NVIDIA DeepStream SDK API Reference

6.4 Release
GstNvInferBatch Struct Reference

Detailed Description

Holds information about the batch of frames to be inferred.

Definition at line 93 of file gstnvinfer_impl.h.

Data Fields

std::vector< GstNvInferFrameframes
 Vector of frames in the batch. More...
 
GstBufferinbuf = nullptr
 Pointer to the input GstBuffer. More...
 
gulong inbuf_batch_num = 0
 Batch number of the input batch. More...
 
gboolean push_buffer = FALSE
 Boolean indicating that the output thread should only push the buffer to downstream element. More...
 
gboolean event_marker = FALSE
 Boolean marking this batch as an event marker. More...
 
GstBufferconv_buf = nullptr
 Buffer containing the intermediate conversion output for the batch. More...
 
nvtxRangeId_t nvtx_complete_buf_range = 0
 
NvBufSurfTransformSyncObj_t sync_obj = NULL
 Sync object for allowing asynchronous call to nvbufsurftransform API Wait and Destroy to be done before preprocess call of nvinfer. More...
 
std::vector< GstNvInferObjHistory_MetaPairobjs_pending_meta_attach
 List of objects not inferred on in the current batch but pending attachment of lastest available classification metadata. More...
 

Field Documentation

◆ conv_buf

GstBuffer* GstNvInferBatch::conv_buf = nullptr

Buffer containing the intermediate conversion output for the batch.

Definition at line 110 of file gstnvinfer_impl.h.

◆ event_marker

gboolean GstNvInferBatch::event_marker = FALSE

Boolean marking this batch as an event marker.

This is only used for synchronization. The output loop does not process on the batch.

Definition at line 108 of file gstnvinfer_impl.h.

◆ frames

std::vector<GstNvInferFrame> GstNvInferBatch::frames

Vector of frames in the batch.

Definition at line 95 of file gstnvinfer_impl.h.

◆ inbuf

GstBuffer* GstNvInferBatch::inbuf = nullptr

Pointer to the input GstBuffer.

Definition at line 97 of file gstnvinfer_impl.h.

◆ inbuf_batch_num

gulong GstNvInferBatch::inbuf_batch_num = 0

Batch number of the input batch.

Definition at line 99 of file gstnvinfer_impl.h.

◆ nvtx_complete_buf_range

nvtxRangeId_t GstNvInferBatch::nvtx_complete_buf_range = 0

Definition at line 111 of file gstnvinfer_impl.h.

◆ objs_pending_meta_attach

std::vector<GstNvInferObjHistory_MetaPair> GstNvInferBatch::objs_pending_meta_attach

List of objects not inferred on in the current batch but pending attachment of lastest available classification metadata.

Definition at line 119 of file gstnvinfer_impl.h.

◆ push_buffer

gboolean GstNvInferBatch::push_buffer = FALSE

Boolean indicating that the output thread should only push the buffer to downstream element.

If set to true, a corresponding batch has not been queued at the input of NvDsInferContext and hence dequeuing of output is not required.

Definition at line 104 of file gstnvinfer_impl.h.

◆ sync_obj

NvBufSurfTransformSyncObj_t GstNvInferBatch::sync_obj = NULL

Sync object for allowing asynchronous call to nvbufsurftransform API Wait and Destroy to be done before preprocess call of nvinfer.

Definition at line 115 of file gstnvinfer_impl.h.


The documentation for this struct was generated from the following file: