Segformer
Browse files- config.json +80 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +259 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "nvidia/segformer-b3-finetuned-ade-512-512",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"SegformerForSemanticSegmentation"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"classifier_dropout_prob": 0.1,
|
| 8 |
+
"decoder_hidden_size": 768,
|
| 9 |
+
"depths": [
|
| 10 |
+
3,
|
| 11 |
+
4,
|
| 12 |
+
18,
|
| 13 |
+
3
|
| 14 |
+
],
|
| 15 |
+
"downsampling_rates": [
|
| 16 |
+
1,
|
| 17 |
+
4,
|
| 18 |
+
8,
|
| 19 |
+
16
|
| 20 |
+
],
|
| 21 |
+
"drop_path_rate": 0.1,
|
| 22 |
+
"hidden_act": "gelu",
|
| 23 |
+
"hidden_dropout_prob": 0.0,
|
| 24 |
+
"hidden_sizes": [
|
| 25 |
+
64,
|
| 26 |
+
128,
|
| 27 |
+
320,
|
| 28 |
+
512
|
| 29 |
+
],
|
| 30 |
+
"id2label": {
|
| 31 |
+
"0": "building",
|
| 32 |
+
"1": "door",
|
| 33 |
+
"2": "window"
|
| 34 |
+
},
|
| 35 |
+
"image_size": 224,
|
| 36 |
+
"initializer_range": 0.02,
|
| 37 |
+
"label2id": {
|
| 38 |
+
"building": 0,
|
| 39 |
+
"door": 1,
|
| 40 |
+
"window": 2
|
| 41 |
+
},
|
| 42 |
+
"layer_norm_eps": 1e-06,
|
| 43 |
+
"mlp_ratios": [
|
| 44 |
+
4,
|
| 45 |
+
4,
|
| 46 |
+
4,
|
| 47 |
+
4
|
| 48 |
+
],
|
| 49 |
+
"model_type": "segformer",
|
| 50 |
+
"num_attention_heads": [
|
| 51 |
+
1,
|
| 52 |
+
2,
|
| 53 |
+
5,
|
| 54 |
+
8
|
| 55 |
+
],
|
| 56 |
+
"num_channels": 3,
|
| 57 |
+
"num_encoder_blocks": 4,
|
| 58 |
+
"patch_sizes": [
|
| 59 |
+
7,
|
| 60 |
+
3,
|
| 61 |
+
3,
|
| 62 |
+
3
|
| 63 |
+
],
|
| 64 |
+
"reshape_last_stage": true,
|
| 65 |
+
"semantic_loss_ignore_index": 255,
|
| 66 |
+
"sr_ratios": [
|
| 67 |
+
8,
|
| 68 |
+
4,
|
| 69 |
+
2,
|
| 70 |
+
1
|
| 71 |
+
],
|
| 72 |
+
"strides": [
|
| 73 |
+
4,
|
| 74 |
+
2,
|
| 75 |
+
2,
|
| 76 |
+
2
|
| 77 |
+
],
|
| 78 |
+
"torch_dtype": "float32",
|
| 79 |
+
"transformers_version": "4.27.2"
|
| 80 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:383a91adaeafe63d0faac95af8a998a3c40c630c1ac5f97effee36d22d542acb
|
| 3 |
+
size 378175711
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:856c5309063362ee0729503113c384c0e6f35ce6ebc798980a2e2e980c1d7ad1
|
| 3 |
+
size 189128349
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:18e44d71690af658199ec19d1a46c14d400b81c55874b2088e85f13140705043
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:779be1a3bafdb7db1f4511b02816c8bf0a7868c7281d127b3e09c53223034706
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,259 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.5169988276670576,
|
| 5 |
+
"global_step": 3000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 5e-05,
|
| 13 |
+
"loss": 0.5593,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.23,
|
| 18 |
+
"learning_rate": 5e-05,
|
| 19 |
+
"loss": 0.3226,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.35,
|
| 24 |
+
"learning_rate": 5e-05,
|
| 25 |
+
"loss": 0.2389,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.35,
|
| 30 |
+
"eval_loss": 0.22972644865512848,
|
| 31 |
+
"eval_mean_iou": 0.5079903689346261,
|
| 32 |
+
"eval_runtime": 224.7647,
|
| 33 |
+
"eval_samples_per_second": 0.948,
|
| 34 |
+
"eval_steps_per_second": 0.948,
|
| 35 |
+
"step": 300
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"epoch": 0.47,
|
| 39 |
+
"learning_rate": 5e-05,
|
| 40 |
+
"loss": 0.2187,
|
| 41 |
+
"step": 400
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.59,
|
| 45 |
+
"learning_rate": 5e-05,
|
| 46 |
+
"loss": 0.1961,
|
| 47 |
+
"step": 500
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 0.7,
|
| 51 |
+
"learning_rate": 5e-05,
|
| 52 |
+
"loss": 0.1873,
|
| 53 |
+
"step": 600
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.7,
|
| 57 |
+
"eval_loss": 0.18838702142238617,
|
| 58 |
+
"eval_mean_iou": 0.6565456095704835,
|
| 59 |
+
"eval_runtime": 223.134,
|
| 60 |
+
"eval_samples_per_second": 0.955,
|
| 61 |
+
"eval_steps_per_second": 0.955,
|
| 62 |
+
"step": 600
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.82,
|
| 66 |
+
"learning_rate": 5e-05,
|
| 67 |
+
"loss": 0.172,
|
| 68 |
+
"step": 700
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.94,
|
| 72 |
+
"learning_rate": 5e-05,
|
| 73 |
+
"loss": 0.2013,
|
| 74 |
+
"step": 800
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.06,
|
| 78 |
+
"learning_rate": 5e-05,
|
| 79 |
+
"loss": 0.1549,
|
| 80 |
+
"step": 900
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.06,
|
| 84 |
+
"eval_loss": 0.1733293980360031,
|
| 85 |
+
"eval_mean_iou": 0.6822287052114181,
|
| 86 |
+
"eval_runtime": 226.0674,
|
| 87 |
+
"eval_samples_per_second": 0.942,
|
| 88 |
+
"eval_steps_per_second": 0.942,
|
| 89 |
+
"step": 900
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.17,
|
| 93 |
+
"learning_rate": 5e-05,
|
| 94 |
+
"loss": 0.1401,
|
| 95 |
+
"step": 1000
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 1.29,
|
| 99 |
+
"learning_rate": 5e-05,
|
| 100 |
+
"loss": 0.1365,
|
| 101 |
+
"step": 1100
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.41,
|
| 105 |
+
"learning_rate": 5e-05,
|
| 106 |
+
"loss": 0.151,
|
| 107 |
+
"step": 1200
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.41,
|
| 111 |
+
"eval_loss": 0.17005586624145508,
|
| 112 |
+
"eval_mean_iou": 0.6851204106483545,
|
| 113 |
+
"eval_runtime": 217.995,
|
| 114 |
+
"eval_samples_per_second": 0.977,
|
| 115 |
+
"eval_steps_per_second": 0.977,
|
| 116 |
+
"step": 1200
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.52,
|
| 120 |
+
"learning_rate": 5e-05,
|
| 121 |
+
"loss": 0.1461,
|
| 122 |
+
"step": 1300
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.64,
|
| 126 |
+
"learning_rate": 5e-05,
|
| 127 |
+
"loss": 0.1385,
|
| 128 |
+
"step": 1400
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.76,
|
| 132 |
+
"learning_rate": 5e-05,
|
| 133 |
+
"loss": 0.1737,
|
| 134 |
+
"step": 1500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.76,
|
| 138 |
+
"eval_loss": 0.16983915865421295,
|
| 139 |
+
"eval_mean_iou": 0.6635217840293423,
|
| 140 |
+
"eval_runtime": 232.534,
|
| 141 |
+
"eval_samples_per_second": 0.916,
|
| 142 |
+
"eval_steps_per_second": 0.916,
|
| 143 |
+
"step": 1500
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.88,
|
| 147 |
+
"learning_rate": 5e-05,
|
| 148 |
+
"loss": 0.1277,
|
| 149 |
+
"step": 1600
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.99,
|
| 153 |
+
"learning_rate": 5e-05,
|
| 154 |
+
"loss": 0.1366,
|
| 155 |
+
"step": 1700
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 2.0,
|
| 159 |
+
"eval_loss": 0.13025489449501038,
|
| 160 |
+
"eval_mean_iou": 0.7721716611246382,
|
| 161 |
+
"eval_runtime": 227.8304,
|
| 162 |
+
"eval_samples_per_second": 0.935,
|
| 163 |
+
"eval_steps_per_second": 0.935,
|
| 164 |
+
"step": 1706
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.11,
|
| 168 |
+
"learning_rate": 5e-05,
|
| 169 |
+
"loss": 0.1235,
|
| 170 |
+
"step": 1800
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 2.23,
|
| 174 |
+
"learning_rate": 5e-05,
|
| 175 |
+
"loss": 0.1202,
|
| 176 |
+
"step": 1900
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 2.34,
|
| 180 |
+
"learning_rate": 5e-05,
|
| 181 |
+
"loss": 0.1199,
|
| 182 |
+
"step": 2000
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 2.46,
|
| 186 |
+
"learning_rate": 5e-05,
|
| 187 |
+
"loss": 0.1254,
|
| 188 |
+
"step": 2100
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 2.58,
|
| 192 |
+
"learning_rate": 5e-05,
|
| 193 |
+
"loss": 0.1214,
|
| 194 |
+
"step": 2200
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.7,
|
| 198 |
+
"learning_rate": 5e-05,
|
| 199 |
+
"loss": 0.1228,
|
| 200 |
+
"step": 2300
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 2.81,
|
| 204 |
+
"learning_rate": 5e-05,
|
| 205 |
+
"loss": 0.123,
|
| 206 |
+
"step": 2400
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.93,
|
| 210 |
+
"learning_rate": 5e-05,
|
| 211 |
+
"loss": 0.1113,
|
| 212 |
+
"step": 2500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.0,
|
| 216 |
+
"eval_loss": 0.1386958658695221,
|
| 217 |
+
"eval_mean_iou": 0.7781228705976458,
|
| 218 |
+
"eval_runtime": 228.3799,
|
| 219 |
+
"eval_samples_per_second": 0.933,
|
| 220 |
+
"eval_steps_per_second": 0.933,
|
| 221 |
+
"step": 2559
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 3.05,
|
| 225 |
+
"learning_rate": 5e-05,
|
| 226 |
+
"loss": 0.1162,
|
| 227 |
+
"step": 2600
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 3.17,
|
| 231 |
+
"learning_rate": 5e-05,
|
| 232 |
+
"loss": 0.0971,
|
| 233 |
+
"step": 2700
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 3.28,
|
| 237 |
+
"learning_rate": 5e-05,
|
| 238 |
+
"loss": 0.097,
|
| 239 |
+
"step": 2800
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.4,
|
| 243 |
+
"learning_rate": 5e-05,
|
| 244 |
+
"loss": 0.091,
|
| 245 |
+
"step": 2900
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 3.52,
|
| 249 |
+
"learning_rate": 5e-05,
|
| 250 |
+
"loss": 0.0887,
|
| 251 |
+
"step": 3000
|
| 252 |
+
}
|
| 253 |
+
],
|
| 254 |
+
"max_steps": 170600,
|
| 255 |
+
"num_train_epochs": 200,
|
| 256 |
+
"total_flos": 6.68503279927296e+17,
|
| 257 |
+
"trial_name": null,
|
| 258 |
+
"trial_params": null
|
| 259 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e02bb2def3ec1407bed61a388338418c1e79cc077cbb8f28e3ade86d34c1666b
|
| 3 |
+
size 3579
|