Hugging Face's logo

AXERA-TECH
/
bevformer

Object Detection
ONNX
English
ONNX
Axera
NPU
Pulsar2
BEVFormer
Autonomous-Driving
Bird-Eye-View
Model card Files
xet
Community
bevformer
File size: 2,201 Bytes 
e2845fc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c8a2d3c
e2845fc
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
 
{
  "output_name": "bevformer_tiny.axmodel",
  "model_type": "ONNX",
  "target_hardware": "AX650",
  "npu_mode": "NPU3",
  "quant": {
    "input_configs": [
      {
        "tensor_name": "img",
        "calibration_dataset": "./calibration/calibration_img.tar.gz",
        "calibration_format": "Numpy",
        "calibration_size": 64,
        "calibration_mean": [0, 0, 0],
        "calibration_std": [1, 1, 1]
      },
      {
        "tensor_name": "lidar2img",
        "calibration_dataset": "./calibration/calibration_lidar2img.tar.gz",
        "calibration_format": "Numpy",
        "calibration_size": 64
      },
      {
        "tensor_name": "can_bus",
        "calibration_dataset": "./calibration/calibration_can_bus.tar.gz",
        "calibration_format": "Numpy",
        "calibration_size": 64
      },
      {
        "tensor_name": "prev_bev",
        "calibration_dataset": "./calibration/calibration_prev_bev.tar.gz",
        "calibration_format": "Numpy",
        "calibration_size": 64
      }
    ],
    "calibration_method": "MSE",
    "precision_analysis": true,
    "precision_analysis_method": "EndToEnd",
    "enable_smooth_quant": true,
    "smooth_quant_threshold": 0.15,
    "smooth_quant_strength": 0.5,
    "refine_weight_threshold": 1e-6,
    "conv_bias_data_type": "FP32",
    "ln_scale_data_type": "FP32",
    "layer_configs": [
      {
        "layer_name": "/pts_bbox_head/transformer/encoder/Reshape_10",
        "data_type": "FP32"
      },
      {
        "layer_name": "/pts_bbox_head/transformer/encoder/Tile_3",
        "data_type": "FP32"
      },
      {
        "layer_name": "/pts_bbox_head/transformer/encoder/MatMul",
        "data_type": "FP32",
        "output_data_type": "FP32"
      },
      {
        "layer_name": "/pts_bbox_head/transformer/encoder/Squeeze",
        "data_type": "FP32"
      },
      {
        "layer_name": "/pts_bbox_head/transformer/encoder/Slice_10",
        "data_type": "FP32"
      },
      {
        "layer_name": "/pts_bbox_head/transformer/encoder/Slice_9",
        "data_type": "FP32"
      },
      {
        "layer_name": "/pts_bbox_head/transformer/encoder/Div_2",
        "data_type": "FP32"
      }
    ]
  }
}