administrator/frigate
1
0
Fork 0
mirror of https://github.com/blakeblackshear/frigate.git synced 2026-02-09 12:45:25 +03:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

configure number of cores rather than a core mask for better user experience

Browse Source
This commit is contained in:
MarcA711 2024-03-23 19:44:10 +00:00
parent 4c8c196c32
commit 0f07c2bdc8
2 changed files with 7 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
docs/docs/configuration/object_detectors.md
View File

@ -406,8 +406,11 @@ This `config.yml` shows all relevant options to configure the detector and expla
detectors: # required detectors: # required
rknn: # required rknn: # required
type: rknn # required type: rknn # required
# core mask for npu # number of NPU cores used for detection
core_mask: 0b111 # 0 means automatic selection, 1 is one core etc
# only relevant for SoCs with multicore NPU like rk3588/s
# increase to 3 on rk3588 SoC for best performance
num_cores: 0
model: # required model: # required
# name of yolov8 model or path to your own .rknn model file # name of yolov8 model or path to your own .rknn model file
@ -429,13 +432,6 @@ model: # required
input_tensor: nhwc input_tensor: nhwc
``` ```
Explanation for rknn specific options:
- **core mask** controls which cores of your NPU should be used. This option applies only to SoCs with a multicore NPU (at the time of writing this in only the RK3588/S). The easiest way is to pass the value as a binary number. To do so, use the prefix `0b` and write a `0` to disable a core and a `1` to enable a core, whereas the last digit corresponds to core0, the second last to core1, etc. You also have to use the cores in ascending order (so you can't use core0 and core2; but you can use core0 and core1). Enabling more cores can reduce the inference speed, especially when using bigger models (see section below). Examples:
- `core_mask: 0b000` or just `core_mask: 0` let the NPU decide which cores should be used.
- `core_mask: 0b001` use only core0.
- `core_mask: 0b111` use core0, core1 and core2. Fastest and default value.
- `core_mask: 0b110` use core1 and core2. **This does not** work, since core0 is disabled.
### Choosing a model ### Choosing a model

4
frigate/detectors/plugins/rknn.py
View File

@ -18,7 +18,7 @@ supported_socs = ["rk3562", "rk3566", "rk3568", "rk3588"]
class RknnDetectorConfig(BaseDetectorConfig): class RknnDetectorConfig(BaseDetectorConfig):
type: Literal[DETECTOR_KEY] type: Literal[DETECTOR_KEY]
core_mask: int = Field(default=7, ge=0, le=7, title="Core mask for NPU.") num_cores: int = Field(default=0, ge=0, le=3, title="Number of NPU cores.")
class Rknn(DetectionApi): class Rknn(DetectionApi):
@ -29,7 +29,7 @@ class Rknn(DetectionApi):
self.width = config.model.width self.width = config.model.width
soc = self.get_soc() soc = self.get_soc()
core_mask = config.core_mask core_mask = 2**config.num_cores - 1
model_properties = self.get_model_properties( model_properties = self.get_model_properties(
config.model.path or "default-yolov8n", soc config.model.path or "default-yolov8n", soc