Update tiny models for CvtForImageClassification
Browse files- config.json +102 -0
- model.safetensors +3 -0
- preprocessor_config.json +27 -0
config.json
ADDED
|
@@ -0,0 +1,102 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"CvtForImageClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_drop_rate": [
|
| 6 |
+
0.0,
|
| 7 |
+
0.0,
|
| 8 |
+
0.0
|
| 9 |
+
],
|
| 10 |
+
"cls_token": [
|
| 11 |
+
false,
|
| 12 |
+
false,
|
| 13 |
+
true
|
| 14 |
+
],
|
| 15 |
+
"depth": [
|
| 16 |
+
1,
|
| 17 |
+
2,
|
| 18 |
+
10
|
| 19 |
+
],
|
| 20 |
+
"drop_path_rate": [
|
| 21 |
+
0.0,
|
| 22 |
+
0.0,
|
| 23 |
+
0.1
|
| 24 |
+
],
|
| 25 |
+
"drop_rate": [
|
| 26 |
+
0.0,
|
| 27 |
+
0.0,
|
| 28 |
+
0.0
|
| 29 |
+
],
|
| 30 |
+
"dtype": "float32",
|
| 31 |
+
"embed_dim": [
|
| 32 |
+
16,
|
| 33 |
+
32,
|
| 34 |
+
48
|
| 35 |
+
],
|
| 36 |
+
"image_size": 64,
|
| 37 |
+
"initializer_range": 0.02,
|
| 38 |
+
"kernel_qkv": [
|
| 39 |
+
3,
|
| 40 |
+
3,
|
| 41 |
+
3
|
| 42 |
+
],
|
| 43 |
+
"layer_norm_eps": 1e-12,
|
| 44 |
+
"mlp_ratio": [
|
| 45 |
+
4.0,
|
| 46 |
+
4.0,
|
| 47 |
+
4.0
|
| 48 |
+
],
|
| 49 |
+
"model_type": "cvt",
|
| 50 |
+
"num_channels": 3,
|
| 51 |
+
"num_heads": [
|
| 52 |
+
1,
|
| 53 |
+
2,
|
| 54 |
+
3
|
| 55 |
+
],
|
| 56 |
+
"padding_kv": [
|
| 57 |
+
1,
|
| 58 |
+
1,
|
| 59 |
+
1
|
| 60 |
+
],
|
| 61 |
+
"padding_q": [
|
| 62 |
+
1,
|
| 63 |
+
1,
|
| 64 |
+
1
|
| 65 |
+
],
|
| 66 |
+
"patch_padding": [
|
| 67 |
+
2,
|
| 68 |
+
1,
|
| 69 |
+
1
|
| 70 |
+
],
|
| 71 |
+
"patch_sizes": [
|
| 72 |
+
7,
|
| 73 |
+
3,
|
| 74 |
+
3
|
| 75 |
+
],
|
| 76 |
+
"patch_stride": [
|
| 77 |
+
4,
|
| 78 |
+
2,
|
| 79 |
+
2
|
| 80 |
+
],
|
| 81 |
+
"qkv_bias": [
|
| 82 |
+
true,
|
| 83 |
+
true,
|
| 84 |
+
true
|
| 85 |
+
],
|
| 86 |
+
"qkv_projection_method": [
|
| 87 |
+
"dw_bn",
|
| 88 |
+
"dw_bn",
|
| 89 |
+
"dw_bn"
|
| 90 |
+
],
|
| 91 |
+
"stride_kv": [
|
| 92 |
+
2,
|
| 93 |
+
2,
|
| 94 |
+
2
|
| 95 |
+
],
|
| 96 |
+
"stride_q": [
|
| 97 |
+
1,
|
| 98 |
+
1,
|
| 99 |
+
1
|
| 100 |
+
],
|
| 101 |
+
"transformers_version": "5.16.0.dev0"
|
| 102 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fb6a24cd6ebd11dbc248fab2d5267d9c4b658009ff310e443c8f3c63faebc3a6
|
| 3 |
+
size 1490272
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_pct": 0.875,
|
| 3 |
+
"crop_size": {
|
| 4 |
+
"longest_edge": 64,
|
| 5 |
+
"shortest_edge": 64
|
| 6 |
+
},
|
| 7 |
+
"do_normalize": true,
|
| 8 |
+
"do_rescale": true,
|
| 9 |
+
"do_resize": true,
|
| 10 |
+
"image_mean": [
|
| 11 |
+
0.5,
|
| 12 |
+
0.5,
|
| 13 |
+
0.5
|
| 14 |
+
],
|
| 15 |
+
"image_processor_type": "ConvNextImageProcessor",
|
| 16 |
+
"image_std": [
|
| 17 |
+
0.5,
|
| 18 |
+
0.5,
|
| 19 |
+
0.5
|
| 20 |
+
],
|
| 21 |
+
"resample": 3,
|
| 22 |
+
"rescale_factor": 0.00392156862745098,
|
| 23 |
+
"size": {
|
| 24 |
+
"longest_edge": 64,
|
| 25 |
+
"shortest_edge": 64
|
| 26 |
+
}
|
| 27 |
+
}
|