Commit
•
14756bb
1
Parent(s):
116c902
Update makefile
Browse files
Makefile
CHANGED
@@ -18,7 +18,8 @@ trt:
|
|
18 |
"trtexec --onnx=onnx/model.onnx --saveEngine=$(MODEL_PATH)/model.plan \
|
19 |
--minShapes=input:1x3x560x560 \
|
20 |
--optShapes=input:2x3x560x560 \
|
21 |
-
--maxShapes=input:8x3x560x560
|
|
|
22 |
|
23 |
|
24 |
triton:
|
@@ -26,9 +27,9 @@ triton:
|
|
26 |
--name dinov2_vitl14_triton \
|
27 |
-p 8000:8000 --gpus "device=0" -p 8001:8001 -p 8002:8002 \
|
28 |
-v $(PWD)/model_repository:/models \
|
29 |
-
nvcr.io/nvidia/tritonserver:23.
|
30 |
tritonserver --model-repository=/models
|
31 |
|
32 |
perf:
|
33 |
-
docker run --gpus all --rm -it --net host nvcr.io/nvidia/tritonserver:23.
|
34 |
|
|
|
18 |
"trtexec --onnx=onnx/model.onnx --saveEngine=$(MODEL_PATH)/model.plan \
|
19 |
--minShapes=input:1x3x560x560 \
|
20 |
--optShapes=input:2x3x560x560 \
|
21 |
+
--maxShapes=input:8x3x560x560 \
|
22 |
+
--fp16"
|
23 |
|
24 |
|
25 |
triton:
|
|
|
27 |
--name dinov2_vitl14_triton \
|
28 |
-p 8000:8000 --gpus "device=0" -p 8001:8001 -p 8002:8002 \
|
29 |
-v $(PWD)/model_repository:/models \
|
30 |
+
nvcr.io/nvidia/tritonserver:23.06-py3 \
|
31 |
tritonserver --model-repository=/models
|
32 |
|
33 |
perf:
|
34 |
+
docker run --gpus all --rm -it --net host nvcr.io/nvidia/tritonserver:23.06-py3-sdk perf_analyzer -m dinov2_vitl14 --percentile=95 -i grpc -u 0.0.0.0:8001 --concurrency-range 16:16 --shape input:3,560,560
|
35 |
|