.sh

337 Bytes

Apply constant folding 7 months ago
convert_to_fp16.py

1.25 kB

Re-quantize models in FP16 in keep positional encoding in FP32 to avoid bad accuracy 7 months ago
convert_to_fp16_all.py

1.36 kB

Bugfix: correct batch dimension fixation then regenerate ONNX models 7 months ago
convert_to_mixed.py

5.26 kB

Improve Debug Message Layout in Validate Funciton 7 months ago
demo.sh

56 Bytes

Remove models with fixed batch size as dynamic batch is generally supported 7 months ago
demo_batch.sh

71 Bytes

Unify batched and non-batched versions 5 months ago
onnx_to_engine.py

7.05 kB

Use dynamic batch size by default 5 months ago
rtmo-l.fp16.onnx
88 MB
LFS

Re-quantize models in FP16 in keep positional encoding in FP32 to avoid bad accuracy 7 months ago
rtmo-l.onnx
176 MB
LFS

Apply constant folding 7 months ago
rtmo-m.fp16.onnx
44.7 MB
LFS

Re-quantize models in FP16 in keep positional encoding in FP32 to avoid bad accuracy 7 months ago
rtmo-m.onnx
89.3 MB
LFS

Apply constant folding 7 months ago
rtmo-s.fp16.onnx
19.9 MB
LFS

Re-quantize models in FP16 in keep positional encoding in FP32 to avoid bad accuracy 7 months ago
rtmo-s.onnx
39.6 MB
LFS

Apply constant folding 7 months ago
rtmo-t.fp16.onnx
13.7 MB
LFS

Re-quantize models in FP16 in keep positional encoding in FP32 to avoid bad accuracy 7 months ago
rtmo-t.onnx
27.4 MB
LFS

Apply constant folding 7 months ago
rtmo_demo.py

2.63 kB

Show bounding box on screen too 5 months ago
rtmo_demo_batch.py

3.46 kB

Show bounding box on screen too 5 months ago
rtmo_gpu.py

24 kB

Deal with multiple cameras 5 months ago
symbolic_shape_infer.py

138 kB

Make ONNX models compatible with ONNXruntime's TensorrtExecutionProvider 7 months ago
symbolic_shape_infer_all.sh

135 Bytes

Make ONNX models compatible with ONNXruntime's TensorrtExecutionProvider 7 months ago