qubvel-hf HF staff commited on
Commit
77837db
1 Parent(s): 6aea1b5

End of training

Browse files
Files changed (5) hide show
  1. README.md +4 -1
  2. all_results.json +28 -0
  3. test_results.json +23 -0
  4. train_results.json +8 -0
  5. trainer_state.json +1282 -0
README.md CHANGED
@@ -2,6 +2,9 @@
2
  license: other
3
  base_model: facebook/mask2former-swin-tiny-coco-instance
4
  tags:
 
 
 
5
  - generated_from_trainer
6
  model-index:
7
  - name: finetune-instance-segmentation-ade20k-mini-mask2former-v1
@@ -14,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
14
  [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/qubvel-hf-co/huggingface/runs/tpf8jz7h)
15
  # finetune-instance-segmentation-ade20k-mini-mask2former-v1
16
 
17
- This model is a fine-tuned version of [facebook/mask2former-swin-tiny-coco-instance](https://huggingface.co/facebook/mask2former-swin-tiny-coco-instance) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
  - Loss: 27.5494
20
  - Map: 0.2315
 
2
  license: other
3
  base_model: facebook/mask2former-swin-tiny-coco-instance
4
  tags:
5
+ - image-segmentation
6
+ - instance-segmentation
7
+ - vision
8
  - generated_from_trainer
9
  model-index:
10
  - name: finetune-instance-segmentation-ade20k-mini-mask2former-v1
 
17
  [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/qubvel-hf-co/huggingface/runs/tpf8jz7h)
18
  # finetune-instance-segmentation-ade20k-mini-mask2former-v1
19
 
20
+ This model is a fine-tuned version of [facebook/mask2former-swin-tiny-coco-instance](https://huggingface.co/facebook/mask2former-swin-tiny-coco-instance) on the qubvel-hf/ade20k-mini dataset.
21
  It achieves the following results on the evaluation set:
22
  - Loss: 27.5494
23
  - Map: 0.2315
all_results.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 40.0,
3
+ "test_loss": 27.620899200439453,
4
+ "test_map": 0.2315,
5
+ "test_map_50": 0.4495,
6
+ "test_map_75": 0.2185,
7
+ "test_map_car": 0.3003,
8
+ "test_map_large": 0.8161,
9
+ "test_map_medium": 0.6606,
10
+ "test_map_person": 0.1626,
11
+ "test_map_small": 0.1535,
12
+ "test_mar_1": 0.0981,
13
+ "test_mar_10": 0.2576,
14
+ "test_mar_100": 0.3,
15
+ "test_mar_100_car": 0.3776,
16
+ "test_mar_100_person": 0.2224,
17
+ "test_mar_large": 0.8618,
18
+ "test_mar_medium": 0.7189,
19
+ "test_mar_small": 0.2272,
20
+ "test_runtime": 112.5346,
21
+ "test_samples_per_second": 2.932,
22
+ "test_steps_per_second": 0.373,
23
+ "total_flos": 3.574777289637888e+18,
24
+ "train_loss": 23.846087463378908,
25
+ "train_runtime": 9658.9738,
26
+ "train_samples_per_second": 6.626,
27
+ "train_steps_per_second": 0.414
28
+ }
test_results.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 40.0,
3
+ "test_loss": 27.620899200439453,
4
+ "test_map": 0.2315,
5
+ "test_map_50": 0.4495,
6
+ "test_map_75": 0.2185,
7
+ "test_map_car": 0.3003,
8
+ "test_map_large": 0.8161,
9
+ "test_map_medium": 0.6606,
10
+ "test_map_person": 0.1626,
11
+ "test_map_small": 0.1535,
12
+ "test_mar_1": 0.0981,
13
+ "test_mar_10": 0.2576,
14
+ "test_mar_100": 0.3,
15
+ "test_mar_100_car": 0.3776,
16
+ "test_mar_100_person": 0.2224,
17
+ "test_mar_large": 0.8618,
18
+ "test_mar_medium": 0.7189,
19
+ "test_mar_small": 0.2272,
20
+ "test_runtime": 112.5346,
21
+ "test_samples_per_second": 2.932,
22
+ "test_steps_per_second": 0.373
23
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 40.0,
3
+ "total_flos": 3.574777289637888e+18,
4
+ "train_loss": 23.846087463378908,
5
+ "train_runtime": 9658.9738,
6
+ "train_samples_per_second": 6.626,
7
+ "train_steps_per_second": 0.414
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1282 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 40.0,
5
+ "eval_steps": 500,
6
+ "global_step": 4000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "grad_norm": 148.08364868164062,
14
+ "learning_rate": 1e-05,
15
+ "loss": 36.7831,
16
+ "step": 100
17
+ },
18
+ {
19
+ "epoch": 1.0,
20
+ "eval_loss": 33.27680206298828,
21
+ "eval_map": 0.1838,
22
+ "eval_map_50": 0.3677,
23
+ "eval_map_75": 0.174,
24
+ "eval_map_car": 0.25,
25
+ "eval_map_large": 0.7974,
26
+ "eval_map_medium": 0.6012,
27
+ "eval_map_person": 0.1175,
28
+ "eval_map_small": 0.1175,
29
+ "eval_mar_1": 0.0884,
30
+ "eval_mar_10": 0.2431,
31
+ "eval_mar_100": 0.284,
32
+ "eval_mar_100_car": 0.3665,
33
+ "eval_mar_100_person": 0.2014,
34
+ "eval_mar_large": 0.8712,
35
+ "eval_mar_medium": 0.7053,
36
+ "eval_mar_small": 0.2104,
37
+ "eval_runtime": 82.6291,
38
+ "eval_samples_per_second": 3.994,
39
+ "eval_steps_per_second": 0.508,
40
+ "step": 100
41
+ },
42
+ {
43
+ "epoch": 2.0,
44
+ "grad_norm": 200.38975524902344,
45
+ "learning_rate": 1e-05,
46
+ "loss": 30.2324,
47
+ "step": 200
48
+ },
49
+ {
50
+ "epoch": 2.0,
51
+ "eval_loss": 30.826847076416016,
52
+ "eval_map": 0.198,
53
+ "eval_map_50": 0.4007,
54
+ "eval_map_75": 0.1831,
55
+ "eval_map_car": 0.263,
56
+ "eval_map_large": 0.8028,
57
+ "eval_map_medium": 0.6183,
58
+ "eval_map_person": 0.1331,
59
+ "eval_map_small": 0.1321,
60
+ "eval_mar_1": 0.0916,
61
+ "eval_mar_10": 0.25,
62
+ "eval_mar_100": 0.2885,
63
+ "eval_mar_100_car": 0.3691,
64
+ "eval_mar_100_person": 0.2079,
65
+ "eval_mar_large": 0.8354,
66
+ "eval_mar_medium": 0.7125,
67
+ "eval_mar_small": 0.2151,
68
+ "eval_runtime": 76.2113,
69
+ "eval_samples_per_second": 4.33,
70
+ "eval_steps_per_second": 0.551,
71
+ "step": 200
72
+ },
73
+ {
74
+ "epoch": 3.0,
75
+ "grad_norm": 156.3461456298828,
76
+ "learning_rate": 1e-05,
77
+ "loss": 28.4136,
78
+ "step": 300
79
+ },
80
+ {
81
+ "epoch": 3.0,
82
+ "eval_loss": 29.826114654541016,
83
+ "eval_map": 0.2036,
84
+ "eval_map_50": 0.416,
85
+ "eval_map_75": 0.1849,
86
+ "eval_map_car": 0.269,
87
+ "eval_map_large": 0.7969,
88
+ "eval_map_medium": 0.6332,
89
+ "eval_map_person": 0.1381,
90
+ "eval_map_small": 0.1337,
91
+ "eval_mar_1": 0.0934,
92
+ "eval_mar_10": 0.2472,
93
+ "eval_mar_100": 0.2905,
94
+ "eval_mar_100_car": 0.3697,
95
+ "eval_mar_100_person": 0.2112,
96
+ "eval_mar_large": 0.8323,
97
+ "eval_mar_medium": 0.7162,
98
+ "eval_mar_small": 0.2169,
99
+ "eval_runtime": 80.5243,
100
+ "eval_samples_per_second": 4.098,
101
+ "eval_steps_per_second": 0.522,
102
+ "step": 300
103
+ },
104
+ {
105
+ "epoch": 4.0,
106
+ "grad_norm": 239.60546875,
107
+ "learning_rate": 1e-05,
108
+ "loss": 27.5659,
109
+ "step": 400
110
+ },
111
+ {
112
+ "epoch": 4.0,
113
+ "eval_loss": 29.292625427246094,
114
+ "eval_map": 0.2101,
115
+ "eval_map_50": 0.4176,
116
+ "eval_map_75": 0.1918,
117
+ "eval_map_car": 0.2745,
118
+ "eval_map_large": 0.8051,
119
+ "eval_map_medium": 0.6352,
120
+ "eval_map_person": 0.1456,
121
+ "eval_map_small": 0.1371,
122
+ "eval_mar_1": 0.094,
123
+ "eval_mar_10": 0.25,
124
+ "eval_mar_100": 0.2884,
125
+ "eval_mar_100_car": 0.3661,
126
+ "eval_mar_100_person": 0.2107,
127
+ "eval_mar_large": 0.8354,
128
+ "eval_mar_medium": 0.7174,
129
+ "eval_mar_small": 0.2143,
130
+ "eval_runtime": 80.9874,
131
+ "eval_samples_per_second": 4.075,
132
+ "eval_steps_per_second": 0.519,
133
+ "step": 400
134
+ },
135
+ {
136
+ "epoch": 5.0,
137
+ "grad_norm": 188.27593994140625,
138
+ "learning_rate": 1e-05,
139
+ "loss": 26.9971,
140
+ "step": 500
141
+ },
142
+ {
143
+ "epoch": 5.0,
144
+ "eval_loss": 28.8044376373291,
145
+ "eval_map": 0.213,
146
+ "eval_map_50": 0.4209,
147
+ "eval_map_75": 0.2016,
148
+ "eval_map_car": 0.2785,
149
+ "eval_map_large": 0.8094,
150
+ "eval_map_medium": 0.6419,
151
+ "eval_map_person": 0.1475,
152
+ "eval_map_small": 0.1379,
153
+ "eval_mar_1": 0.093,
154
+ "eval_mar_10": 0.2499,
155
+ "eval_mar_100": 0.2894,
156
+ "eval_mar_100_car": 0.3692,
157
+ "eval_mar_100_person": 0.2096,
158
+ "eval_mar_large": 0.8441,
159
+ "eval_mar_medium": 0.7207,
160
+ "eval_mar_small": 0.2148,
161
+ "eval_runtime": 80.6175,
162
+ "eval_samples_per_second": 4.093,
163
+ "eval_steps_per_second": 0.521,
164
+ "step": 500
165
+ },
166
+ {
167
+ "epoch": 6.0,
168
+ "grad_norm": 289.4519348144531,
169
+ "learning_rate": 1e-05,
170
+ "loss": 26.42,
171
+ "step": 600
172
+ },
173
+ {
174
+ "epoch": 6.0,
175
+ "eval_loss": 28.48480796813965,
176
+ "eval_map": 0.2196,
177
+ "eval_map_50": 0.4224,
178
+ "eval_map_75": 0.2062,
179
+ "eval_map_car": 0.2892,
180
+ "eval_map_large": 0.8046,
181
+ "eval_map_medium": 0.647,
182
+ "eval_map_person": 0.15,
183
+ "eval_map_small": 0.1426,
184
+ "eval_mar_1": 0.0944,
185
+ "eval_mar_10": 0.2523,
186
+ "eval_mar_100": 0.2925,
187
+ "eval_mar_100_car": 0.3745,
188
+ "eval_mar_100_person": 0.2106,
189
+ "eval_mar_large": 0.8354,
190
+ "eval_mar_medium": 0.7196,
191
+ "eval_mar_small": 0.2188,
192
+ "eval_runtime": 83.776,
193
+ "eval_samples_per_second": 3.939,
194
+ "eval_steps_per_second": 0.501,
195
+ "step": 600
196
+ },
197
+ {
198
+ "epoch": 7.0,
199
+ "grad_norm": 94.67891693115234,
200
+ "learning_rate": 1e-05,
201
+ "loss": 25.9065,
202
+ "step": 700
203
+ },
204
+ {
205
+ "epoch": 7.0,
206
+ "eval_loss": 28.260093688964844,
207
+ "eval_map": 0.2212,
208
+ "eval_map_50": 0.4261,
209
+ "eval_map_75": 0.207,
210
+ "eval_map_car": 0.2893,
211
+ "eval_map_large": 0.8049,
212
+ "eval_map_medium": 0.6442,
213
+ "eval_map_person": 0.153,
214
+ "eval_map_small": 0.1444,
215
+ "eval_mar_1": 0.0943,
216
+ "eval_mar_10": 0.2527,
217
+ "eval_mar_100": 0.2902,
218
+ "eval_mar_100_car": 0.3703,
219
+ "eval_mar_100_person": 0.2102,
220
+ "eval_mar_large": 0.8323,
221
+ "eval_mar_medium": 0.7103,
222
+ "eval_mar_small": 0.2176,
223
+ "eval_runtime": 76.8134,
224
+ "eval_samples_per_second": 4.296,
225
+ "eval_steps_per_second": 0.547,
226
+ "step": 700
227
+ },
228
+ {
229
+ "epoch": 8.0,
230
+ "grad_norm": 111.85684967041016,
231
+ "learning_rate": 1e-05,
232
+ "loss": 25.6766,
233
+ "step": 800
234
+ },
235
+ {
236
+ "epoch": 8.0,
237
+ "eval_loss": 28.25814437866211,
238
+ "eval_map": 0.2209,
239
+ "eval_map_50": 0.4276,
240
+ "eval_map_75": 0.2076,
241
+ "eval_map_car": 0.2885,
242
+ "eval_map_large": 0.8201,
243
+ "eval_map_medium": 0.6485,
244
+ "eval_map_person": 0.1532,
245
+ "eval_map_small": 0.1434,
246
+ "eval_mar_1": 0.0943,
247
+ "eval_mar_10": 0.2532,
248
+ "eval_mar_100": 0.294,
249
+ "eval_mar_100_car": 0.3758,
250
+ "eval_mar_100_person": 0.2122,
251
+ "eval_mar_large": 0.8681,
252
+ "eval_mar_medium": 0.7212,
253
+ "eval_mar_small": 0.2197,
254
+ "eval_runtime": 80.2804,
255
+ "eval_samples_per_second": 4.111,
256
+ "eval_steps_per_second": 0.523,
257
+ "step": 800
258
+ },
259
+ {
260
+ "epoch": 9.0,
261
+ "grad_norm": 152.9182586669922,
262
+ "learning_rate": 1e-05,
263
+ "loss": 25.3111,
264
+ "step": 900
265
+ },
266
+ {
267
+ "epoch": 9.0,
268
+ "eval_loss": 27.862319946289062,
269
+ "eval_map": 0.2234,
270
+ "eval_map_50": 0.4318,
271
+ "eval_map_75": 0.2163,
272
+ "eval_map_car": 0.2907,
273
+ "eval_map_large": 0.8252,
274
+ "eval_map_medium": 0.649,
275
+ "eval_map_person": 0.1561,
276
+ "eval_map_small": 0.1451,
277
+ "eval_mar_1": 0.0951,
278
+ "eval_mar_10": 0.2519,
279
+ "eval_mar_100": 0.2953,
280
+ "eval_mar_100_car": 0.3757,
281
+ "eval_mar_100_person": 0.2148,
282
+ "eval_mar_large": 0.8649,
283
+ "eval_mar_medium": 0.721,
284
+ "eval_mar_small": 0.2212,
285
+ "eval_runtime": 81.4979,
286
+ "eval_samples_per_second": 4.049,
287
+ "eval_steps_per_second": 0.515,
288
+ "step": 900
289
+ },
290
+ {
291
+ "epoch": 10.0,
292
+ "grad_norm": 137.48260498046875,
293
+ "learning_rate": 1e-05,
294
+ "loss": 24.9424,
295
+ "step": 1000
296
+ },
297
+ {
298
+ "epoch": 10.0,
299
+ "eval_loss": 27.892526626586914,
300
+ "eval_map": 0.2256,
301
+ "eval_map_50": 0.4367,
302
+ "eval_map_75": 0.2129,
303
+ "eval_map_car": 0.2923,
304
+ "eval_map_large": 0.8314,
305
+ "eval_map_medium": 0.6476,
306
+ "eval_map_person": 0.1588,
307
+ "eval_map_small": 0.1479,
308
+ "eval_mar_1": 0.0953,
309
+ "eval_mar_10": 0.2556,
310
+ "eval_mar_100": 0.2973,
311
+ "eval_mar_100_car": 0.3793,
312
+ "eval_mar_100_person": 0.2153,
313
+ "eval_mar_large": 0.8712,
314
+ "eval_mar_medium": 0.7159,
315
+ "eval_mar_small": 0.2244,
316
+ "eval_runtime": 82.7931,
317
+ "eval_samples_per_second": 3.986,
318
+ "eval_steps_per_second": 0.507,
319
+ "step": 1000
320
+ },
321
+ {
322
+ "epoch": 11.0,
323
+ "grad_norm": 122.72390747070312,
324
+ "learning_rate": 1e-05,
325
+ "loss": 24.6502,
326
+ "step": 1100
327
+ },
328
+ {
329
+ "epoch": 11.0,
330
+ "eval_loss": 27.75235939025879,
331
+ "eval_map": 0.2254,
332
+ "eval_map_50": 0.441,
333
+ "eval_map_75": 0.2163,
334
+ "eval_map_car": 0.2929,
335
+ "eval_map_large": 0.8186,
336
+ "eval_map_medium": 0.6468,
337
+ "eval_map_person": 0.1578,
338
+ "eval_map_small": 0.1486,
339
+ "eval_mar_1": 0.0952,
340
+ "eval_mar_10": 0.2556,
341
+ "eval_mar_100": 0.2963,
342
+ "eval_mar_100_car": 0.3772,
343
+ "eval_mar_100_person": 0.2153,
344
+ "eval_mar_large": 0.8681,
345
+ "eval_mar_medium": 0.7167,
346
+ "eval_mar_small": 0.2231,
347
+ "eval_runtime": 82.907,
348
+ "eval_samples_per_second": 3.98,
349
+ "eval_steps_per_second": 0.507,
350
+ "step": 1100
351
+ },
352
+ {
353
+ "epoch": 12.0,
354
+ "grad_norm": 184.76023864746094,
355
+ "learning_rate": 1e-05,
356
+ "loss": 24.5278,
357
+ "step": 1200
358
+ },
359
+ {
360
+ "epoch": 12.0,
361
+ "eval_loss": 27.71224594116211,
362
+ "eval_map": 0.2252,
363
+ "eval_map_50": 0.4349,
364
+ "eval_map_75": 0.2167,
365
+ "eval_map_car": 0.2921,
366
+ "eval_map_large": 0.8237,
367
+ "eval_map_medium": 0.6462,
368
+ "eval_map_person": 0.1583,
369
+ "eval_map_small": 0.1473,
370
+ "eval_mar_1": 0.0927,
371
+ "eval_mar_10": 0.2549,
372
+ "eval_mar_100": 0.2979,
373
+ "eval_mar_100_car": 0.3793,
374
+ "eval_mar_100_person": 0.2165,
375
+ "eval_mar_large": 0.8649,
376
+ "eval_mar_medium": 0.7162,
377
+ "eval_mar_small": 0.2251,
378
+ "eval_runtime": 83.1034,
379
+ "eval_samples_per_second": 3.971,
380
+ "eval_steps_per_second": 0.505,
381
+ "step": 1200
382
+ },
383
+ {
384
+ "epoch": 13.0,
385
+ "grad_norm": 259.48358154296875,
386
+ "learning_rate": 1e-05,
387
+ "loss": 24.3514,
388
+ "step": 1300
389
+ },
390
+ {
391
+ "epoch": 13.0,
392
+ "eval_loss": 27.538196563720703,
393
+ "eval_map": 0.224,
394
+ "eval_map_50": 0.4345,
395
+ "eval_map_75": 0.2156,
396
+ "eval_map_car": 0.2897,
397
+ "eval_map_large": 0.8324,
398
+ "eval_map_medium": 0.6554,
399
+ "eval_map_person": 0.1583,
400
+ "eval_map_small": 0.1459,
401
+ "eval_mar_1": 0.0958,
402
+ "eval_mar_10": 0.2554,
403
+ "eval_mar_100": 0.2988,
404
+ "eval_mar_100_car": 0.3785,
405
+ "eval_mar_100_person": 0.2191,
406
+ "eval_mar_large": 0.8806,
407
+ "eval_mar_medium": 0.722,
408
+ "eval_mar_small": 0.2251,
409
+ "eval_runtime": 81.6693,
410
+ "eval_samples_per_second": 4.041,
411
+ "eval_steps_per_second": 0.514,
412
+ "step": 1300
413
+ },
414
+ {
415
+ "epoch": 14.0,
416
+ "grad_norm": 129.53614807128906,
417
+ "learning_rate": 1e-05,
418
+ "loss": 24.3422,
419
+ "step": 1400
420
+ },
421
+ {
422
+ "epoch": 14.0,
423
+ "eval_loss": 27.5665283203125,
424
+ "eval_map": 0.226,
425
+ "eval_map_50": 0.4374,
426
+ "eval_map_75": 0.2172,
427
+ "eval_map_car": 0.2928,
428
+ "eval_map_large": 0.8059,
429
+ "eval_map_medium": 0.6505,
430
+ "eval_map_person": 0.1592,
431
+ "eval_map_small": 0.1488,
432
+ "eval_mar_1": 0.0974,
433
+ "eval_mar_10": 0.2551,
434
+ "eval_mar_100": 0.2964,
435
+ "eval_mar_100_car": 0.377,
436
+ "eval_mar_100_person": 0.2158,
437
+ "eval_mar_large": 0.8434,
438
+ "eval_mar_medium": 0.7141,
439
+ "eval_mar_small": 0.2241,
440
+ "eval_runtime": 83.3796,
441
+ "eval_samples_per_second": 3.958,
442
+ "eval_steps_per_second": 0.504,
443
+ "step": 1400
444
+ },
445
+ {
446
+ "epoch": 15.0,
447
+ "grad_norm": 94.8553466796875,
448
+ "learning_rate": 1e-05,
449
+ "loss": 23.9768,
450
+ "step": 1500
451
+ },
452
+ {
453
+ "epoch": 15.0,
454
+ "eval_loss": 27.777042388916016,
455
+ "eval_map": 0.2281,
456
+ "eval_map_50": 0.4379,
457
+ "eval_map_75": 0.2215,
458
+ "eval_map_car": 0.2963,
459
+ "eval_map_large": 0.8188,
460
+ "eval_map_medium": 0.6553,
461
+ "eval_map_person": 0.1599,
462
+ "eval_map_small": 0.1499,
463
+ "eval_mar_1": 0.096,
464
+ "eval_mar_10": 0.2553,
465
+ "eval_mar_100": 0.2978,
466
+ "eval_mar_100_car": 0.3793,
467
+ "eval_mar_100_person": 0.2163,
468
+ "eval_mar_large": 0.8632,
469
+ "eval_mar_medium": 0.72,
470
+ "eval_mar_small": 0.2244,
471
+ "eval_runtime": 84.3745,
472
+ "eval_samples_per_second": 3.911,
473
+ "eval_steps_per_second": 0.498,
474
+ "step": 1500
475
+ },
476
+ {
477
+ "epoch": 16.0,
478
+ "grad_norm": 217.70468139648438,
479
+ "learning_rate": 1e-05,
480
+ "loss": 23.7005,
481
+ "step": 1600
482
+ },
483
+ {
484
+ "epoch": 16.0,
485
+ "eval_loss": 27.55354881286621,
486
+ "eval_map": 0.227,
487
+ "eval_map_50": 0.4392,
488
+ "eval_map_75": 0.2167,
489
+ "eval_map_car": 0.2932,
490
+ "eval_map_large": 0.8165,
491
+ "eval_map_medium": 0.6509,
492
+ "eval_map_person": 0.1608,
493
+ "eval_map_small": 0.1485,
494
+ "eval_mar_1": 0.0965,
495
+ "eval_mar_10": 0.255,
496
+ "eval_mar_100": 0.2972,
497
+ "eval_mar_100_car": 0.3779,
498
+ "eval_mar_100_person": 0.2164,
499
+ "eval_mar_large": 0.8656,
500
+ "eval_mar_medium": 0.7175,
501
+ "eval_mar_small": 0.2241,
502
+ "eval_runtime": 82.802,
503
+ "eval_samples_per_second": 3.985,
504
+ "eval_steps_per_second": 0.507,
505
+ "step": 1600
506
+ },
507
+ {
508
+ "epoch": 17.0,
509
+ "grad_norm": 121.01888275146484,
510
+ "learning_rate": 1e-05,
511
+ "loss": 23.579,
512
+ "step": 1700
513
+ },
514
+ {
515
+ "epoch": 17.0,
516
+ "eval_loss": 27.489389419555664,
517
+ "eval_map": 0.2286,
518
+ "eval_map_50": 0.44,
519
+ "eval_map_75": 0.2209,
520
+ "eval_map_car": 0.2952,
521
+ "eval_map_large": 0.8152,
522
+ "eval_map_medium": 0.6488,
523
+ "eval_map_person": 0.162,
524
+ "eval_map_small": 0.1511,
525
+ "eval_mar_1": 0.097,
526
+ "eval_mar_10": 0.2583,
527
+ "eval_mar_100": 0.2965,
528
+ "eval_mar_100_car": 0.3785,
529
+ "eval_mar_100_person": 0.2144,
530
+ "eval_mar_large": 0.8601,
531
+ "eval_mar_medium": 0.7113,
532
+ "eval_mar_small": 0.2243,
533
+ "eval_runtime": 82.3337,
534
+ "eval_samples_per_second": 4.008,
535
+ "eval_steps_per_second": 0.51,
536
+ "step": 1700
537
+ },
538
+ {
539
+ "epoch": 18.0,
540
+ "grad_norm": 84.91890716552734,
541
+ "learning_rate": 1e-05,
542
+ "loss": 23.5004,
543
+ "step": 1800
544
+ },
545
+ {
546
+ "epoch": 18.0,
547
+ "eval_loss": 27.218843460083008,
548
+ "eval_map": 0.2274,
549
+ "eval_map_50": 0.4374,
550
+ "eval_map_75": 0.216,
551
+ "eval_map_car": 0.2933,
552
+ "eval_map_large": 0.7954,
553
+ "eval_map_medium": 0.6512,
554
+ "eval_map_person": 0.1614,
555
+ "eval_map_small": 0.1498,
556
+ "eval_mar_1": 0.0962,
557
+ "eval_mar_10": 0.2562,
558
+ "eval_mar_100": 0.2969,
559
+ "eval_mar_100_car": 0.3788,
560
+ "eval_mar_100_person": 0.215,
561
+ "eval_mar_large": 0.8323,
562
+ "eval_mar_medium": 0.712,
563
+ "eval_mar_small": 0.2251,
564
+ "eval_runtime": 82.0132,
565
+ "eval_samples_per_second": 4.024,
566
+ "eval_steps_per_second": 0.512,
567
+ "step": 1800
568
+ },
569
+ {
570
+ "epoch": 19.0,
571
+ "grad_norm": 152.5535125732422,
572
+ "learning_rate": 1e-05,
573
+ "loss": 23.1744,
574
+ "step": 1900
575
+ },
576
+ {
577
+ "epoch": 19.0,
578
+ "eval_loss": 27.352279663085938,
579
+ "eval_map": 0.2286,
580
+ "eval_map_50": 0.4391,
581
+ "eval_map_75": 0.2166,
582
+ "eval_map_car": 0.297,
583
+ "eval_map_large": 0.8203,
584
+ "eval_map_medium": 0.6559,
585
+ "eval_map_person": 0.1602,
586
+ "eval_map_small": 0.1494,
587
+ "eval_mar_1": 0.0962,
588
+ "eval_mar_10": 0.2565,
589
+ "eval_mar_100": 0.2998,
590
+ "eval_mar_100_car": 0.3821,
591
+ "eval_mar_100_person": 0.2174,
592
+ "eval_mar_large": 0.8656,
593
+ "eval_mar_medium": 0.7156,
594
+ "eval_mar_small": 0.2274,
595
+ "eval_runtime": 83.3155,
596
+ "eval_samples_per_second": 3.961,
597
+ "eval_steps_per_second": 0.504,
598
+ "step": 1900
599
+ },
600
+ {
601
+ "epoch": 20.0,
602
+ "grad_norm": 122.66116333007812,
603
+ "learning_rate": 1e-05,
604
+ "loss": 23.1884,
605
+ "step": 2000
606
+ },
607
+ {
608
+ "epoch": 20.0,
609
+ "eval_loss": 27.118486404418945,
610
+ "eval_map": 0.2304,
611
+ "eval_map_50": 0.4395,
612
+ "eval_map_75": 0.2204,
613
+ "eval_map_car": 0.2998,
614
+ "eval_map_large": 0.8004,
615
+ "eval_map_medium": 0.6533,
616
+ "eval_map_person": 0.1611,
617
+ "eval_map_small": 0.1521,
618
+ "eval_mar_1": 0.0968,
619
+ "eval_mar_10": 0.2558,
620
+ "eval_mar_100": 0.299,
621
+ "eval_mar_100_car": 0.3809,
622
+ "eval_mar_100_person": 0.217,
623
+ "eval_mar_large": 0.8347,
624
+ "eval_mar_medium": 0.7131,
625
+ "eval_mar_small": 0.2273,
626
+ "eval_runtime": 82.4333,
627
+ "eval_samples_per_second": 4.003,
628
+ "eval_steps_per_second": 0.51,
629
+ "step": 2000
630
+ },
631
+ {
632
+ "epoch": 21.0,
633
+ "grad_norm": 247.2215118408203,
634
+ "learning_rate": 1e-05,
635
+ "loss": 22.9136,
636
+ "step": 2100
637
+ },
638
+ {
639
+ "epoch": 21.0,
640
+ "eval_loss": 27.429582595825195,
641
+ "eval_map": 0.2301,
642
+ "eval_map_50": 0.4386,
643
+ "eval_map_75": 0.2197,
644
+ "eval_map_car": 0.2992,
645
+ "eval_map_large": 0.8185,
646
+ "eval_map_medium": 0.6545,
647
+ "eval_map_person": 0.1609,
648
+ "eval_map_small": 0.1518,
649
+ "eval_mar_1": 0.0968,
650
+ "eval_mar_10": 0.2552,
651
+ "eval_mar_100": 0.2979,
652
+ "eval_mar_100_car": 0.3778,
653
+ "eval_mar_100_person": 0.2179,
654
+ "eval_mar_large": 0.8712,
655
+ "eval_mar_medium": 0.7123,
656
+ "eval_mar_small": 0.2256,
657
+ "eval_runtime": 81.5801,
658
+ "eval_samples_per_second": 4.045,
659
+ "eval_steps_per_second": 0.515,
660
+ "step": 2100
661
+ },
662
+ {
663
+ "epoch": 22.0,
664
+ "grad_norm": 102.28423309326172,
665
+ "learning_rate": 1e-05,
666
+ "loss": 22.6863,
667
+ "step": 2200
668
+ },
669
+ {
670
+ "epoch": 22.0,
671
+ "eval_loss": 26.9977970123291,
672
+ "eval_map": 0.2309,
673
+ "eval_map_50": 0.444,
674
+ "eval_map_75": 0.2196,
675
+ "eval_map_car": 0.2994,
676
+ "eval_map_large": 0.7955,
677
+ "eval_map_medium": 0.657,
678
+ "eval_map_person": 0.1624,
679
+ "eval_map_small": 0.1519,
680
+ "eval_mar_1": 0.0976,
681
+ "eval_mar_10": 0.2543,
682
+ "eval_mar_100": 0.2982,
683
+ "eval_mar_100_car": 0.3784,
684
+ "eval_mar_100_person": 0.2181,
685
+ "eval_mar_large": 0.8316,
686
+ "eval_mar_medium": 0.714,
687
+ "eval_mar_small": 0.2264,
688
+ "eval_runtime": 82.3586,
689
+ "eval_samples_per_second": 4.007,
690
+ "eval_steps_per_second": 0.51,
691
+ "step": 2200
692
+ },
693
+ {
694
+ "epoch": 23.0,
695
+ "grad_norm": 280.4220886230469,
696
+ "learning_rate": 1e-05,
697
+ "loss": 22.7741,
698
+ "step": 2300
699
+ },
700
+ {
701
+ "epoch": 23.0,
702
+ "eval_loss": 27.07025146484375,
703
+ "eval_map": 0.23,
704
+ "eval_map_50": 0.4436,
705
+ "eval_map_75": 0.2183,
706
+ "eval_map_car": 0.2979,
707
+ "eval_map_large": 0.8029,
708
+ "eval_map_medium": 0.6508,
709
+ "eval_map_person": 0.162,
710
+ "eval_map_small": 0.1519,
711
+ "eval_mar_1": 0.0966,
712
+ "eval_mar_10": 0.2562,
713
+ "eval_mar_100": 0.3001,
714
+ "eval_mar_100_car": 0.3823,
715
+ "eval_mar_100_person": 0.218,
716
+ "eval_mar_large": 0.8434,
717
+ "eval_mar_medium": 0.7106,
718
+ "eval_mar_small": 0.229,
719
+ "eval_runtime": 82.5934,
720
+ "eval_samples_per_second": 3.995,
721
+ "eval_steps_per_second": 0.509,
722
+ "step": 2300
723
+ },
724
+ {
725
+ "epoch": 24.0,
726
+ "grad_norm": 110.03192138671875,
727
+ "learning_rate": 1e-05,
728
+ "loss": 22.4779,
729
+ "step": 2400
730
+ },
731
+ {
732
+ "epoch": 24.0,
733
+ "eval_loss": 27.03939437866211,
734
+ "eval_map": 0.2335,
735
+ "eval_map_50": 0.4521,
736
+ "eval_map_75": 0.2252,
737
+ "eval_map_car": 0.3045,
738
+ "eval_map_large": 0.8318,
739
+ "eval_map_medium": 0.656,
740
+ "eval_map_person": 0.1624,
741
+ "eval_map_small": 0.1552,
742
+ "eval_mar_1": 0.0962,
743
+ "eval_mar_10": 0.2598,
744
+ "eval_mar_100": 0.3026,
745
+ "eval_mar_100_car": 0.3865,
746
+ "eval_mar_100_person": 0.2187,
747
+ "eval_mar_large": 0.8601,
748
+ "eval_mar_medium": 0.7143,
749
+ "eval_mar_small": 0.231,
750
+ "eval_runtime": 82.8838,
751
+ "eval_samples_per_second": 3.981,
752
+ "eval_steps_per_second": 0.507,
753
+ "step": 2400
754
+ },
755
+ {
756
+ "epoch": 25.0,
757
+ "grad_norm": 135.92689514160156,
758
+ "learning_rate": 1e-05,
759
+ "loss": 22.357,
760
+ "step": 2500
761
+ },
762
+ {
763
+ "epoch": 25.0,
764
+ "eval_loss": 27.148326873779297,
765
+ "eval_map": 0.2304,
766
+ "eval_map_50": 0.4456,
767
+ "eval_map_75": 0.2189,
768
+ "eval_map_car": 0.2989,
769
+ "eval_map_large": 0.8065,
770
+ "eval_map_medium": 0.6586,
771
+ "eval_map_person": 0.162,
772
+ "eval_map_small": 0.1517,
773
+ "eval_mar_1": 0.0967,
774
+ "eval_mar_10": 0.2554,
775
+ "eval_mar_100": 0.2996,
776
+ "eval_mar_100_car": 0.3805,
777
+ "eval_mar_100_person": 0.2187,
778
+ "eval_mar_large": 0.8378,
779
+ "eval_mar_medium": 0.7143,
780
+ "eval_mar_small": 0.2278,
781
+ "eval_runtime": 82.5665,
782
+ "eval_samples_per_second": 3.997,
783
+ "eval_steps_per_second": 0.509,
784
+ "step": 2500
785
+ },
786
+ {
787
+ "epoch": 26.0,
788
+ "grad_norm": 580.8344116210938,
789
+ "learning_rate": 1e-05,
790
+ "loss": 22.3167,
791
+ "step": 2600
792
+ },
793
+ {
794
+ "epoch": 26.0,
795
+ "eval_loss": 27.329933166503906,
796
+ "eval_map": 0.232,
797
+ "eval_map_50": 0.4438,
798
+ "eval_map_75": 0.2193,
799
+ "eval_map_car": 0.3016,
800
+ "eval_map_large": 0.8221,
801
+ "eval_map_medium": 0.6572,
802
+ "eval_map_person": 0.1624,
803
+ "eval_map_small": 0.1534,
804
+ "eval_mar_1": 0.0977,
805
+ "eval_mar_10": 0.2564,
806
+ "eval_mar_100": 0.2989,
807
+ "eval_mar_100_car": 0.3802,
808
+ "eval_mar_100_person": 0.2176,
809
+ "eval_mar_large": 0.8681,
810
+ "eval_mar_medium": 0.7134,
811
+ "eval_mar_small": 0.2267,
812
+ "eval_runtime": 81.9253,
813
+ "eval_samples_per_second": 4.028,
814
+ "eval_steps_per_second": 0.513,
815
+ "step": 2600
816
+ },
817
+ {
818
+ "epoch": 27.0,
819
+ "grad_norm": 223.89878845214844,
820
+ "learning_rate": 1e-05,
821
+ "loss": 22.0958,
822
+ "step": 2700
823
+ },
824
+ {
825
+ "epoch": 27.0,
826
+ "eval_loss": 27.25712013244629,
827
+ "eval_map": 0.232,
828
+ "eval_map_50": 0.4438,
829
+ "eval_map_75": 0.2171,
830
+ "eval_map_car": 0.3018,
831
+ "eval_map_large": 0.8268,
832
+ "eval_map_medium": 0.6539,
833
+ "eval_map_person": 0.1622,
834
+ "eval_map_small": 0.1535,
835
+ "eval_mar_1": 0.0974,
836
+ "eval_mar_10": 0.2591,
837
+ "eval_mar_100": 0.2986,
838
+ "eval_mar_100_car": 0.3788,
839
+ "eval_mar_100_person": 0.2185,
840
+ "eval_mar_large": 0.8774,
841
+ "eval_mar_medium": 0.7153,
842
+ "eval_mar_small": 0.226,
843
+ "eval_runtime": 82.6409,
844
+ "eval_samples_per_second": 3.993,
845
+ "eval_steps_per_second": 0.508,
846
+ "step": 2700
847
+ },
848
+ {
849
+ "epoch": 28.0,
850
+ "grad_norm": 131.83126831054688,
851
+ "learning_rate": 1e-05,
852
+ "loss": 22.0902,
853
+ "step": 2800
854
+ },
855
+ {
856
+ "epoch": 28.0,
857
+ "eval_loss": 27.515573501586914,
858
+ "eval_map": 0.2315,
859
+ "eval_map_50": 0.4482,
860
+ "eval_map_75": 0.2177,
861
+ "eval_map_car": 0.3004,
862
+ "eval_map_large": 0.8265,
863
+ "eval_map_medium": 0.6566,
864
+ "eval_map_person": 0.1626,
865
+ "eval_map_small": 0.1539,
866
+ "eval_mar_1": 0.0978,
867
+ "eval_mar_10": 0.2583,
868
+ "eval_mar_100": 0.3021,
869
+ "eval_mar_100_car": 0.3842,
870
+ "eval_mar_100_person": 0.22,
871
+ "eval_mar_large": 0.8719,
872
+ "eval_mar_medium": 0.716,
873
+ "eval_mar_small": 0.23,
874
+ "eval_runtime": 84.8147,
875
+ "eval_samples_per_second": 3.891,
876
+ "eval_steps_per_second": 0.495,
877
+ "step": 2800
878
+ },
879
+ {
880
+ "epoch": 29.0,
881
+ "grad_norm": 131.6576690673828,
882
+ "learning_rate": 1e-05,
883
+ "loss": 21.9943,
884
+ "step": 2900
885
+ },
886
+ {
887
+ "epoch": 29.0,
888
+ "eval_loss": 27.014198303222656,
889
+ "eval_map": 0.2288,
890
+ "eval_map_50": 0.4449,
891
+ "eval_map_75": 0.2155,
892
+ "eval_map_car": 0.2961,
893
+ "eval_map_large": 0.8176,
894
+ "eval_map_medium": 0.6536,
895
+ "eval_map_person": 0.1616,
896
+ "eval_map_small": 0.1511,
897
+ "eval_mar_1": 0.097,
898
+ "eval_mar_10": 0.2557,
899
+ "eval_mar_100": 0.2984,
900
+ "eval_mar_100_car": 0.3766,
901
+ "eval_mar_100_person": 0.2202,
902
+ "eval_mar_large": 0.8569,
903
+ "eval_mar_medium": 0.7169,
904
+ "eval_mar_small": 0.2257,
905
+ "eval_runtime": 84.758,
906
+ "eval_samples_per_second": 3.893,
907
+ "eval_steps_per_second": 0.496,
908
+ "step": 2900
909
+ },
910
+ {
911
+ "epoch": 30.0,
912
+ "grad_norm": 81.99681091308594,
913
+ "learning_rate": 1e-05,
914
+ "loss": 21.8843,
915
+ "step": 3000
916
+ },
917
+ {
918
+ "epoch": 30.0,
919
+ "eval_loss": 27.1738338470459,
920
+ "eval_map": 0.2314,
921
+ "eval_map_50": 0.4456,
922
+ "eval_map_75": 0.2192,
923
+ "eval_map_car": 0.2999,
924
+ "eval_map_large": 0.8263,
925
+ "eval_map_medium": 0.6557,
926
+ "eval_map_person": 0.1629,
927
+ "eval_map_small": 0.1534,
928
+ "eval_mar_1": 0.0973,
929
+ "eval_mar_10": 0.2587,
930
+ "eval_mar_100": 0.3026,
931
+ "eval_mar_100_car": 0.3848,
932
+ "eval_mar_100_person": 0.2203,
933
+ "eval_mar_large": 0.8625,
934
+ "eval_mar_medium": 0.7204,
935
+ "eval_mar_small": 0.23,
936
+ "eval_runtime": 83.1696,
937
+ "eval_samples_per_second": 3.968,
938
+ "eval_steps_per_second": 0.505,
939
+ "step": 3000
940
+ },
941
+ {
942
+ "epoch": 31.0,
943
+ "grad_norm": 161.1930389404297,
944
+ "learning_rate": 1e-05,
945
+ "loss": 21.8635,
946
+ "step": 3100
947
+ },
948
+ {
949
+ "epoch": 31.0,
950
+ "eval_loss": 27.065763473510742,
951
+ "eval_map": 0.2316,
952
+ "eval_map_50": 0.4461,
953
+ "eval_map_75": 0.22,
954
+ "eval_map_car": 0.3003,
955
+ "eval_map_large": 0.8166,
956
+ "eval_map_medium": 0.6582,
957
+ "eval_map_person": 0.163,
958
+ "eval_map_small": 0.1534,
959
+ "eval_mar_1": 0.0987,
960
+ "eval_mar_10": 0.2581,
961
+ "eval_mar_100": 0.3013,
962
+ "eval_mar_100_car": 0.3838,
963
+ "eval_mar_100_person": 0.2188,
964
+ "eval_mar_large": 0.8625,
965
+ "eval_mar_medium": 0.7156,
966
+ "eval_mar_small": 0.2292,
967
+ "eval_runtime": 81.8266,
968
+ "eval_samples_per_second": 4.033,
969
+ "eval_steps_per_second": 0.513,
970
+ "step": 3100
971
+ },
972
+ {
973
+ "epoch": 32.0,
974
+ "grad_norm": 125.2677993774414,
975
+ "learning_rate": 1e-05,
976
+ "loss": 21.473,
977
+ "step": 3200
978
+ },
979
+ {
980
+ "epoch": 32.0,
981
+ "eval_loss": 27.135364532470703,
982
+ "eval_map": 0.2323,
983
+ "eval_map_50": 0.4493,
984
+ "eval_map_75": 0.219,
985
+ "eval_map_car": 0.3026,
986
+ "eval_map_large": 0.8077,
987
+ "eval_map_medium": 0.6569,
988
+ "eval_map_person": 0.1619,
989
+ "eval_map_small": 0.1545,
990
+ "eval_mar_1": 0.0966,
991
+ "eval_mar_10": 0.259,
992
+ "eval_mar_100": 0.3024,
993
+ "eval_mar_100_car": 0.3866,
994
+ "eval_mar_100_person": 0.2182,
995
+ "eval_mar_large": 0.8507,
996
+ "eval_mar_medium": 0.7172,
997
+ "eval_mar_small": 0.2305,
998
+ "eval_runtime": 82.9012,
999
+ "eval_samples_per_second": 3.981,
1000
+ "eval_steps_per_second": 0.507,
1001
+ "step": 3200
1002
+ },
1003
+ {
1004
+ "epoch": 33.0,
1005
+ "grad_norm": 349.8231201171875,
1006
+ "learning_rate": 1e-05,
1007
+ "loss": 21.6879,
1008
+ "step": 3300
1009
+ },
1010
+ {
1011
+ "epoch": 33.0,
1012
+ "eval_loss": 26.981035232543945,
1013
+ "eval_map": 0.2306,
1014
+ "eval_map_50": 0.4461,
1015
+ "eval_map_75": 0.2178,
1016
+ "eval_map_car": 0.2989,
1017
+ "eval_map_large": 0.8095,
1018
+ "eval_map_medium": 0.6572,
1019
+ "eval_map_person": 0.1624,
1020
+ "eval_map_small": 0.1533,
1021
+ "eval_mar_1": 0.0983,
1022
+ "eval_mar_10": 0.2581,
1023
+ "eval_mar_100": 0.3004,
1024
+ "eval_mar_100_car": 0.3814,
1025
+ "eval_mar_100_person": 0.2194,
1026
+ "eval_mar_large": 0.8476,
1027
+ "eval_mar_medium": 0.7146,
1028
+ "eval_mar_small": 0.2285,
1029
+ "eval_runtime": 83.6457,
1030
+ "eval_samples_per_second": 3.945,
1031
+ "eval_steps_per_second": 0.502,
1032
+ "step": 3300
1033
+ },
1034
+ {
1035
+ "epoch": 34.0,
1036
+ "grad_norm": 114.99758911132812,
1037
+ "learning_rate": 1e-05,
1038
+ "loss": 21.3771,
1039
+ "step": 3400
1040
+ },
1041
+ {
1042
+ "epoch": 34.0,
1043
+ "eval_loss": 27.532331466674805,
1044
+ "eval_map": 0.23,
1045
+ "eval_map_50": 0.4476,
1046
+ "eval_map_75": 0.2149,
1047
+ "eval_map_car": 0.298,
1048
+ "eval_map_large": 0.8185,
1049
+ "eval_map_medium": 0.6593,
1050
+ "eval_map_person": 0.162,
1051
+ "eval_map_small": 0.1536,
1052
+ "eval_mar_1": 0.0968,
1053
+ "eval_mar_10": 0.2577,
1054
+ "eval_mar_100": 0.2996,
1055
+ "eval_mar_100_car": 0.3781,
1056
+ "eval_mar_100_person": 0.2212,
1057
+ "eval_mar_large": 0.8618,
1058
+ "eval_mar_medium": 0.7204,
1059
+ "eval_mar_small": 0.2265,
1060
+ "eval_runtime": 82.9885,
1061
+ "eval_samples_per_second": 3.976,
1062
+ "eval_steps_per_second": 0.506,
1063
+ "step": 3400
1064
+ },
1065
+ {
1066
+ "epoch": 35.0,
1067
+ "grad_norm": 183.1493377685547,
1068
+ "learning_rate": 1e-05,
1069
+ "loss": 21.2772,
1070
+ "step": 3500
1071
+ },
1072
+ {
1073
+ "epoch": 35.0,
1074
+ "eval_loss": 27.14505958557129,
1075
+ "eval_map": 0.2327,
1076
+ "eval_map_50": 0.4465,
1077
+ "eval_map_75": 0.2172,
1078
+ "eval_map_car": 0.3038,
1079
+ "eval_map_large": 0.8195,
1080
+ "eval_map_medium": 0.6641,
1081
+ "eval_map_person": 0.1616,
1082
+ "eval_map_small": 0.1544,
1083
+ "eval_mar_1": 0.0988,
1084
+ "eval_mar_10": 0.2597,
1085
+ "eval_mar_100": 0.3028,
1086
+ "eval_mar_100_car": 0.3847,
1087
+ "eval_mar_100_person": 0.221,
1088
+ "eval_mar_large": 0.8594,
1089
+ "eval_mar_medium": 0.7262,
1090
+ "eval_mar_small": 0.2294,
1091
+ "eval_runtime": 82.073,
1092
+ "eval_samples_per_second": 4.021,
1093
+ "eval_steps_per_second": 0.512,
1094
+ "step": 3500
1095
+ },
1096
+ {
1097
+ "epoch": 36.0,
1098
+ "grad_norm": 115.38822174072266,
1099
+ "learning_rate": 1e-05,
1100
+ "loss": 21.3682,
1101
+ "step": 3600
1102
+ },
1103
+ {
1104
+ "epoch": 36.0,
1105
+ "eval_loss": 27.46976089477539,
1106
+ "eval_map": 0.2334,
1107
+ "eval_map_50": 0.4503,
1108
+ "eval_map_75": 0.2184,
1109
+ "eval_map_car": 0.3011,
1110
+ "eval_map_large": 0.8088,
1111
+ "eval_map_medium": 0.6608,
1112
+ "eval_map_person": 0.1657,
1113
+ "eval_map_small": 0.155,
1114
+ "eval_mar_1": 0.0985,
1115
+ "eval_mar_10": 0.2574,
1116
+ "eval_mar_100": 0.3013,
1117
+ "eval_mar_100_car": 0.3797,
1118
+ "eval_mar_100_person": 0.223,
1119
+ "eval_mar_large": 0.8594,
1120
+ "eval_mar_medium": 0.7164,
1121
+ "eval_mar_small": 0.2292,
1122
+ "eval_runtime": 82.7041,
1123
+ "eval_samples_per_second": 3.99,
1124
+ "eval_steps_per_second": 0.508,
1125
+ "step": 3600
1126
+ },
1127
+ {
1128
+ "epoch": 37.0,
1129
+ "grad_norm": 195.50607299804688,
1130
+ "learning_rate": 1e-05,
1131
+ "loss": 21.0417,
1132
+ "step": 3700
1133
+ },
1134
+ {
1135
+ "epoch": 37.0,
1136
+ "eval_loss": 27.249868392944336,
1137
+ "eval_map": 0.2354,
1138
+ "eval_map_50": 0.4523,
1139
+ "eval_map_75": 0.2211,
1140
+ "eval_map_car": 0.3054,
1141
+ "eval_map_large": 0.8224,
1142
+ "eval_map_medium": 0.6643,
1143
+ "eval_map_person": 0.1654,
1144
+ "eval_map_small": 0.1569,
1145
+ "eval_mar_1": 0.0998,
1146
+ "eval_mar_10": 0.2604,
1147
+ "eval_mar_100": 0.3037,
1148
+ "eval_mar_100_car": 0.3865,
1149
+ "eval_mar_100_person": 0.2209,
1150
+ "eval_mar_large": 0.8562,
1151
+ "eval_mar_medium": 0.7243,
1152
+ "eval_mar_small": 0.2307,
1153
+ "eval_runtime": 93.0917,
1154
+ "eval_samples_per_second": 3.545,
1155
+ "eval_steps_per_second": 0.451,
1156
+ "step": 3700
1157
+ },
1158
+ {
1159
+ "epoch": 38.0,
1160
+ "grad_norm": 175.56350708007812,
1161
+ "learning_rate": 1e-05,
1162
+ "loss": 21.0664,
1163
+ "step": 3800
1164
+ },
1165
+ {
1166
+ "epoch": 38.0,
1167
+ "eval_loss": 27.342622756958008,
1168
+ "eval_map": 0.2304,
1169
+ "eval_map_50": 0.4437,
1170
+ "eval_map_75": 0.2159,
1171
+ "eval_map_car": 0.2967,
1172
+ "eval_map_large": 0.8071,
1173
+ "eval_map_medium": 0.6568,
1174
+ "eval_map_person": 0.1641,
1175
+ "eval_map_small": 0.1516,
1176
+ "eval_mar_1": 0.0986,
1177
+ "eval_mar_10": 0.2566,
1178
+ "eval_mar_100": 0.2993,
1179
+ "eval_mar_100_car": 0.3788,
1180
+ "eval_mar_100_person": 0.2198,
1181
+ "eval_mar_large": 0.8451,
1182
+ "eval_mar_medium": 0.7164,
1183
+ "eval_mar_small": 0.227,
1184
+ "eval_runtime": 85.2034,
1185
+ "eval_samples_per_second": 3.873,
1186
+ "eval_steps_per_second": 0.493,
1187
+ "step": 3800
1188
+ },
1189
+ {
1190
+ "epoch": 39.0,
1191
+ "grad_norm": 115.73571014404297,
1192
+ "learning_rate": 1e-05,
1193
+ "loss": 21.0042,
1194
+ "step": 3900
1195
+ },
1196
+ {
1197
+ "epoch": 39.0,
1198
+ "eval_loss": 27.7719783782959,
1199
+ "eval_map": 0.2315,
1200
+ "eval_map_50": 0.4449,
1201
+ "eval_map_75": 0.2182,
1202
+ "eval_map_car": 0.3026,
1203
+ "eval_map_large": 0.8214,
1204
+ "eval_map_medium": 0.6611,
1205
+ "eval_map_person": 0.1604,
1206
+ "eval_map_small": 0.1528,
1207
+ "eval_mar_1": 0.0994,
1208
+ "eval_mar_10": 0.2594,
1209
+ "eval_mar_100": 0.2994,
1210
+ "eval_mar_100_car": 0.3827,
1211
+ "eval_mar_100_person": 0.2161,
1212
+ "eval_mar_large": 0.8594,
1213
+ "eval_mar_medium": 0.7191,
1214
+ "eval_mar_small": 0.2265,
1215
+ "eval_runtime": 88.1265,
1216
+ "eval_samples_per_second": 3.745,
1217
+ "eval_steps_per_second": 0.477,
1218
+ "step": 3900
1219
+ },
1220
+ {
1221
+ "epoch": 40.0,
1222
+ "grad_norm": 103.04826354980469,
1223
+ "learning_rate": 1e-05,
1224
+ "loss": 20.8548,
1225
+ "step": 4000
1226
+ },
1227
+ {
1228
+ "epoch": 40.0,
1229
+ "eval_loss": 27.549421310424805,
1230
+ "eval_map": 0.2315,
1231
+ "eval_map_50": 0.4495,
1232
+ "eval_map_75": 0.2185,
1233
+ "eval_map_car": 0.3003,
1234
+ "eval_map_large": 0.8161,
1235
+ "eval_map_medium": 0.6606,
1236
+ "eval_map_person": 0.1626,
1237
+ "eval_map_small": 0.1535,
1238
+ "eval_mar_1": 0.0981,
1239
+ "eval_mar_10": 0.2576,
1240
+ "eval_mar_100": 0.3,
1241
+ "eval_mar_100_car": 0.3776,
1242
+ "eval_mar_100_person": 0.2224,
1243
+ "eval_mar_large": 0.8618,
1244
+ "eval_mar_medium": 0.7189,
1245
+ "eval_mar_small": 0.2272,
1246
+ "eval_runtime": 102.0298,
1247
+ "eval_samples_per_second": 3.234,
1248
+ "eval_steps_per_second": 0.412,
1249
+ "step": 4000
1250
+ },
1251
+ {
1252
+ "epoch": 40.0,
1253
+ "step": 4000,
1254
+ "total_flos": 3.574777289637888e+18,
1255
+ "train_loss": 23.846087463378908,
1256
+ "train_runtime": 9658.9738,
1257
+ "train_samples_per_second": 6.626,
1258
+ "train_steps_per_second": 0.414
1259
+ }
1260
+ ],
1261
+ "logging_steps": 500,
1262
+ "max_steps": 4000,
1263
+ "num_input_tokens_seen": 0,
1264
+ "num_train_epochs": 40,
1265
+ "save_steps": 500,
1266
+ "stateful_callbacks": {
1267
+ "TrainerControl": {
1268
+ "args": {
1269
+ "should_epoch_stop": false,
1270
+ "should_evaluate": false,
1271
+ "should_log": false,
1272
+ "should_save": true,
1273
+ "should_training_stop": true
1274
+ },
1275
+ "attributes": {}
1276
+ }
1277
+ },
1278
+ "total_flos": 3.574777289637888e+18,
1279
+ "train_batch_size": 8,
1280
+ "trial_name": null,
1281
+ "trial_params": null
1282
+ }