use nms in the model itself, simplify and make postprocess faster.

make inference in batches, fix c# handling, add overlap handling
This commit is contained in:
Alex Bezdieniezhnykh
2025-02-10 14:55:00 +02:00
parent ba3e3b4a55
commit c1b5b5fee2
19 changed files with 259 additions and 140 deletions
+2 -1
View File
@@ -9,4 +9,5 @@ cdef str TOKEN_FILE # Name of the token file where temporary token wo
cdef str QUEUE_CONFIG_FILENAME # queue config filename to load from api
cdef str AI_MODEL_FILE # AI Model file
cdef bytes DONE_SIGNAL
cdef bytes DONE_SIGNAL
cdef int MODEL_BATCH_SIZE