clean up modelframe

selfdrive/modeld/modeld.py

@@ -56,7 +56,7 @@ class ModelState:
   prev_desire: np.ndarray  # for tracking the rising edge of the pulse
 
   def __init__(self, context: CLContext):
-    self.frames = {'input_imgs': DrivingModelFrame(context), 'big_input_imgs': DrivingModelFrame(context)}
+    self.frames = {'input_imgs': DrivingModelFrame(context, ModelConstants.TEMPORAL_SKIP), 'big_input_imgs': DrivingModelFrame(context, ModelConstants.TEMPORAL_SKIP)}
     self.prev_desire = np.zeros(ModelConstants.DESIRE_LEN, dtype=np.float32)
 
     self.full_features_buffer = np.zeros((1, ModelConstants.FULL_HISTORY_BUFFER_LEN,  ModelConstants.FEATURE_LEN), dtype=np.float32)

selfdrive/modeld/models/commonmodel.cc

@@ -5,11 +5,11 @@
 
 #include "common/clutil.h"
 
-DrivingModelFrame::DrivingModelFrame(cl_device_id device_id, cl_context context) : ModelFrame(device_id, context) {
+DrivingModelFrame::DrivingModelFrame(cl_device_id device_id, cl_context context, int temporal_skip) : ModelFrame(device_id, context) {
   input_frames = std::make_unique<uint8_t[]>(buf_size);
   input_frames_cl = CL_CHECK_ERR(clCreateBuffer(context, CL_MEM_READ_WRITE, buf_size, NULL, &err));
-  img_buffer_20hz_cl = CL_CHECK_ERR(clCreateBuffer(context, CL_MEM_READ_WRITE, 5*frame_size_bytes, NULL, &err));
-  region.origin = 4 * frame_size_bytes;
+  img_buffer_20hz_cl = CL_CHECK_ERR(clCreateBuffer(context, CL_MEM_READ_WRITE, (temporal_skip+1)*frame_size_bytes, NULL, &err));
+  region.origin = temporal_skip * frame_size_bytes;
   region.size = frame_size_bytes;
   last_img_cl = CL_CHECK_ERR(clCreateSubBuffer(img_buffer_20hz_cl, CL_MEM_READ_WRITE, CL_BUFFER_CREATE_TYPE_REGION, &region, &err));
 
@@ -20,7 +20,7 @@ DrivingModelFrame::DrivingModelFrame(cl_device_id device_id, cl_context context)
 cl_mem* DrivingModelFrame::prepare(cl_mem yuv_cl, int frame_width, int frame_height, int frame_stride, int frame_uv_offset, const mat3& projection) {
   run_transform(yuv_cl, MODEL_WIDTH, MODEL_HEIGHT, frame_width, frame_height, frame_stride, frame_uv_offset, projection);
 
-  for (int i = 0; i < 4; i++) {
+  for (int i = 0; i < temporal_skip; i++) {
     CL_CHECK(clEnqueueCopyBuffer(q, img_buffer_20hz_cl, img_buffer_20hz_cl, (i+1)*frame_size_bytes, i*frame_size_bytes, frame_size_bytes, 0, nullptr, nullptr));
   }
   loadyuv_queue(&loadyuv, q, y_cl, u_cl, v_cl, last_img_cl);

selfdrive/modeld/models/commonmodel.h

@@ -64,14 +64,14 @@ protected:
 
 class DrivingModelFrame : public ModelFrame {
 public:
-  DrivingModelFrame(cl_device_id device_id, cl_context context);
+  DrivingModelFrame(cl_device_id device_id, cl_context context, int temporal_skip);
   ~DrivingModelFrame();
   cl_mem* prepare(cl_mem yuv_cl, int frame_width, int frame_height, int frame_stride, int frame_uv_offset, const mat3& projection);
 
   const int MODEL_WIDTH = 512;
   const int MODEL_HEIGHT = 256;
   const int MODEL_FRAME_SIZE = MODEL_WIDTH * MODEL_HEIGHT * 3 / 2;
-  const int buf_size = MODEL_FRAME_SIZE * 2;
+  const int buf_size = MODEL_FRAME_SIZE * 2; // 2 frames are temporal_skip frames apart
   const size_t frame_size_bytes = MODEL_FRAME_SIZE * sizeof(uint8_t);
 
 private:

selfdrive/modeld/models/commonmodel.pxd

@@ -20,7 +20,7 @@ cdef extern from "selfdrive/modeld/models/commonmodel.h":
 
   cppclass DrivingModelFrame:
     int buf_size
-    DrivingModelFrame(cl_device_id, cl_context)
+    DrivingModelFrame(cl_device_id, cl_context, int)
 
   cppclass MonitoringModelFrame:
     int buf_size

selfdrive/modeld/models/commonmodel_pyx.pyx

@@ -59,8 +59,8 @@ cdef class ModelFrame:
 cdef class DrivingModelFrame(ModelFrame):
   cdef cppDrivingModelFrame * _frame
 
-  def __cinit__(self, CLContext context):
-    self._frame = new cppDrivingModelFrame(context.device_id, context.context)
+  def __cinit__(self, CLContext context, int temporal_skip):
+    self._frame = new cppDrivingModelFrame(context.device_id, context.context, temporal_skip)
     self.frame = <cppModelFrame*>(self._frame)
     self.buf_size = self._frame.buf_size