I have trained my model on ocrnet resnet50 as backbone
with this character dict
0
1
2
3
4
5
6
7
8
9
a
b
c
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
s
t
u
v
w
x
y
z
/
Note : didn’t got any issue while exporting or training
but while deploying it and inferencing it with the tritonserver i get this that triton getting Error saying :
"(1536, 2048, 3)
new_buffer (1, 736, 1280, 3) 1280 736
[ERROR] Character dict is not compatible with OCRNet TRT engine.
[nvOCDR] batch images infer time: : 399 ms
[nvOCDR] Infer done, got 39 texts
"
How can i fix this issue ? @Morganh