Hugging Face's logo

jnmrr
/
rtdetr-v2-voucher-classifier

Object Detection
TensorBoard
Safetensors
rt_detr_v2
computer-vision
voucher-classification
rt-detr
rtdetrv2
Model card Files
xet
Metrics Community
jnmrr commited on
Commit
e026564
·
verified ·
1 Parent(s): 2c82cf9

Upload RT-DETRv2 voucher classifier

Browse files
Files changed (45) hide show
  1. README.md +23 -23
  2. checkpoint-2400/config.json +129 -0
  3. checkpoint-2400/model.safetensors +3 -0
  4. checkpoint-2400/optimizer.pt +3 -0
  5. checkpoint-2400/preprocessor_config.json +36 -0
  6. checkpoint-2400/rng_state.pth +3 -0
  7. checkpoint-2400/scheduler.pt +3 -0
  8. checkpoint-2400/trainer_state.json +454 -0
  9. checkpoint-2400/training_args.bin +3 -0
  10. checkpoint-2600/config.json +129 -0
  11. checkpoint-2600/model.safetensors +3 -0
  12. checkpoint-2600/optimizer.pt +3 -0
  13. checkpoint-2600/preprocessor_config.json +36 -0
  14. checkpoint-2600/rng_state.pth +3 -0
  15. checkpoint-2600/scheduler.pt +3 -0
  16. checkpoint-2600/trainer_state.json +489 -0
  17. checkpoint-2600/training_args.bin +3 -0
  18. checkpoint-2800/config.json +129 -0
  19. checkpoint-2800/model.safetensors +3 -0
  20. checkpoint-2800/optimizer.pt +3 -0
  21. checkpoint-2800/preprocessor_config.json +36 -0
  22. checkpoint-2800/rng_state.pth +3 -0
  23. checkpoint-2800/scheduler.pt +3 -0
  24. checkpoint-2800/trainer_state.json +524 -0
  25. checkpoint-2800/training_args.bin +3 -0
  26. checkpoint-3000/config.json +129 -0
  27. checkpoint-3000/model.safetensors +3 -0
  28. checkpoint-3000/optimizer.pt +3 -0
  29. checkpoint-3000/preprocessor_config.json +36 -0
  30. checkpoint-3000/rng_state.pth +3 -0
  31. checkpoint-3000/scheduler.pt +3 -0
  32. checkpoint-3000/trainer_state.json +559 -0
  33. checkpoint-3000/training_args.bin +3 -0
  34. checkpoint-3040/config.json +129 -0
  35. checkpoint-3040/model.safetensors +3 -0
  36. checkpoint-3040/optimizer.pt +3 -0
  37. checkpoint-3040/preprocessor_config.json +36 -0
  38. checkpoint-3040/rng_state.pth +3 -0
  39. checkpoint-3040/scheduler.pt +3 -0
  40. checkpoint-3040/trainer_state.json +566 -0
  41. checkpoint-3040/training_args.bin +3 -0
  42. config.json +6 -6
  43. model.safetensors +2 -2
  44. preprocessor_config.json +2 -2
  45. training_args.bin +1 -1
README.md CHANGED
@@ -1,6 +1,6 @@
1
  ---
2
  license: apache-2.0
3
- base_model: PekingU/rtdetr_v2_r50vd
4
  tags:
5
  - object-detection
6
  - computer-vision
@@ -20,13 +20,13 @@ widget:
20
 
21
  # RT-DETRv2 Fine-tuned for Voucher Classification
22
 
23
- This model is a fine-tuned version of [PekingU/rtdetr_v2_r50vd](https://huggingface.co/PekingU/rtdetr_v2_r50vd) for voucher classification and object detection.
24
 
25
  ## Model Details
26
 
27
  ### Model Description
28
  - **Model Type**: Object Detection (RT-DETRv2)
29
- - **Base Model**: PekingU/rtdetr_v2_r50vd
30
  - **Task**: Multi-class voucher classification and detection
31
  - **Classes**: 3 classes
32
  - 0: digital (digital invoices)
@@ -36,19 +36,19 @@ This model is a fine-tuned version of [PekingU/rtdetr_v2_r50vd](https://huggingf
36
  ### Training Details
37
 
38
  **Training Dataset:**
39
- - **Total Samples**: 404
40
  - **Class Distribution**:
41
- - **fisico** (id: 1): 138 samples (34.2%)
42
- - **tesoreria** (id: 2): 141 samples (34.9%)
43
- - **digital** (id: 0): 125 samples (30.9%)
44
 
45
 
46
  **Training Configuration:**
47
- - **Image Size**: 665x665
48
- - **Batch Size**: 14
49
- - **Learning Rate**: 1.4e-05
50
- - **Weight Decay**: 0.0001
51
- - **Epochs**: 2
52
  - **Validation Split**: 0.2
53
 
54
  **Data Processing:**
@@ -86,7 +86,7 @@ This model is a fine-tuned version of [PekingU/rtdetr_v2_r50vd](https://huggingf
86
  - **Tesoreria receipts**: 0.0000 (needs improvement)
87
 
88
  **Model Confidence:**
89
- - **Digital invoices mean confidence**: 0.5884 (low)
90
  - **Fisico receipts mean confidence**: 0.0000 (low)
91
  - **Tesoreria receipts mean confidence**: 0.0000 (low)
92
 
@@ -105,25 +105,25 @@ This model is a fine-tuned version of [PekingU/rtdetr_v2_r50vd](https://huggingf
105
  - **Base model**: PekingU/rtdetr_v2_r101vd
106
  - **Architecture**: rtdetr_v2_r101vd
107
  - **Input resolution**: 832×832 pixels
108
- - **Training epochs**: 2
109
- - **Batch size**: 28
110
 
111
  **Training Hardware:**
112
- - **GPU**: NVIDIA L40S
113
- - **VRAM**: 44.4 GB
114
- - **RAM**: 62.8 GB
115
- - **GPU configuration**: L40S optimized
116
 
117
- **Training Time**: 0.6 minutes
118
 
119
  **Training Summary:**
120
- - **Final training loss**: 1032.4518
121
- - **Final learning rate**: 3.44e-07
122
 
123
 
124
  ### MLflow Tracking
125
 
126
- - **MLflow Run ID**: 37c9e6eabd1a46c6851d1a53af3f67ac
127
  - **MLflow Experiment**: RT-DETRv2_Voucher_Classification
128
 
129
 
 
1
  ---
2
  license: apache-2.0
3
+ base_model: PekingU/rtdetr_v2_r101vd
4
  tags:
5
  - object-detection
6
  - computer-vision
 
20
 
21
  # RT-DETRv2 Fine-tuned for Voucher Classification
22
 
23
+ This model is a fine-tuned version of [PekingU/rtdetr_v2_r101vd](https://huggingface.co/PekingU/rtdetr_v2_r101vd) for voucher classification and object detection.
24
 
25
  ## Model Details
26
 
27
  ### Model Description
28
  - **Model Type**: Object Detection (RT-DETRv2)
29
+ - **Base Model**: PekingU/rtdetr_v2_r101vd
30
  - **Task**: Multi-class voucher classification and detection
31
  - **Classes**: 3 classes
32
  - 0: digital (digital invoices)
 
36
  ### Training Details
37
 
38
  **Training Dataset:**
39
+ - **Total Samples**: 1227
40
  - **Class Distribution**:
41
+ - **tesoreria** (id: 2): 405 samples (33.0%)
42
+ - **fisico** (id: 1): 416 samples (33.9%)
43
+ - **digital** (id: 0): 406 samples (33.1%)
44
 
45
 
46
  **Training Configuration:**
47
+ - **Image Size**: 832x832
48
+ - **Batch Size**: 32
49
+ - **Learning Rate**: 3e-05
50
+ - **Weight Decay**: 0.01
51
+ - **Epochs**: 80
52
  - **Validation Split**: 0.2
53
 
54
  **Data Processing:**
 
86
  - **Tesoreria receipts**: 0.0000 (needs improvement)
87
 
88
  **Model Confidence:**
89
+ - **Digital invoices mean confidence**: 0.4346 (low)
90
  - **Fisico receipts mean confidence**: 0.0000 (low)
91
  - **Tesoreria receipts mean confidence**: 0.0000 (low)
92
 
 
105
  - **Base model**: PekingU/rtdetr_v2_r101vd
106
  - **Architecture**: rtdetr_v2_r101vd
107
  - **Input resolution**: 832×832 pixels
108
+ - **Training epochs**: 80
109
+ - **Batch size**: 32
110
 
111
  **Training Hardware:**
112
+ - **GPU**: NVIDIA H100 80GB HBM3
113
+ - **VRAM**: 79.2 GB
114
+ - **RAM**: 235.9 GB
115
+ - **GPU configuration**: H100 optimized
116
 
117
+ **Training Time**: 39.6 minutes
118
 
119
  **Training Summary:**
120
+ - **Final training loss**: 4.9881
121
+ - **Final learning rate**: 2.08e-08
122
 
123
 
124
  ### MLflow Tracking
125
 
126
+ - **MLflow Run ID**: 1690d8d04ea74ca99f0fea73a8466f83
127
  - **MLflow Experiment**: RT-DETRv2_Voucher_Classification
128
 
129
 
checkpoint-2400/config.json ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "silu",
4
+ "anchor_image_size": null,
5
+ "architectures": [
6
+ "RTDetrV2ForObjectDetection"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "auxiliary_loss": true,
10
+ "backbone": null,
11
+ "backbone_config": {
12
+ "depths": [
13
+ 3,
14
+ 4,
15
+ 23,
16
+ 3
17
+ ],
18
+ "downsample_in_bottleneck": false,
19
+ "downsample_in_first_stage": false,
20
+ "embedding_size": 64,
21
+ "hidden_act": "relu",
22
+ "hidden_sizes": [
23
+ 256,
24
+ 512,
25
+ 1024,
26
+ 2048
27
+ ],
28
+ "layer_type": "bottleneck",
29
+ "model_type": "rt_detr_resnet",
30
+ "num_channels": 3,
31
+ "out_features": [
32
+ "stage2",
33
+ "stage3",
34
+ "stage4"
35
+ ],
36
+ "out_indices": [
37
+ 2,
38
+ 3,
39
+ 4
40
+ ],
41
+ "stage_names": [
42
+ "stem",
43
+ "stage1",
44
+ "stage2",
45
+ "stage3",
46
+ "stage4"
47
+ ],
48
+ "torch_dtype": "float32"
49
+ },
50
+ "backbone_kwargs": null,
51
+ "batch_norm_eps": 1e-05,
52
+ "box_noise_scale": 1.0,
53
+ "d_model": 256,
54
+ "decoder_activation_function": "relu",
55
+ "decoder_attention_heads": 8,
56
+ "decoder_ffn_dim": 1024,
57
+ "decoder_in_channels": [
58
+ 384,
59
+ 384,
60
+ 384
61
+ ],
62
+ "decoder_layers": 6,
63
+ "decoder_method": "default",
64
+ "decoder_n_levels": 3,
65
+ "decoder_n_points": 4,
66
+ "decoder_offset_scale": 0.5,
67
+ "disable_custom_kernels": true,
68
+ "dropout": 0.0,
69
+ "encode_proj_layers": [
70
+ 2
71
+ ],
72
+ "encoder_activation_function": "gelu",
73
+ "encoder_attention_heads": 8,
74
+ "encoder_ffn_dim": 2048,
75
+ "encoder_hidden_dim": 384,
76
+ "encoder_in_channels": [
77
+ 512,
78
+ 1024,
79
+ 2048
80
+ ],
81
+ "encoder_layers": 1,
82
+ "eos_coefficient": 0.0001,
83
+ "eval_size": null,
84
+ "feat_strides": [
85
+ 8,
86
+ 16,
87
+ 32
88
+ ],
89
+ "focal_loss_alpha": 0.75,
90
+ "focal_loss_gamma": 2.0,
91
+ "freeze_backbone_batch_norms": true,
92
+ "hidden_expansion": 1.0,
93
+ "id2label": {
94
+ "0": "digital",
95
+ "1": "fisico",
96
+ "2": "tesoreria"
97
+ },
98
+ "initializer_bias_prior_prob": null,
99
+ "initializer_range": 0.01,
100
+ "is_encoder_decoder": true,
101
+ "label2id": {
102
+ "digital": 0,
103
+ "fisico": 1,
104
+ "tesoreria": 2
105
+ },
106
+ "label_noise_ratio": 0.5,
107
+ "layer_norm_eps": 1e-05,
108
+ "learn_initial_query": false,
109
+ "matcher_alpha": 0.25,
110
+ "matcher_bbox_cost": 5.0,
111
+ "matcher_class_cost": 2.0,
112
+ "matcher_gamma": 2.0,
113
+ "matcher_giou_cost": 2.0,
114
+ "model_type": "rt_detr_v2",
115
+ "normalize_before": false,
116
+ "num_denoising": 100,
117
+ "num_feature_levels": 3,
118
+ "num_queries": 300,
119
+ "positional_encoding_temperature": 10000,
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.55.2",
122
+ "use_focal_loss": true,
123
+ "use_pretrained_backbone": false,
124
+ "use_timm_backbone": false,
125
+ "weight_loss_bbox": 5.0,
126
+ "weight_loss_giou": 2.0,
127
+ "weight_loss_vfl": 1.0,
128
+ "with_box_refine": true
129
+ }
checkpoint-2400/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7958de6d52bccdb53feb8d8ee22b493e29cbc2482680ecbdf2f542ffa20dfc7
3
+ size 306699044
checkpoint-2400/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9406c2c81fc162abc8e60f14b83c38383bd677e0cef15921a7a3988ce9ae8a6
3
+ size 611580898
checkpoint-2400/preprocessor_config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": false,
5
+ "device": null,
6
+ "disable_grouping": null,
7
+ "do_center_crop": null,
8
+ "do_convert_annotations": true,
9
+ "do_convert_rgb": null,
10
+ "do_normalize": false,
11
+ "do_pad": false,
12
+ "do_rescale": true,
13
+ "do_resize": true,
14
+ "format": "coco_detection",
15
+ "image_mean": [
16
+ 0.485,
17
+ 0.456,
18
+ 0.406
19
+ ],
20
+ "image_processor_type": "RTDetrImageProcessorFast",
21
+ "image_std": [
22
+ 0.229,
23
+ 0.224,
24
+ 0.225
25
+ ],
26
+ "input_data_format": null,
27
+ "pad_size": null,
28
+ "resample": 2,
29
+ "rescale_factor": 0.00392156862745098,
30
+ "return_segmentation_masks": null,
31
+ "return_tensors": null,
32
+ "size": {
33
+ "height": 832,
34
+ "width": 832
35
+ }
36
+ }
checkpoint-2400/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f346ba2198f89be9de19ef055630cfa547efce46fa785baa5f84493468e2fd3
3
+ size 14645
checkpoint-2400/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fdfc302a3829065b6dfe3e7d55bb2c36cf51ab0d6b08f8db7206a7f9999d327a
3
+ size 1465
checkpoint-2400/trainer_state.json ADDED
@@ -0,0 +1,454 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 63.1578947368421,
6
+ "eval_steps": 500,
7
+ "global_step": 2400,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 1.0526315789473684,
14
+ "grad_norm": 8679.8642578125,
15
+ "learning_rate": 7.312500000000001e-07,
16
+ "loss": 1384.9168,
17
+ "step": 40
18
+ },
19
+ {
20
+ "epoch": 2.1052631578947367,
21
+ "grad_norm": 4251.486328125,
22
+ "learning_rate": 1.48125e-06,
23
+ "loss": 1290.1081,
24
+ "step": 80
25
+ },
26
+ {
27
+ "epoch": 3.1578947368421053,
28
+ "grad_norm": 4015.17138671875,
29
+ "learning_rate": 2.23125e-06,
30
+ "loss": 1097.6352,
31
+ "step": 120
32
+ },
33
+ {
34
+ "epoch": 4.2105263157894735,
35
+ "grad_norm": 2110.156982421875,
36
+ "learning_rate": 2.98125e-06,
37
+ "loss": 846.727,
38
+ "step": 160
39
+ },
40
+ {
41
+ "epoch": 5.2631578947368425,
42
+ "grad_norm": 1568.4326171875,
43
+ "learning_rate": 3.73125e-06,
44
+ "loss": 578.2698,
45
+ "step": 200
46
+ },
47
+ {
48
+ "epoch": 6.315789473684211,
49
+ "grad_norm": 1924.275146484375,
50
+ "learning_rate": 4.4812500000000005e-06,
51
+ "loss": 412.4179,
52
+ "step": 240
53
+ },
54
+ {
55
+ "epoch": 7.368421052631579,
56
+ "grad_norm": 862.5057373046875,
57
+ "learning_rate": 5.23125e-06,
58
+ "loss": 311.6504,
59
+ "step": 280
60
+ },
61
+ {
62
+ "epoch": 8.421052631578947,
63
+ "grad_norm": 1463.6671142578125,
64
+ "learning_rate": 5.98125e-06,
65
+ "loss": 251.8019,
66
+ "step": 320
67
+ },
68
+ {
69
+ "epoch": 9.473684210526315,
70
+ "grad_norm": 1956.26708984375,
71
+ "learning_rate": 6.73125e-06,
72
+ "loss": 197.8886,
73
+ "step": 360
74
+ },
75
+ {
76
+ "epoch": 10.526315789473685,
77
+ "grad_norm": 404.4883728027344,
78
+ "learning_rate": 7.481250000000001e-06,
79
+ "loss": 155.3046,
80
+ "step": 400
81
+ },
82
+ {
83
+ "epoch": 11.578947368421053,
84
+ "grad_norm": 319.11956787109375,
85
+ "learning_rate": 8.23125e-06,
86
+ "loss": 127.7263,
87
+ "step": 440
88
+ },
89
+ {
90
+ "epoch": 12.631578947368421,
91
+ "grad_norm": 289.92144775390625,
92
+ "learning_rate": 8.98125e-06,
93
+ "loss": 101.4958,
94
+ "step": 480
95
+ },
96
+ {
97
+ "epoch": 13.68421052631579,
98
+ "grad_norm": 238.7841033935547,
99
+ "learning_rate": 9.731250000000001e-06,
100
+ "loss": 80.8442,
101
+ "step": 520
102
+ },
103
+ {
104
+ "epoch": 14.736842105263158,
105
+ "grad_norm": 194.82733154296875,
106
+ "learning_rate": 1.048125e-05,
107
+ "loss": 62.7129,
108
+ "step": 560
109
+ },
110
+ {
111
+ "epoch": 15.789473684210526,
112
+ "grad_norm": 154.37205505371094,
113
+ "learning_rate": 1.123125e-05,
114
+ "loss": 50.3538,
115
+ "step": 600
116
+ },
117
+ {
118
+ "epoch": 16.842105263157894,
119
+ "grad_norm": 208.374267578125,
120
+ "learning_rate": 1.198125e-05,
121
+ "loss": 41.4226,
122
+ "step": 640
123
+ },
124
+ {
125
+ "epoch": 17.894736842105264,
126
+ "grad_norm": 101.1446304321289,
127
+ "learning_rate": 1.2731250000000001e-05,
128
+ "loss": 32.4618,
129
+ "step": 680
130
+ },
131
+ {
132
+ "epoch": 18.94736842105263,
133
+ "grad_norm": 147.55865478515625,
134
+ "learning_rate": 1.348125e-05,
135
+ "loss": 27.0828,
136
+ "step": 720
137
+ },
138
+ {
139
+ "epoch": 20.0,
140
+ "grad_norm": 112.81380462646484,
141
+ "learning_rate": 1.423125e-05,
142
+ "loss": 22.4964,
143
+ "step": 760
144
+ },
145
+ {
146
+ "epoch": 21.05263157894737,
147
+ "grad_norm": 73.30062866210938,
148
+ "learning_rate": 1.4981250000000002e-05,
149
+ "loss": 18.7236,
150
+ "step": 800
151
+ },
152
+ {
153
+ "epoch": 22.105263157894736,
154
+ "grad_norm": 64.0571517944336,
155
+ "learning_rate": 1.5731250000000003e-05,
156
+ "loss": 16.3388,
157
+ "step": 840
158
+ },
159
+ {
160
+ "epoch": 23.157894736842106,
161
+ "grad_norm": 76.62981414794922,
162
+ "learning_rate": 1.6481249999999997e-05,
163
+ "loss": 14.0038,
164
+ "step": 880
165
+ },
166
+ {
167
+ "epoch": 24.210526315789473,
168
+ "grad_norm": 83.39350891113281,
169
+ "learning_rate": 1.723125e-05,
170
+ "loss": 12.6697,
171
+ "step": 920
172
+ },
173
+ {
174
+ "epoch": 25.263157894736842,
175
+ "grad_norm": 49.30060577392578,
176
+ "learning_rate": 1.798125e-05,
177
+ "loss": 11.5521,
178
+ "step": 960
179
+ },
180
+ {
181
+ "epoch": 26.31578947368421,
182
+ "grad_norm": 49.78151321411133,
183
+ "learning_rate": 1.873125e-05,
184
+ "loss": 10.8067,
185
+ "step": 1000
186
+ },
187
+ {
188
+ "epoch": 27.36842105263158,
189
+ "grad_norm": 64.97428894042969,
190
+ "learning_rate": 1.9481250000000003e-05,
191
+ "loss": 9.9502,
192
+ "step": 1040
193
+ },
194
+ {
195
+ "epoch": 28.42105263157895,
196
+ "grad_norm": 96.64849853515625,
197
+ "learning_rate": 2.023125e-05,
198
+ "loss": 9.6148,
199
+ "step": 1080
200
+ },
201
+ {
202
+ "epoch": 29.473684210526315,
203
+ "grad_norm": 65.55039978027344,
204
+ "learning_rate": 2.098125e-05,
205
+ "loss": 9.462,
206
+ "step": 1120
207
+ },
208
+ {
209
+ "epoch": 30.526315789473685,
210
+ "grad_norm": 85.08692169189453,
211
+ "learning_rate": 2.173125e-05,
212
+ "loss": 8.8002,
213
+ "step": 1160
214
+ },
215
+ {
216
+ "epoch": 31.57894736842105,
217
+ "grad_norm": 124.7302017211914,
218
+ "learning_rate": 2.248125e-05,
219
+ "loss": 8.7508,
220
+ "step": 1200
221
+ },
222
+ {
223
+ "epoch": 32.63157894736842,
224
+ "grad_norm": 152.3477020263672,
225
+ "learning_rate": 2.3231250000000002e-05,
226
+ "loss": 8.5803,
227
+ "step": 1240
228
+ },
229
+ {
230
+ "epoch": 33.68421052631579,
231
+ "grad_norm": 162.75997924804688,
232
+ "learning_rate": 2.398125e-05,
233
+ "loss": 8.1324,
234
+ "step": 1280
235
+ },
236
+ {
237
+ "epoch": 34.73684210526316,
238
+ "grad_norm": 78.276611328125,
239
+ "learning_rate": 2.4731249999999998e-05,
240
+ "loss": 8.4976,
241
+ "step": 1320
242
+ },
243
+ {
244
+ "epoch": 35.78947368421053,
245
+ "grad_norm": 151.76458740234375,
246
+ "learning_rate": 2.548125e-05,
247
+ "loss": 8.9999,
248
+ "step": 1360
249
+ },
250
+ {
251
+ "epoch": 36.8421052631579,
252
+ "grad_norm": 67.1139144897461,
253
+ "learning_rate": 2.623125e-05,
254
+ "loss": 8.0035,
255
+ "step": 1400
256
+ },
257
+ {
258
+ "epoch": 37.89473684210526,
259
+ "grad_norm": 61.43220901489258,
260
+ "learning_rate": 2.6981250000000002e-05,
261
+ "loss": 7.8756,
262
+ "step": 1440
263
+ },
264
+ {
265
+ "epoch": 38.94736842105263,
266
+ "grad_norm": 68.46588134765625,
267
+ "learning_rate": 2.773125e-05,
268
+ "loss": 7.6236,
269
+ "step": 1480
270
+ },
271
+ {
272
+ "epoch": 40.0,
273
+ "grad_norm": 52.86519241333008,
274
+ "learning_rate": 2.848125e-05,
275
+ "loss": 7.3731,
276
+ "step": 1520
277
+ },
278
+ {
279
+ "epoch": 41.05263157894737,
280
+ "grad_norm": 32.2613525390625,
281
+ "learning_rate": 2.923125e-05,
282
+ "loss": 7.5533,
283
+ "step": 1560
284
+ },
285
+ {
286
+ "epoch": 42.10526315789474,
287
+ "grad_norm": 28.07533836364746,
288
+ "learning_rate": 2.998125e-05,
289
+ "loss": 7.2446,
290
+ "step": 1600
291
+ },
292
+ {
293
+ "epoch": 43.1578947368421,
294
+ "grad_norm": 51.62834167480469,
295
+ "learning_rate": 2.91875e-05,
296
+ "loss": 7.2844,
297
+ "step": 1640
298
+ },
299
+ {
300
+ "epoch": 44.21052631578947,
301
+ "grad_norm": 94.51573944091797,
302
+ "learning_rate": 2.8354166666666667e-05,
303
+ "loss": 7.3347,
304
+ "step": 1680
305
+ },
306
+ {
307
+ "epoch": 45.26315789473684,
308
+ "grad_norm": 33.503990173339844,
309
+ "learning_rate": 2.7520833333333333e-05,
310
+ "loss": 7.1209,
311
+ "step": 1720
312
+ },
313
+ {
314
+ "epoch": 46.31578947368421,
315
+ "grad_norm": 74.56465148925781,
316
+ "learning_rate": 2.6687499999999998e-05,
317
+ "loss": 6.7968,
318
+ "step": 1760
319
+ },
320
+ {
321
+ "epoch": 47.36842105263158,
322
+ "grad_norm": 32.059810638427734,
323
+ "learning_rate": 2.5854166666666667e-05,
324
+ "loss": 6.8781,
325
+ "step": 1800
326
+ },
327
+ {
328
+ "epoch": 48.421052631578945,
329
+ "grad_norm": 97.38982391357422,
330
+ "learning_rate": 2.5020833333333336e-05,
331
+ "loss": 6.7608,
332
+ "step": 1840
333
+ },
334
+ {
335
+ "epoch": 49.473684210526315,
336
+ "grad_norm": 34.4022102355957,
337
+ "learning_rate": 2.4187500000000002e-05,
338
+ "loss": 6.4979,
339
+ "step": 1880
340
+ },
341
+ {
342
+ "epoch": 50.526315789473685,
343
+ "grad_norm": 35.79924011230469,
344
+ "learning_rate": 2.3354166666666667e-05,
345
+ "loss": 6.5747,
346
+ "step": 1920
347
+ },
348
+ {
349
+ "epoch": 51.578947368421055,
350
+ "grad_norm": 100.87942504882812,
351
+ "learning_rate": 2.2520833333333333e-05,
352
+ "loss": 6.1757,
353
+ "step": 1960
354
+ },
355
+ {
356
+ "epoch": 52.63157894736842,
357
+ "grad_norm": 159.83860778808594,
358
+ "learning_rate": 2.16875e-05,
359
+ "loss": 6.4399,
360
+ "step": 2000
361
+ },
362
+ {
363
+ "epoch": 53.68421052631579,
364
+ "grad_norm": 58.15259552001953,
365
+ "learning_rate": 2.0854166666666668e-05,
366
+ "loss": 6.3292,
367
+ "step": 2040
368
+ },
369
+ {
370
+ "epoch": 54.73684210526316,
371
+ "grad_norm": 23.384042739868164,
372
+ "learning_rate": 2.0020833333333333e-05,
373
+ "loss": 6.2464,
374
+ "step": 2080
375
+ },
376
+ {
377
+ "epoch": 55.78947368421053,
378
+ "grad_norm": 22.12693977355957,
379
+ "learning_rate": 1.91875e-05,
380
+ "loss": 6.1714,
381
+ "step": 2120
382
+ },
383
+ {
384
+ "epoch": 56.8421052631579,
385
+ "grad_norm": 17.758420944213867,
386
+ "learning_rate": 1.8354166666666668e-05,
387
+ "loss": 5.9527,
388
+ "step": 2160
389
+ },
390
+ {
391
+ "epoch": 57.89473684210526,
392
+ "grad_norm": 39.71998977661133,
393
+ "learning_rate": 1.7520833333333333e-05,
394
+ "loss": 5.9018,
395
+ "step": 2200
396
+ },
397
+ {
398
+ "epoch": 58.94736842105263,
399
+ "grad_norm": 28.227684020996094,
400
+ "learning_rate": 1.6687500000000002e-05,
401
+ "loss": 5.8687,
402
+ "step": 2240
403
+ },
404
+ {
405
+ "epoch": 60.0,
406
+ "grad_norm": 18.77845573425293,
407
+ "learning_rate": 1.5854166666666668e-05,
408
+ "loss": 5.7978,
409
+ "step": 2280
410
+ },
411
+ {
412
+ "epoch": 61.05263157894737,
413
+ "grad_norm": 25.35810089111328,
414
+ "learning_rate": 1.5020833333333334e-05,
415
+ "loss": 5.8052,
416
+ "step": 2320
417
+ },
418
+ {
419
+ "epoch": 62.10526315789474,
420
+ "grad_norm": 92.96112823486328,
421
+ "learning_rate": 1.41875e-05,
422
+ "loss": 5.7386,
423
+ "step": 2360
424
+ },
425
+ {
426
+ "epoch": 63.1578947368421,
427
+ "grad_norm": 35.39582824707031,
428
+ "learning_rate": 1.3354166666666667e-05,
429
+ "loss": 5.6511,
430
+ "step": 2400
431
+ }
432
+ ],
433
+ "logging_steps": 40,
434
+ "max_steps": 3040,
435
+ "num_input_tokens_seen": 0,
436
+ "num_train_epochs": 80,
437
+ "save_steps": 200,
438
+ "stateful_callbacks": {
439
+ "TrainerControl": {
440
+ "args": {
441
+ "should_epoch_stop": false,
442
+ "should_evaluate": false,
443
+ "should_log": false,
444
+ "should_save": true,
445
+ "should_training_stop": false
446
+ },
447
+ "attributes": {}
448
+ }
449
+ },
450
+ "total_flos": 7.310662478164132e+19,
451
+ "train_batch_size": 32,
452
+ "trial_name": null,
453
+ "trial_params": null
454
+ }
checkpoint-2400/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a669420e271b606473a6402f32de0daffc871134717e7055e89961f40d85f8d
3
+ size 5777
checkpoint-2600/config.json ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "silu",
4
+ "anchor_image_size": null,
5
+ "architectures": [
6
+ "RTDetrV2ForObjectDetection"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "auxiliary_loss": true,
10
+ "backbone": null,
11
+ "backbone_config": {
12
+ "depths": [
13
+ 3,
14
+ 4,
15
+ 23,
16
+ 3
17
+ ],
18
+ "downsample_in_bottleneck": false,
19
+ "downsample_in_first_stage": false,
20
+ "embedding_size": 64,
21
+ "hidden_act": "relu",
22
+ "hidden_sizes": [
23
+ 256,
24
+ 512,
25
+ 1024,
26
+ 2048
27
+ ],
28
+ "layer_type": "bottleneck",
29
+ "model_type": "rt_detr_resnet",
30
+ "num_channels": 3,
31
+ "out_features": [
32
+ "stage2",
33
+ "stage3",
34
+ "stage4"
35
+ ],
36
+ "out_indices": [
37
+ 2,
38
+ 3,
39
+ 4
40
+ ],
41
+ "stage_names": [
42
+ "stem",
43
+ "stage1",
44
+ "stage2",
45
+ "stage3",
46
+ "stage4"
47
+ ],
48
+ "torch_dtype": "float32"
49
+ },
50
+ "backbone_kwargs": null,
51
+ "batch_norm_eps": 1e-05,
52
+ "box_noise_scale": 1.0,
53
+ "d_model": 256,
54
+ "decoder_activation_function": "relu",
55
+ "decoder_attention_heads": 8,
56
+ "decoder_ffn_dim": 1024,
57
+ "decoder_in_channels": [
58
+ 384,
59
+ 384,
60
+ 384
61
+ ],
62
+ "decoder_layers": 6,
63
+ "decoder_method": "default",
64
+ "decoder_n_levels": 3,
65
+ "decoder_n_points": 4,
66
+ "decoder_offset_scale": 0.5,
67
+ "disable_custom_kernels": true,
68
+ "dropout": 0.0,
69
+ "encode_proj_layers": [
70
+ 2
71
+ ],
72
+ "encoder_activation_function": "gelu",
73
+ "encoder_attention_heads": 8,
74
+ "encoder_ffn_dim": 2048,
75
+ "encoder_hidden_dim": 384,
76
+ "encoder_in_channels": [
77
+ 512,
78
+ 1024,
79
+ 2048
80
+ ],
81
+ "encoder_layers": 1,
82
+ "eos_coefficient": 0.0001,
83
+ "eval_size": null,
84
+ "feat_strides": [
85
+ 8,
86
+ 16,
87
+ 32
88
+ ],
89
+ "focal_loss_alpha": 0.75,
90
+ "focal_loss_gamma": 2.0,
91
+ "freeze_backbone_batch_norms": true,
92
+ "hidden_expansion": 1.0,
93
+ "id2label": {
94
+ "0": "digital",
95
+ "1": "fisico",
96
+ "2": "tesoreria"
97
+ },
98
+ "initializer_bias_prior_prob": null,
99
+ "initializer_range": 0.01,
100
+ "is_encoder_decoder": true,
101
+ "label2id": {
102
+ "digital": 0,
103
+ "fisico": 1,
104
+ "tesoreria": 2
105
+ },
106
+ "label_noise_ratio": 0.5,
107
+ "layer_norm_eps": 1e-05,
108
+ "learn_initial_query": false,
109
+ "matcher_alpha": 0.25,
110
+ "matcher_bbox_cost": 5.0,
111
+ "matcher_class_cost": 2.0,
112
+ "matcher_gamma": 2.0,
113
+ "matcher_giou_cost": 2.0,
114
+ "model_type": "rt_detr_v2",
115
+ "normalize_before": false,
116
+ "num_denoising": 100,
117
+ "num_feature_levels": 3,
118
+ "num_queries": 300,
119
+ "positional_encoding_temperature": 10000,
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.55.2",
122
+ "use_focal_loss": true,
123
+ "use_pretrained_backbone": false,
124
+ "use_timm_backbone": false,
125
+ "weight_loss_bbox": 5.0,
126
+ "weight_loss_giou": 2.0,
127
+ "weight_loss_vfl": 1.0,
128
+ "with_box_refine": true
129
+ }
checkpoint-2600/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:407e71227c927acccacc98c51fd74ab85477e49646edeb1260d5177330d1324d
3
+ size 306699044
checkpoint-2600/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a66836dfcd7995cd018e7229c1bc220162a654211d692794c1c86334448bcda
3
+ size 611580898
checkpoint-2600/preprocessor_config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": false,
5
+ "device": null,
6
+ "disable_grouping": null,
7
+ "do_center_crop": null,
8
+ "do_convert_annotations": true,
9
+ "do_convert_rgb": null,
10
+ "do_normalize": false,
11
+ "do_pad": false,
12
+ "do_rescale": true,
13
+ "do_resize": true,
14
+ "format": "coco_detection",
15
+ "image_mean": [
16
+ 0.485,
17
+ 0.456,
18
+ 0.406
19
+ ],
20
+ "image_processor_type": "RTDetrImageProcessorFast",
21
+ "image_std": [
22
+ 0.229,
23
+ 0.224,
24
+ 0.225
25
+ ],
26
+ "input_data_format": null,
27
+ "pad_size": null,
28
+ "resample": 2,
29
+ "rescale_factor": 0.00392156862745098,
30
+ "return_segmentation_masks": null,
31
+ "return_tensors": null,
32
+ "size": {
33
+ "height": 832,
34
+ "width": 832
35
+ }
36
+ }
checkpoint-2600/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1f0662f3673d67a6a2182ba94c23332a7f3d815c3be3454967df6bc3583c60b
3
+ size 14645
checkpoint-2600/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9c00028a1341c9a20e88a97e440a9cf3a65df6c7539855962a5e9f6a23971d3
3
+ size 1465
checkpoint-2600/trainer_state.json ADDED
@@ -0,0 +1,489 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 68.42105263157895,
6
+ "eval_steps": 500,
7
+ "global_step": 2600,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 1.0526315789473684,
14
+ "grad_norm": 8679.8642578125,
15
+ "learning_rate": 7.312500000000001e-07,
16
+ "loss": 1384.9168,
17
+ "step": 40
18
+ },
19
+ {
20
+ "epoch": 2.1052631578947367,
21
+ "grad_norm": 4251.486328125,
22
+ "learning_rate": 1.48125e-06,
23
+ "loss": 1290.1081,
24
+ "step": 80
25
+ },
26
+ {
27
+ "epoch": 3.1578947368421053,
28
+ "grad_norm": 4015.17138671875,
29
+ "learning_rate": 2.23125e-06,
30
+ "loss": 1097.6352,
31
+ "step": 120
32
+ },
33
+ {
34
+ "epoch": 4.2105263157894735,
35
+ "grad_norm": 2110.156982421875,
36
+ "learning_rate": 2.98125e-06,
37
+ "loss": 846.727,
38
+ "step": 160
39
+ },
40
+ {
41
+ "epoch": 5.2631578947368425,
42
+ "grad_norm": 1568.4326171875,
43
+ "learning_rate": 3.73125e-06,
44
+ "loss": 578.2698,
45
+ "step": 200
46
+ },
47
+ {
48
+ "epoch": 6.315789473684211,
49
+ "grad_norm": 1924.275146484375,
50
+ "learning_rate": 4.4812500000000005e-06,
51
+ "loss": 412.4179,
52
+ "step": 240
53
+ },
54
+ {
55
+ "epoch": 7.368421052631579,
56
+ "grad_norm": 862.5057373046875,
57
+ "learning_rate": 5.23125e-06,
58
+ "loss": 311.6504,
59
+ "step": 280
60
+ },
61
+ {
62
+ "epoch": 8.421052631578947,
63
+ "grad_norm": 1463.6671142578125,
64
+ "learning_rate": 5.98125e-06,
65
+ "loss": 251.8019,
66
+ "step": 320
67
+ },
68
+ {
69
+ "epoch": 9.473684210526315,
70
+ "grad_norm": 1956.26708984375,
71
+ "learning_rate": 6.73125e-06,
72
+ "loss": 197.8886,
73
+ "step": 360
74
+ },
75
+ {
76
+ "epoch": 10.526315789473685,
77
+ "grad_norm": 404.4883728027344,
78
+ "learning_rate": 7.481250000000001e-06,
79
+ "loss": 155.3046,
80
+ "step": 400
81
+ },
82
+ {
83
+ "epoch": 11.578947368421053,
84
+ "grad_norm": 319.11956787109375,
85
+ "learning_rate": 8.23125e-06,
86
+ "loss": 127.7263,
87
+ "step": 440
88
+ },
89
+ {
90
+ "epoch": 12.631578947368421,
91
+ "grad_norm": 289.92144775390625,
92
+ "learning_rate": 8.98125e-06,
93
+ "loss": 101.4958,
94
+ "step": 480
95
+ },
96
+ {
97
+ "epoch": 13.68421052631579,
98
+ "grad_norm": 238.7841033935547,
99
+ "learning_rate": 9.731250000000001e-06,
100
+ "loss": 80.8442,
101
+ "step": 520
102
+ },
103
+ {
104
+ "epoch": 14.736842105263158,
105
+ "grad_norm": 194.82733154296875,
106
+ "learning_rate": 1.048125e-05,
107
+ "loss": 62.7129,
108
+ "step": 560
109
+ },
110
+ {
111
+ "epoch": 15.789473684210526,
112
+ "grad_norm": 154.37205505371094,
113
+ "learning_rate": 1.123125e-05,
114
+ "loss": 50.3538,
115
+ "step": 600
116
+ },
117
+ {
118
+ "epoch": 16.842105263157894,
119
+ "grad_norm": 208.374267578125,
120
+ "learning_rate": 1.198125e-05,
121
+ "loss": 41.4226,
122
+ "step": 640
123
+ },
124
+ {
125
+ "epoch": 17.894736842105264,
126
+ "grad_norm": 101.1446304321289,
127
+ "learning_rate": 1.2731250000000001e-05,
128
+ "loss": 32.4618,
129
+ "step": 680
130
+ },
131
+ {
132
+ "epoch": 18.94736842105263,
133
+ "grad_norm": 147.55865478515625,
134
+ "learning_rate": 1.348125e-05,
135
+ "loss": 27.0828,
136
+ "step": 720
137
+ },
138
+ {
139
+ "epoch": 20.0,
140
+ "grad_norm": 112.81380462646484,
141
+ "learning_rate": 1.423125e-05,
142
+ "loss": 22.4964,
143
+ "step": 760
144
+ },
145
+ {
146
+ "epoch": 21.05263157894737,
147
+ "grad_norm": 73.30062866210938,
148
+ "learning_rate": 1.4981250000000002e-05,
149
+ "loss": 18.7236,
150
+ "step": 800
151
+ },
152
+ {
153
+ "epoch": 22.105263157894736,
154
+ "grad_norm": 64.0571517944336,
155
+ "learning_rate": 1.5731250000000003e-05,
156
+ "loss": 16.3388,
157
+ "step": 840
158
+ },
159
+ {
160
+ "epoch": 23.157894736842106,
161
+ "grad_norm": 76.62981414794922,
162
+ "learning_rate": 1.6481249999999997e-05,
163
+ "loss": 14.0038,
164
+ "step": 880
165
+ },
166
+ {
167
+ "epoch": 24.210526315789473,
168
+ "grad_norm": 83.39350891113281,
169
+ "learning_rate": 1.723125e-05,
170
+ "loss": 12.6697,
171
+ "step": 920
172
+ },
173
+ {
174
+ "epoch": 25.263157894736842,
175
+ "grad_norm": 49.30060577392578,
176
+ "learning_rate": 1.798125e-05,
177
+ "loss": 11.5521,
178
+ "step": 960
179
+ },
180
+ {
181
+ "epoch": 26.31578947368421,
182
+ "grad_norm": 49.78151321411133,
183
+ "learning_rate": 1.873125e-05,
184
+ "loss": 10.8067,
185
+ "step": 1000
186
+ },
187
+ {
188
+ "epoch": 27.36842105263158,
189
+ "grad_norm": 64.97428894042969,
190
+ "learning_rate": 1.9481250000000003e-05,
191
+ "loss": 9.9502,
192
+ "step": 1040
193
+ },
194
+ {
195
+ "epoch": 28.42105263157895,
196
+ "grad_norm": 96.64849853515625,
197
+ "learning_rate": 2.023125e-05,
198
+ "loss": 9.6148,
199
+ "step": 1080
200
+ },
201
+ {
202
+ "epoch": 29.473684210526315,
203
+ "grad_norm": 65.55039978027344,
204
+ "learning_rate": 2.098125e-05,
205
+ "loss": 9.462,
206
+ "step": 1120
207
+ },
208
+ {
209
+ "epoch": 30.526315789473685,
210
+ "grad_norm": 85.08692169189453,
211
+ "learning_rate": 2.173125e-05,
212
+ "loss": 8.8002,
213
+ "step": 1160
214
+ },
215
+ {
216
+ "epoch": 31.57894736842105,
217
+ "grad_norm": 124.7302017211914,
218
+ "learning_rate": 2.248125e-05,
219
+ "loss": 8.7508,
220
+ "step": 1200
221
+ },
222
+ {
223
+ "epoch": 32.63157894736842,
224
+ "grad_norm": 152.3477020263672,
225
+ "learning_rate": 2.3231250000000002e-05,
226
+ "loss": 8.5803,
227
+ "step": 1240
228
+ },
229
+ {
230
+ "epoch": 33.68421052631579,
231
+ "grad_norm": 162.75997924804688,
232
+ "learning_rate": 2.398125e-05,
233
+ "loss": 8.1324,
234
+ "step": 1280
235
+ },
236
+ {
237
+ "epoch": 34.73684210526316,
238
+ "grad_norm": 78.276611328125,
239
+ "learning_rate": 2.4731249999999998e-05,
240
+ "loss": 8.4976,
241
+ "step": 1320
242
+ },
243
+ {
244
+ "epoch": 35.78947368421053,
245
+ "grad_norm": 151.76458740234375,
246
+ "learning_rate": 2.548125e-05,
247
+ "loss": 8.9999,
248
+ "step": 1360
249
+ },
250
+ {
251
+ "epoch": 36.8421052631579,
252
+ "grad_norm": 67.1139144897461,
253
+ "learning_rate": 2.623125e-05,
254
+ "loss": 8.0035,
255
+ "step": 1400
256
+ },
257
+ {
258
+ "epoch": 37.89473684210526,
259
+ "grad_norm": 61.43220901489258,
260
+ "learning_rate": 2.6981250000000002e-05,
261
+ "loss": 7.8756,
262
+ "step": 1440
263
+ },
264
+ {
265
+ "epoch": 38.94736842105263,
266
+ "grad_norm": 68.46588134765625,
267
+ "learning_rate": 2.773125e-05,
268
+ "loss": 7.6236,
269
+ "step": 1480
270
+ },
271
+ {
272
+ "epoch": 40.0,
273
+ "grad_norm": 52.86519241333008,
274
+ "learning_rate": 2.848125e-05,
275
+ "loss": 7.3731,
276
+ "step": 1520
277
+ },
278
+ {
279
+ "epoch": 41.05263157894737,
280
+ "grad_norm": 32.2613525390625,
281
+ "learning_rate": 2.923125e-05,
282
+ "loss": 7.5533,
283
+ "step": 1560
284
+ },
285
+ {
286
+ "epoch": 42.10526315789474,
287
+ "grad_norm": 28.07533836364746,
288
+ "learning_rate": 2.998125e-05,
289
+ "loss": 7.2446,
290
+ "step": 1600
291
+ },
292
+ {
293
+ "epoch": 43.1578947368421,
294
+ "grad_norm": 51.62834167480469,
295
+ "learning_rate": 2.91875e-05,
296
+ "loss": 7.2844,
297
+ "step": 1640
298
+ },
299
+ {
300
+ "epoch": 44.21052631578947,
301
+ "grad_norm": 94.51573944091797,
302
+ "learning_rate": 2.8354166666666667e-05,
303
+ "loss": 7.3347,
304
+ "step": 1680
305
+ },
306
+ {
307
+ "epoch": 45.26315789473684,
308
+ "grad_norm": 33.503990173339844,
309
+ "learning_rate": 2.7520833333333333e-05,
310
+ "loss": 7.1209,
311
+ "step": 1720
312
+ },
313
+ {
314
+ "epoch": 46.31578947368421,
315
+ "grad_norm": 74.56465148925781,
316
+ "learning_rate": 2.6687499999999998e-05,
317
+ "loss": 6.7968,
318
+ "step": 1760
319
+ },
320
+ {
321
+ "epoch": 47.36842105263158,
322
+ "grad_norm": 32.059810638427734,
323
+ "learning_rate": 2.5854166666666667e-05,
324
+ "loss": 6.8781,
325
+ "step": 1800
326
+ },
327
+ {
328
+ "epoch": 48.421052631578945,
329
+ "grad_norm": 97.38982391357422,
330
+ "learning_rate": 2.5020833333333336e-05,
331
+ "loss": 6.7608,
332
+ "step": 1840
333
+ },
334
+ {
335
+ "epoch": 49.473684210526315,
336
+ "grad_norm": 34.4022102355957,
337
+ "learning_rate": 2.4187500000000002e-05,
338
+ "loss": 6.4979,
339
+ "step": 1880
340
+ },
341
+ {
342
+ "epoch": 50.526315789473685,
343
+ "grad_norm": 35.79924011230469,
344
+ "learning_rate": 2.3354166666666667e-05,
345
+ "loss": 6.5747,
346
+ "step": 1920
347
+ },
348
+ {
349
+ "epoch": 51.578947368421055,
350
+ "grad_norm": 100.87942504882812,
351
+ "learning_rate": 2.2520833333333333e-05,
352
+ "loss": 6.1757,
353
+ "step": 1960
354
+ },
355
+ {
356
+ "epoch": 52.63157894736842,
357
+ "grad_norm": 159.83860778808594,
358
+ "learning_rate": 2.16875e-05,
359
+ "loss": 6.4399,
360
+ "step": 2000
361
+ },
362
+ {
363
+ "epoch": 53.68421052631579,
364
+ "grad_norm": 58.15259552001953,
365
+ "learning_rate": 2.0854166666666668e-05,
366
+ "loss": 6.3292,
367
+ "step": 2040
368
+ },
369
+ {
370
+ "epoch": 54.73684210526316,
371
+ "grad_norm": 23.384042739868164,
372
+ "learning_rate": 2.0020833333333333e-05,
373
+ "loss": 6.2464,
374
+ "step": 2080
375
+ },
376
+ {
377
+ "epoch": 55.78947368421053,
378
+ "grad_norm": 22.12693977355957,
379
+ "learning_rate": 1.91875e-05,
380
+ "loss": 6.1714,
381
+ "step": 2120
382
+ },
383
+ {
384
+ "epoch": 56.8421052631579,
385
+ "grad_norm": 17.758420944213867,
386
+ "learning_rate": 1.8354166666666668e-05,
387
+ "loss": 5.9527,
388
+ "step": 2160
389
+ },
390
+ {
391
+ "epoch": 57.89473684210526,
392
+ "grad_norm": 39.71998977661133,
393
+ "learning_rate": 1.7520833333333333e-05,
394
+ "loss": 5.9018,
395
+ "step": 2200
396
+ },
397
+ {
398
+ "epoch": 58.94736842105263,
399
+ "grad_norm": 28.227684020996094,
400
+ "learning_rate": 1.6687500000000002e-05,
401
+ "loss": 5.8687,
402
+ "step": 2240
403
+ },
404
+ {
405
+ "epoch": 60.0,
406
+ "grad_norm": 18.77845573425293,
407
+ "learning_rate": 1.5854166666666668e-05,
408
+ "loss": 5.7978,
409
+ "step": 2280
410
+ },
411
+ {
412
+ "epoch": 61.05263157894737,
413
+ "grad_norm": 25.35810089111328,
414
+ "learning_rate": 1.5020833333333334e-05,
415
+ "loss": 5.8052,
416
+ "step": 2320
417
+ },
418
+ {
419
+ "epoch": 62.10526315789474,
420
+ "grad_norm": 92.96112823486328,
421
+ "learning_rate": 1.41875e-05,
422
+ "loss": 5.7386,
423
+ "step": 2360
424
+ },
425
+ {
426
+ "epoch": 63.1578947368421,
427
+ "grad_norm": 35.39582824707031,
428
+ "learning_rate": 1.3354166666666667e-05,
429
+ "loss": 5.6511,
430
+ "step": 2400
431
+ },
432
+ {
433
+ "epoch": 64.21052631578948,
434
+ "grad_norm": 16.94559669494629,
435
+ "learning_rate": 1.2520833333333334e-05,
436
+ "loss": 5.5223,
437
+ "step": 2440
438
+ },
439
+ {
440
+ "epoch": 65.26315789473684,
441
+ "grad_norm": 25.390945434570312,
442
+ "learning_rate": 1.1687500000000001e-05,
443
+ "loss": 5.4784,
444
+ "step": 2480
445
+ },
446
+ {
447
+ "epoch": 66.3157894736842,
448
+ "grad_norm": 81.70599365234375,
449
+ "learning_rate": 1.0854166666666667e-05,
450
+ "loss": 5.4446,
451
+ "step": 2520
452
+ },
453
+ {
454
+ "epoch": 67.36842105263158,
455
+ "grad_norm": 30.602882385253906,
456
+ "learning_rate": 1.0020833333333332e-05,
457
+ "loss": 5.3098,
458
+ "step": 2560
459
+ },
460
+ {
461
+ "epoch": 68.42105263157895,
462
+ "grad_norm": 23.538394927978516,
463
+ "learning_rate": 9.187500000000001e-06,
464
+ "loss": 5.4437,
465
+ "step": 2600
466
+ }
467
+ ],
468
+ "logging_steps": 40,
469
+ "max_steps": 3040,
470
+ "num_input_tokens_seen": 0,
471
+ "num_train_epochs": 80,
472
+ "save_steps": 200,
473
+ "stateful_callbacks": {
474
+ "TrainerControl": {
475
+ "args": {
476
+ "should_epoch_stop": false,
477
+ "should_evaluate": false,
478
+ "should_log": false,
479
+ "should_save": true,
480
+ "should_training_stop": false
481
+ },
482
+ "attributes": {}
483
+ }
484
+ },
485
+ "total_flos": 7.919884351344476e+19,
486
+ "train_batch_size": 32,
487
+ "trial_name": null,
488
+ "trial_params": null
489
+ }
checkpoint-2600/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a669420e271b606473a6402f32de0daffc871134717e7055e89961f40d85f8d
3
+ size 5777
checkpoint-2800/config.json ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "silu",
4
+ "anchor_image_size": null,
5
+ "architectures": [
6
+ "RTDetrV2ForObjectDetection"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "auxiliary_loss": true,
10
+ "backbone": null,
11
+ "backbone_config": {
12
+ "depths": [
13
+ 3,
14
+ 4,
15
+ 23,
16
+ 3
17
+ ],
18
+ "downsample_in_bottleneck": false,
19
+ "downsample_in_first_stage": false,
20
+ "embedding_size": 64,
21
+ "hidden_act": "relu",
22
+ "hidden_sizes": [
23
+ 256,
24
+ 512,
25
+ 1024,
26
+ 2048
27
+ ],
28
+ "layer_type": "bottleneck",
29
+ "model_type": "rt_detr_resnet",
30
+ "num_channels": 3,
31
+ "out_features": [
32
+ "stage2",
33
+ "stage3",
34
+ "stage4"
35
+ ],
36
+ "out_indices": [
37
+ 2,
38
+ 3,
39
+ 4
40
+ ],
41
+ "stage_names": [
42
+ "stem",
43
+ "stage1",
44
+ "stage2",
45
+ "stage3",
46
+ "stage4"
47
+ ],
48
+ "torch_dtype": "float32"
49
+ },
50
+ "backbone_kwargs": null,
51
+ "batch_norm_eps": 1e-05,
52
+ "box_noise_scale": 1.0,
53
+ "d_model": 256,
54
+ "decoder_activation_function": "relu",
55
+ "decoder_attention_heads": 8,
56
+ "decoder_ffn_dim": 1024,
57
+ "decoder_in_channels": [
58
+ 384,
59
+ 384,
60
+ 384
61
+ ],
62
+ "decoder_layers": 6,
63
+ "decoder_method": "default",
64
+ "decoder_n_levels": 3,
65
+ "decoder_n_points": 4,
66
+ "decoder_offset_scale": 0.5,
67
+ "disable_custom_kernels": true,
68
+ "dropout": 0.0,
69
+ "encode_proj_layers": [
70
+ 2
71
+ ],
72
+ "encoder_activation_function": "gelu",
73
+ "encoder_attention_heads": 8,
74
+ "encoder_ffn_dim": 2048,
75
+ "encoder_hidden_dim": 384,
76
+ "encoder_in_channels": [
77
+ 512,
78
+ 1024,
79
+ 2048
80
+ ],
81
+ "encoder_layers": 1,
82
+ "eos_coefficient": 0.0001,
83
+ "eval_size": null,
84
+ "feat_strides": [
85
+ 8,
86
+ 16,
87
+ 32
88
+ ],
89
+ "focal_loss_alpha": 0.75,
90
+ "focal_loss_gamma": 2.0,
91
+ "freeze_backbone_batch_norms": true,
92
+ "hidden_expansion": 1.0,
93
+ "id2label": {
94
+ "0": "digital",
95
+ "1": "fisico",
96
+ "2": "tesoreria"
97
+ },
98
+ "initializer_bias_prior_prob": null,
99
+ "initializer_range": 0.01,
100
+ "is_encoder_decoder": true,
101
+ "label2id": {
102
+ "digital": 0,
103
+ "fisico": 1,
104
+ "tesoreria": 2
105
+ },
106
+ "label_noise_ratio": 0.5,
107
+ "layer_norm_eps": 1e-05,
108
+ "learn_initial_query": false,
109
+ "matcher_alpha": 0.25,
110
+ "matcher_bbox_cost": 5.0,
111
+ "matcher_class_cost": 2.0,
112
+ "matcher_gamma": 2.0,
113
+ "matcher_giou_cost": 2.0,
114
+ "model_type": "rt_detr_v2",
115
+ "normalize_before": false,
116
+ "num_denoising": 100,
117
+ "num_feature_levels": 3,
118
+ "num_queries": 300,
119
+ "positional_encoding_temperature": 10000,
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.55.2",
122
+ "use_focal_loss": true,
123
+ "use_pretrained_backbone": false,
124
+ "use_timm_backbone": false,
125
+ "weight_loss_bbox": 5.0,
126
+ "weight_loss_giou": 2.0,
127
+ "weight_loss_vfl": 1.0,
128
+ "with_box_refine": true
129
+ }
checkpoint-2800/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6926c4ed5177d2ec0df73212edad08233ef3b70768049cfe0c363da3d08f9d2e
3
+ size 306699044
checkpoint-2800/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f11fdd3d5b6fbab28704b1046893ce743d77ae3a54ab5bf4c1bfe8389290e074
3
+ size 611580898
checkpoint-2800/preprocessor_config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": false,
5
+ "device": null,
6
+ "disable_grouping": null,
7
+ "do_center_crop": null,
8
+ "do_convert_annotations": true,
9
+ "do_convert_rgb": null,
10
+ "do_normalize": false,
11
+ "do_pad": false,
12
+ "do_rescale": true,
13
+ "do_resize": true,
14
+ "format": "coco_detection",
15
+ "image_mean": [
16
+ 0.485,
17
+ 0.456,
18
+ 0.406
19
+ ],
20
+ "image_processor_type": "RTDetrImageProcessorFast",
21
+ "image_std": [
22
+ 0.229,
23
+ 0.224,
24
+ 0.225
25
+ ],
26
+ "input_data_format": null,
27
+ "pad_size": null,
28
+ "resample": 2,
29
+ "rescale_factor": 0.00392156862745098,
30
+ "return_segmentation_masks": null,
31
+ "return_tensors": null,
32
+ "size": {
33
+ "height": 832,
34
+ "width": 832
35
+ }
36
+ }
checkpoint-2800/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef3f32481ff4aabb3fb8d7f7d2ea172d87fd9b1e03a8ef201de84fcc38902614
3
+ size 14645
checkpoint-2800/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f1f18dea35c43f2a8dc132aa7863cec2de0113089db053fe03a0467037552f9f
3
+ size 1465
checkpoint-2800/trainer_state.json ADDED
@@ -0,0 +1,524 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 73.6842105263158,
6
+ "eval_steps": 500,
7
+ "global_step": 2800,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 1.0526315789473684,
14
+ "grad_norm": 8679.8642578125,
15
+ "learning_rate": 7.312500000000001e-07,
16
+ "loss": 1384.9168,
17
+ "step": 40
18
+ },
19
+ {
20
+ "epoch": 2.1052631578947367,
21
+ "grad_norm": 4251.486328125,
22
+ "learning_rate": 1.48125e-06,
23
+ "loss": 1290.1081,
24
+ "step": 80
25
+ },
26
+ {
27
+ "epoch": 3.1578947368421053,
28
+ "grad_norm": 4015.17138671875,
29
+ "learning_rate": 2.23125e-06,
30
+ "loss": 1097.6352,
31
+ "step": 120
32
+ },
33
+ {
34
+ "epoch": 4.2105263157894735,
35
+ "grad_norm": 2110.156982421875,
36
+ "learning_rate": 2.98125e-06,
37
+ "loss": 846.727,
38
+ "step": 160
39
+ },
40
+ {
41
+ "epoch": 5.2631578947368425,
42
+ "grad_norm": 1568.4326171875,
43
+ "learning_rate": 3.73125e-06,
44
+ "loss": 578.2698,
45
+ "step": 200
46
+ },
47
+ {
48
+ "epoch": 6.315789473684211,
49
+ "grad_norm": 1924.275146484375,
50
+ "learning_rate": 4.4812500000000005e-06,
51
+ "loss": 412.4179,
52
+ "step": 240
53
+ },
54
+ {
55
+ "epoch": 7.368421052631579,
56
+ "grad_norm": 862.5057373046875,
57
+ "learning_rate": 5.23125e-06,
58
+ "loss": 311.6504,
59
+ "step": 280
60
+ },
61
+ {
62
+ "epoch": 8.421052631578947,
63
+ "grad_norm": 1463.6671142578125,
64
+ "learning_rate": 5.98125e-06,
65
+ "loss": 251.8019,
66
+ "step": 320
67
+ },
68
+ {
69
+ "epoch": 9.473684210526315,
70
+ "grad_norm": 1956.26708984375,
71
+ "learning_rate": 6.73125e-06,
72
+ "loss": 197.8886,
73
+ "step": 360
74
+ },
75
+ {
76
+ "epoch": 10.526315789473685,
77
+ "grad_norm": 404.4883728027344,
78
+ "learning_rate": 7.481250000000001e-06,
79
+ "loss": 155.3046,
80
+ "step": 400
81
+ },
82
+ {
83
+ "epoch": 11.578947368421053,
84
+ "grad_norm": 319.11956787109375,
85
+ "learning_rate": 8.23125e-06,
86
+ "loss": 127.7263,
87
+ "step": 440
88
+ },
89
+ {
90
+ "epoch": 12.631578947368421,
91
+ "grad_norm": 289.92144775390625,
92
+ "learning_rate": 8.98125e-06,
93
+ "loss": 101.4958,
94
+ "step": 480
95
+ },
96
+ {
97
+ "epoch": 13.68421052631579,
98
+ "grad_norm": 238.7841033935547,
99
+ "learning_rate": 9.731250000000001e-06,
100
+ "loss": 80.8442,
101
+ "step": 520
102
+ },
103
+ {
104
+ "epoch": 14.736842105263158,
105
+ "grad_norm": 194.82733154296875,
106
+ "learning_rate": 1.048125e-05,
107
+ "loss": 62.7129,
108
+ "step": 560
109
+ },
110
+ {
111
+ "epoch": 15.789473684210526,
112
+ "grad_norm": 154.37205505371094,
113
+ "learning_rate": 1.123125e-05,
114
+ "loss": 50.3538,
115
+ "step": 600
116
+ },
117
+ {
118
+ "epoch": 16.842105263157894,
119
+ "grad_norm": 208.374267578125,
120
+ "learning_rate": 1.198125e-05,
121
+ "loss": 41.4226,
122
+ "step": 640
123
+ },
124
+ {
125
+ "epoch": 17.894736842105264,
126
+ "grad_norm": 101.1446304321289,
127
+ "learning_rate": 1.2731250000000001e-05,
128
+ "loss": 32.4618,
129
+ "step": 680
130
+ },
131
+ {
132
+ "epoch": 18.94736842105263,
133
+ "grad_norm": 147.55865478515625,
134
+ "learning_rate": 1.348125e-05,
135
+ "loss": 27.0828,
136
+ "step": 720
137
+ },
138
+ {
139
+ "epoch": 20.0,
140
+ "grad_norm": 112.81380462646484,
141
+ "learning_rate": 1.423125e-05,
142
+ "loss": 22.4964,
143
+ "step": 760
144
+ },
145
+ {
146
+ "epoch": 21.05263157894737,
147
+ "grad_norm": 73.30062866210938,
148
+ "learning_rate": 1.4981250000000002e-05,
149
+ "loss": 18.7236,
150
+ "step": 800
151
+ },
152
+ {
153
+ "epoch": 22.105263157894736,
154
+ "grad_norm": 64.0571517944336,
155
+ "learning_rate": 1.5731250000000003e-05,
156
+ "loss": 16.3388,
157
+ "step": 840
158
+ },
159
+ {
160
+ "epoch": 23.157894736842106,
161
+ "grad_norm": 76.62981414794922,
162
+ "learning_rate": 1.6481249999999997e-05,
163
+ "loss": 14.0038,
164
+ "step": 880
165
+ },
166
+ {
167
+ "epoch": 24.210526315789473,
168
+ "grad_norm": 83.39350891113281,
169
+ "learning_rate": 1.723125e-05,
170
+ "loss": 12.6697,
171
+ "step": 920
172
+ },
173
+ {
174
+ "epoch": 25.263157894736842,
175
+ "grad_norm": 49.30060577392578,
176
+ "learning_rate": 1.798125e-05,
177
+ "loss": 11.5521,
178
+ "step": 960
179
+ },
180
+ {
181
+ "epoch": 26.31578947368421,
182
+ "grad_norm": 49.78151321411133,
183
+ "learning_rate": 1.873125e-05,
184
+ "loss": 10.8067,
185
+ "step": 1000
186
+ },
187
+ {
188
+ "epoch": 27.36842105263158,
189
+ "grad_norm": 64.97428894042969,
190
+ "learning_rate": 1.9481250000000003e-05,
191
+ "loss": 9.9502,
192
+ "step": 1040
193
+ },
194
+ {
195
+ "epoch": 28.42105263157895,
196
+ "grad_norm": 96.64849853515625,
197
+ "learning_rate": 2.023125e-05,
198
+ "loss": 9.6148,
199
+ "step": 1080
200
+ },
201
+ {
202
+ "epoch": 29.473684210526315,
203
+ "grad_norm": 65.55039978027344,
204
+ "learning_rate": 2.098125e-05,
205
+ "loss": 9.462,
206
+ "step": 1120
207
+ },
208
+ {
209
+ "epoch": 30.526315789473685,
210
+ "grad_norm": 85.08692169189453,
211
+ "learning_rate": 2.173125e-05,
212
+ "loss": 8.8002,
213
+ "step": 1160
214
+ },
215
+ {
216
+ "epoch": 31.57894736842105,
217
+ "grad_norm": 124.7302017211914,
218
+ "learning_rate": 2.248125e-05,
219
+ "loss": 8.7508,
220
+ "step": 1200
221
+ },
222
+ {
223
+ "epoch": 32.63157894736842,
224
+ "grad_norm": 152.3477020263672,
225
+ "learning_rate": 2.3231250000000002e-05,
226
+ "loss": 8.5803,
227
+ "step": 1240
228
+ },
229
+ {
230
+ "epoch": 33.68421052631579,
231
+ "grad_norm": 162.75997924804688,
232
+ "learning_rate": 2.398125e-05,
233
+ "loss": 8.1324,
234
+ "step": 1280
235
+ },
236
+ {
237
+ "epoch": 34.73684210526316,
238
+ "grad_norm": 78.276611328125,
239
+ "learning_rate": 2.4731249999999998e-05,
240
+ "loss": 8.4976,
241
+ "step": 1320
242
+ },
243
+ {
244
+ "epoch": 35.78947368421053,
245
+ "grad_norm": 151.76458740234375,
246
+ "learning_rate": 2.548125e-05,
247
+ "loss": 8.9999,
248
+ "step": 1360
249
+ },
250
+ {
251
+ "epoch": 36.8421052631579,
252
+ "grad_norm": 67.1139144897461,
253
+ "learning_rate": 2.623125e-05,
254
+ "loss": 8.0035,
255
+ "step": 1400
256
+ },
257
+ {
258
+ "epoch": 37.89473684210526,
259
+ "grad_norm": 61.43220901489258,
260
+ "learning_rate": 2.6981250000000002e-05,
261
+ "loss": 7.8756,
262
+ "step": 1440
263
+ },
264
+ {
265
+ "epoch": 38.94736842105263,
266
+ "grad_norm": 68.46588134765625,
267
+ "learning_rate": 2.773125e-05,
268
+ "loss": 7.6236,
269
+ "step": 1480
270
+ },
271
+ {
272
+ "epoch": 40.0,
273
+ "grad_norm": 52.86519241333008,
274
+ "learning_rate": 2.848125e-05,
275
+ "loss": 7.3731,
276
+ "step": 1520
277
+ },
278
+ {
279
+ "epoch": 41.05263157894737,
280
+ "grad_norm": 32.2613525390625,
281
+ "learning_rate": 2.923125e-05,
282
+ "loss": 7.5533,
283
+ "step": 1560
284
+ },
285
+ {
286
+ "epoch": 42.10526315789474,
287
+ "grad_norm": 28.07533836364746,
288
+ "learning_rate": 2.998125e-05,
289
+ "loss": 7.2446,
290
+ "step": 1600
291
+ },
292
+ {
293
+ "epoch": 43.1578947368421,
294
+ "grad_norm": 51.62834167480469,
295
+ "learning_rate": 2.91875e-05,
296
+ "loss": 7.2844,
297
+ "step": 1640
298
+ },
299
+ {
300
+ "epoch": 44.21052631578947,
301
+ "grad_norm": 94.51573944091797,
302
+ "learning_rate": 2.8354166666666667e-05,
303
+ "loss": 7.3347,
304
+ "step": 1680
305
+ },
306
+ {
307
+ "epoch": 45.26315789473684,
308
+ "grad_norm": 33.503990173339844,
309
+ "learning_rate": 2.7520833333333333e-05,
310
+ "loss": 7.1209,
311
+ "step": 1720
312
+ },
313
+ {
314
+ "epoch": 46.31578947368421,
315
+ "grad_norm": 74.56465148925781,
316
+ "learning_rate": 2.6687499999999998e-05,
317
+ "loss": 6.7968,
318
+ "step": 1760
319
+ },
320
+ {
321
+ "epoch": 47.36842105263158,
322
+ "grad_norm": 32.059810638427734,
323
+ "learning_rate": 2.5854166666666667e-05,
324
+ "loss": 6.8781,
325
+ "step": 1800
326
+ },
327
+ {
328
+ "epoch": 48.421052631578945,
329
+ "grad_norm": 97.38982391357422,
330
+ "learning_rate": 2.5020833333333336e-05,
331
+ "loss": 6.7608,
332
+ "step": 1840
333
+ },
334
+ {
335
+ "epoch": 49.473684210526315,
336
+ "grad_norm": 34.4022102355957,
337
+ "learning_rate": 2.4187500000000002e-05,
338
+ "loss": 6.4979,
339
+ "step": 1880
340
+ },
341
+ {
342
+ "epoch": 50.526315789473685,
343
+ "grad_norm": 35.79924011230469,
344
+ "learning_rate": 2.3354166666666667e-05,
345
+ "loss": 6.5747,
346
+ "step": 1920
347
+ },
348
+ {
349
+ "epoch": 51.578947368421055,
350
+ "grad_norm": 100.87942504882812,
351
+ "learning_rate": 2.2520833333333333e-05,
352
+ "loss": 6.1757,
353
+ "step": 1960
354
+ },
355
+ {
356
+ "epoch": 52.63157894736842,
357
+ "grad_norm": 159.83860778808594,
358
+ "learning_rate": 2.16875e-05,
359
+ "loss": 6.4399,
360
+ "step": 2000
361
+ },
362
+ {
363
+ "epoch": 53.68421052631579,
364
+ "grad_norm": 58.15259552001953,
365
+ "learning_rate": 2.0854166666666668e-05,
366
+ "loss": 6.3292,
367
+ "step": 2040
368
+ },
369
+ {
370
+ "epoch": 54.73684210526316,
371
+ "grad_norm": 23.384042739868164,
372
+ "learning_rate": 2.0020833333333333e-05,
373
+ "loss": 6.2464,
374
+ "step": 2080
375
+ },
376
+ {
377
+ "epoch": 55.78947368421053,
378
+ "grad_norm": 22.12693977355957,
379
+ "learning_rate": 1.91875e-05,
380
+ "loss": 6.1714,
381
+ "step": 2120
382
+ },
383
+ {
384
+ "epoch": 56.8421052631579,
385
+ "grad_norm": 17.758420944213867,
386
+ "learning_rate": 1.8354166666666668e-05,
387
+ "loss": 5.9527,
388
+ "step": 2160
389
+ },
390
+ {
391
+ "epoch": 57.89473684210526,
392
+ "grad_norm": 39.71998977661133,
393
+ "learning_rate": 1.7520833333333333e-05,
394
+ "loss": 5.9018,
395
+ "step": 2200
396
+ },
397
+ {
398
+ "epoch": 58.94736842105263,
399
+ "grad_norm": 28.227684020996094,
400
+ "learning_rate": 1.6687500000000002e-05,
401
+ "loss": 5.8687,
402
+ "step": 2240
403
+ },
404
+ {
405
+ "epoch": 60.0,
406
+ "grad_norm": 18.77845573425293,
407
+ "learning_rate": 1.5854166666666668e-05,
408
+ "loss": 5.7978,
409
+ "step": 2280
410
+ },
411
+ {
412
+ "epoch": 61.05263157894737,
413
+ "grad_norm": 25.35810089111328,
414
+ "learning_rate": 1.5020833333333334e-05,
415
+ "loss": 5.8052,
416
+ "step": 2320
417
+ },
418
+ {
419
+ "epoch": 62.10526315789474,
420
+ "grad_norm": 92.96112823486328,
421
+ "learning_rate": 1.41875e-05,
422
+ "loss": 5.7386,
423
+ "step": 2360
424
+ },
425
+ {
426
+ "epoch": 63.1578947368421,
427
+ "grad_norm": 35.39582824707031,
428
+ "learning_rate": 1.3354166666666667e-05,
429
+ "loss": 5.6511,
430
+ "step": 2400
431
+ },
432
+ {
433
+ "epoch": 64.21052631578948,
434
+ "grad_norm": 16.94559669494629,
435
+ "learning_rate": 1.2520833333333334e-05,
436
+ "loss": 5.5223,
437
+ "step": 2440
438
+ },
439
+ {
440
+ "epoch": 65.26315789473684,
441
+ "grad_norm": 25.390945434570312,
442
+ "learning_rate": 1.1687500000000001e-05,
443
+ "loss": 5.4784,
444
+ "step": 2480
445
+ },
446
+ {
447
+ "epoch": 66.3157894736842,
448
+ "grad_norm": 81.70599365234375,
449
+ "learning_rate": 1.0854166666666667e-05,
450
+ "loss": 5.4446,
451
+ "step": 2520
452
+ },
453
+ {
454
+ "epoch": 67.36842105263158,
455
+ "grad_norm": 30.602882385253906,
456
+ "learning_rate": 1.0020833333333332e-05,
457
+ "loss": 5.3098,
458
+ "step": 2560
459
+ },
460
+ {
461
+ "epoch": 68.42105263157895,
462
+ "grad_norm": 23.538394927978516,
463
+ "learning_rate": 9.187500000000001e-06,
464
+ "loss": 5.4437,
465
+ "step": 2600
466
+ },
467
+ {
468
+ "epoch": 69.47368421052632,
469
+ "grad_norm": 65.4513168334961,
470
+ "learning_rate": 8.354166666666667e-06,
471
+ "loss": 5.3217,
472
+ "step": 2640
473
+ },
474
+ {
475
+ "epoch": 70.52631578947368,
476
+ "grad_norm": 21.81036376953125,
477
+ "learning_rate": 7.5208333333333335e-06,
478
+ "loss": 5.3911,
479
+ "step": 2680
480
+ },
481
+ {
482
+ "epoch": 71.57894736842105,
483
+ "grad_norm": 21.587907791137695,
484
+ "learning_rate": 6.687500000000001e-06,
485
+ "loss": 5.3211,
486
+ "step": 2720
487
+ },
488
+ {
489
+ "epoch": 72.63157894736842,
490
+ "grad_norm": 69.81565856933594,
491
+ "learning_rate": 5.854166666666667e-06,
492
+ "loss": 5.116,
493
+ "step": 2760
494
+ },
495
+ {
496
+ "epoch": 73.6842105263158,
497
+ "grad_norm": 28.424556732177734,
498
+ "learning_rate": 5.020833333333333e-06,
499
+ "loss": 5.1371,
500
+ "step": 2800
501
+ }
502
+ ],
503
+ "logging_steps": 40,
504
+ "max_steps": 3040,
505
+ "num_input_tokens_seen": 0,
506
+ "num_train_epochs": 80,
507
+ "save_steps": 200,
508
+ "stateful_callbacks": {
509
+ "TrainerControl": {
510
+ "args": {
511
+ "should_epoch_stop": false,
512
+ "should_evaluate": false,
513
+ "should_log": false,
514
+ "should_save": true,
515
+ "should_training_stop": false
516
+ },
517
+ "attributes": {}
518
+ }
519
+ },
520
+ "total_flos": 8.52910622452482e+19,
521
+ "train_batch_size": 32,
522
+ "trial_name": null,
523
+ "trial_params": null
524
+ }
checkpoint-2800/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a669420e271b606473a6402f32de0daffc871134717e7055e89961f40d85f8d
3
+ size 5777
checkpoint-3000/config.json ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "silu",
4
+ "anchor_image_size": null,
5
+ "architectures": [
6
+ "RTDetrV2ForObjectDetection"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "auxiliary_loss": true,
10
+ "backbone": null,
11
+ "backbone_config": {
12
+ "depths": [
13
+ 3,
14
+ 4,
15
+ 23,
16
+ 3
17
+ ],
18
+ "downsample_in_bottleneck": false,
19
+ "downsample_in_first_stage": false,
20
+ "embedding_size": 64,
21
+ "hidden_act": "relu",
22
+ "hidden_sizes": [
23
+ 256,
24
+ 512,
25
+ 1024,
26
+ 2048
27
+ ],
28
+ "layer_type": "bottleneck",
29
+ "model_type": "rt_detr_resnet",
30
+ "num_channels": 3,
31
+ "out_features": [
32
+ "stage2",
33
+ "stage3",
34
+ "stage4"
35
+ ],
36
+ "out_indices": [
37
+ 2,
38
+ 3,
39
+ 4
40
+ ],
41
+ "stage_names": [
42
+ "stem",
43
+ "stage1",
44
+ "stage2",
45
+ "stage3",
46
+ "stage4"
47
+ ],
48
+ "torch_dtype": "float32"
49
+ },
50
+ "backbone_kwargs": null,
51
+ "batch_norm_eps": 1e-05,
52
+ "box_noise_scale": 1.0,
53
+ "d_model": 256,
54
+ "decoder_activation_function": "relu",
55
+ "decoder_attention_heads": 8,
56
+ "decoder_ffn_dim": 1024,
57
+ "decoder_in_channels": [
58
+ 384,
59
+ 384,
60
+ 384
61
+ ],
62
+ "decoder_layers": 6,
63
+ "decoder_method": "default",
64
+ "decoder_n_levels": 3,
65
+ "decoder_n_points": 4,
66
+ "decoder_offset_scale": 0.5,
67
+ "disable_custom_kernels": true,
68
+ "dropout": 0.0,
69
+ "encode_proj_layers": [
70
+ 2
71
+ ],
72
+ "encoder_activation_function": "gelu",
73
+ "encoder_attention_heads": 8,
74
+ "encoder_ffn_dim": 2048,
75
+ "encoder_hidden_dim": 384,
76
+ "encoder_in_channels": [
77
+ 512,
78
+ 1024,
79
+ 2048
80
+ ],
81
+ "encoder_layers": 1,
82
+ "eos_coefficient": 0.0001,
83
+ "eval_size": null,
84
+ "feat_strides": [
85
+ 8,
86
+ 16,
87
+ 32
88
+ ],
89
+ "focal_loss_alpha": 0.75,
90
+ "focal_loss_gamma": 2.0,
91
+ "freeze_backbone_batch_norms": true,
92
+ "hidden_expansion": 1.0,
93
+ "id2label": {
94
+ "0": "digital",
95
+ "1": "fisico",
96
+ "2": "tesoreria"
97
+ },
98
+ "initializer_bias_prior_prob": null,
99
+ "initializer_range": 0.01,
100
+ "is_encoder_decoder": true,
101
+ "label2id": {
102
+ "digital": 0,
103
+ "fisico": 1,
104
+ "tesoreria": 2
105
+ },
106
+ "label_noise_ratio": 0.5,
107
+ "layer_norm_eps": 1e-05,
108
+ "learn_initial_query": false,
109
+ "matcher_alpha": 0.25,
110
+ "matcher_bbox_cost": 5.0,
111
+ "matcher_class_cost": 2.0,
112
+ "matcher_gamma": 2.0,
113
+ "matcher_giou_cost": 2.0,
114
+ "model_type": "rt_detr_v2",
115
+ "normalize_before": false,
116
+ "num_denoising": 100,
117
+ "num_feature_levels": 3,
118
+ "num_queries": 300,
119
+ "positional_encoding_temperature": 10000,
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.55.2",
122
+ "use_focal_loss": true,
123
+ "use_pretrained_backbone": false,
124
+ "use_timm_backbone": false,
125
+ "weight_loss_bbox": 5.0,
126
+ "weight_loss_giou": 2.0,
127
+ "weight_loss_vfl": 1.0,
128
+ "with_box_refine": true
129
+ }
checkpoint-3000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ecde63f0dbb9f4f91b8e6c924c30d7eb273d7fad33a54972c3b6055b64831cb
3
+ size 306699044
checkpoint-3000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bfad71752a7902edf3c4632c95980d86357a18a1bad500311be2e206c5fd1f57
3
+ size 611580898
checkpoint-3000/preprocessor_config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": false,
5
+ "device": null,
6
+ "disable_grouping": null,
7
+ "do_center_crop": null,
8
+ "do_convert_annotations": true,
9
+ "do_convert_rgb": null,
10
+ "do_normalize": false,
11
+ "do_pad": false,
12
+ "do_rescale": true,
13
+ "do_resize": true,
14
+ "format": "coco_detection",
15
+ "image_mean": [
16
+ 0.485,
17
+ 0.456,
18
+ 0.406
19
+ ],
20
+ "image_processor_type": "RTDetrImageProcessorFast",
21
+ "image_std": [
22
+ 0.229,
23
+ 0.224,
24
+ 0.225
25
+ ],
26
+ "input_data_format": null,
27
+ "pad_size": null,
28
+ "resample": 2,
29
+ "rescale_factor": 0.00392156862745098,
30
+ "return_segmentation_masks": null,
31
+ "return_tensors": null,
32
+ "size": {
33
+ "height": 832,
34
+ "width": 832
35
+ }
36
+ }
checkpoint-3000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:895128e4bd46ff43369bc30828f5a536fced230bb7d3000c78e6eec9b56bdbd3
3
+ size 14645
checkpoint-3000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2703b6c8999d616aaff7a42777013df45cedc4c4501383c59ada052ae27c98c
3
+ size 1465
checkpoint-3000/trainer_state.json ADDED
@@ -0,0 +1,559 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 78.94736842105263,
6
+ "eval_steps": 500,
7
+ "global_step": 3000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 1.0526315789473684,
14
+ "grad_norm": 8679.8642578125,
15
+ "learning_rate": 7.312500000000001e-07,
16
+ "loss": 1384.9168,
17
+ "step": 40
18
+ },
19
+ {
20
+ "epoch": 2.1052631578947367,
21
+ "grad_norm": 4251.486328125,
22
+ "learning_rate": 1.48125e-06,
23
+ "loss": 1290.1081,
24
+ "step": 80
25
+ },
26
+ {
27
+ "epoch": 3.1578947368421053,
28
+ "grad_norm": 4015.17138671875,
29
+ "learning_rate": 2.23125e-06,
30
+ "loss": 1097.6352,
31
+ "step": 120
32
+ },
33
+ {
34
+ "epoch": 4.2105263157894735,
35
+ "grad_norm": 2110.156982421875,
36
+ "learning_rate": 2.98125e-06,
37
+ "loss": 846.727,
38
+ "step": 160
39
+ },
40
+ {
41
+ "epoch": 5.2631578947368425,
42
+ "grad_norm": 1568.4326171875,
43
+ "learning_rate": 3.73125e-06,
44
+ "loss": 578.2698,
45
+ "step": 200
46
+ },
47
+ {
48
+ "epoch": 6.315789473684211,
49
+ "grad_norm": 1924.275146484375,
50
+ "learning_rate": 4.4812500000000005e-06,
51
+ "loss": 412.4179,
52
+ "step": 240
53
+ },
54
+ {
55
+ "epoch": 7.368421052631579,
56
+ "grad_norm": 862.5057373046875,
57
+ "learning_rate": 5.23125e-06,
58
+ "loss": 311.6504,
59
+ "step": 280
60
+ },
61
+ {
62
+ "epoch": 8.421052631578947,
63
+ "grad_norm": 1463.6671142578125,
64
+ "learning_rate": 5.98125e-06,
65
+ "loss": 251.8019,
66
+ "step": 320
67
+ },
68
+ {
69
+ "epoch": 9.473684210526315,
70
+ "grad_norm": 1956.26708984375,
71
+ "learning_rate": 6.73125e-06,
72
+ "loss": 197.8886,
73
+ "step": 360
74
+ },
75
+ {
76
+ "epoch": 10.526315789473685,
77
+ "grad_norm": 404.4883728027344,
78
+ "learning_rate": 7.481250000000001e-06,
79
+ "loss": 155.3046,
80
+ "step": 400
81
+ },
82
+ {
83
+ "epoch": 11.578947368421053,
84
+ "grad_norm": 319.11956787109375,
85
+ "learning_rate": 8.23125e-06,
86
+ "loss": 127.7263,
87
+ "step": 440
88
+ },
89
+ {
90
+ "epoch": 12.631578947368421,
91
+ "grad_norm": 289.92144775390625,
92
+ "learning_rate": 8.98125e-06,
93
+ "loss": 101.4958,
94
+ "step": 480
95
+ },
96
+ {
97
+ "epoch": 13.68421052631579,
98
+ "grad_norm": 238.7841033935547,
99
+ "learning_rate": 9.731250000000001e-06,
100
+ "loss": 80.8442,
101
+ "step": 520
102
+ },
103
+ {
104
+ "epoch": 14.736842105263158,
105
+ "grad_norm": 194.82733154296875,
106
+ "learning_rate": 1.048125e-05,
107
+ "loss": 62.7129,
108
+ "step": 560
109
+ },
110
+ {
111
+ "epoch": 15.789473684210526,
112
+ "grad_norm": 154.37205505371094,
113
+ "learning_rate": 1.123125e-05,
114
+ "loss": 50.3538,
115
+ "step": 600
116
+ },
117
+ {
118
+ "epoch": 16.842105263157894,
119
+ "grad_norm": 208.374267578125,
120
+ "learning_rate": 1.198125e-05,
121
+ "loss": 41.4226,
122
+ "step": 640
123
+ },
124
+ {
125
+ "epoch": 17.894736842105264,
126
+ "grad_norm": 101.1446304321289,
127
+ "learning_rate": 1.2731250000000001e-05,
128
+ "loss": 32.4618,
129
+ "step": 680
130
+ },
131
+ {
132
+ "epoch": 18.94736842105263,
133
+ "grad_norm": 147.55865478515625,
134
+ "learning_rate": 1.348125e-05,
135
+ "loss": 27.0828,
136
+ "step": 720
137
+ },
138
+ {
139
+ "epoch": 20.0,
140
+ "grad_norm": 112.81380462646484,
141
+ "learning_rate": 1.423125e-05,
142
+ "loss": 22.4964,
143
+ "step": 760
144
+ },
145
+ {
146
+ "epoch": 21.05263157894737,
147
+ "grad_norm": 73.30062866210938,
148
+ "learning_rate": 1.4981250000000002e-05,
149
+ "loss": 18.7236,
150
+ "step": 800
151
+ },
152
+ {
153
+ "epoch": 22.105263157894736,
154
+ "grad_norm": 64.0571517944336,
155
+ "learning_rate": 1.5731250000000003e-05,
156
+ "loss": 16.3388,
157
+ "step": 840
158
+ },
159
+ {
160
+ "epoch": 23.157894736842106,
161
+ "grad_norm": 76.62981414794922,
162
+ "learning_rate": 1.6481249999999997e-05,
163
+ "loss": 14.0038,
164
+ "step": 880
165
+ },
166
+ {
167
+ "epoch": 24.210526315789473,
168
+ "grad_norm": 83.39350891113281,
169
+ "learning_rate": 1.723125e-05,
170
+ "loss": 12.6697,
171
+ "step": 920
172
+ },
173
+ {
174
+ "epoch": 25.263157894736842,
175
+ "grad_norm": 49.30060577392578,
176
+ "learning_rate": 1.798125e-05,
177
+ "loss": 11.5521,
178
+ "step": 960
179
+ },
180
+ {
181
+ "epoch": 26.31578947368421,
182
+ "grad_norm": 49.78151321411133,
183
+ "learning_rate": 1.873125e-05,
184
+ "loss": 10.8067,
185
+ "step": 1000
186
+ },
187
+ {
188
+ "epoch": 27.36842105263158,
189
+ "grad_norm": 64.97428894042969,
190
+ "learning_rate": 1.9481250000000003e-05,
191
+ "loss": 9.9502,
192
+ "step": 1040
193
+ },
194
+ {
195
+ "epoch": 28.42105263157895,
196
+ "grad_norm": 96.64849853515625,
197
+ "learning_rate": 2.023125e-05,
198
+ "loss": 9.6148,
199
+ "step": 1080
200
+ },
201
+ {
202
+ "epoch": 29.473684210526315,
203
+ "grad_norm": 65.55039978027344,
204
+ "learning_rate": 2.098125e-05,
205
+ "loss": 9.462,
206
+ "step": 1120
207
+ },
208
+ {
209
+ "epoch": 30.526315789473685,
210
+ "grad_norm": 85.08692169189453,
211
+ "learning_rate": 2.173125e-05,
212
+ "loss": 8.8002,
213
+ "step": 1160
214
+ },
215
+ {
216
+ "epoch": 31.57894736842105,
217
+ "grad_norm": 124.7302017211914,
218
+ "learning_rate": 2.248125e-05,
219
+ "loss": 8.7508,
220
+ "step": 1200
221
+ },
222
+ {
223
+ "epoch": 32.63157894736842,
224
+ "grad_norm": 152.3477020263672,
225
+ "learning_rate": 2.3231250000000002e-05,
226
+ "loss": 8.5803,
227
+ "step": 1240
228
+ },
229
+ {
230
+ "epoch": 33.68421052631579,
231
+ "grad_norm": 162.75997924804688,
232
+ "learning_rate": 2.398125e-05,
233
+ "loss": 8.1324,
234
+ "step": 1280
235
+ },
236
+ {
237
+ "epoch": 34.73684210526316,
238
+ "grad_norm": 78.276611328125,
239
+ "learning_rate": 2.4731249999999998e-05,
240
+ "loss": 8.4976,
241
+ "step": 1320
242
+ },
243
+ {
244
+ "epoch": 35.78947368421053,
245
+ "grad_norm": 151.76458740234375,
246
+ "learning_rate": 2.548125e-05,
247
+ "loss": 8.9999,
248
+ "step": 1360
249
+ },
250
+ {
251
+ "epoch": 36.8421052631579,
252
+ "grad_norm": 67.1139144897461,
253
+ "learning_rate": 2.623125e-05,
254
+ "loss": 8.0035,
255
+ "step": 1400
256
+ },
257
+ {
258
+ "epoch": 37.89473684210526,
259
+ "grad_norm": 61.43220901489258,
260
+ "learning_rate": 2.6981250000000002e-05,
261
+ "loss": 7.8756,
262
+ "step": 1440
263
+ },
264
+ {
265
+ "epoch": 38.94736842105263,
266
+ "grad_norm": 68.46588134765625,
267
+ "learning_rate": 2.773125e-05,
268
+ "loss": 7.6236,
269
+ "step": 1480
270
+ },
271
+ {
272
+ "epoch": 40.0,
273
+ "grad_norm": 52.86519241333008,
274
+ "learning_rate": 2.848125e-05,
275
+ "loss": 7.3731,
276
+ "step": 1520
277
+ },
278
+ {
279
+ "epoch": 41.05263157894737,
280
+ "grad_norm": 32.2613525390625,
281
+ "learning_rate": 2.923125e-05,
282
+ "loss": 7.5533,
283
+ "step": 1560
284
+ },
285
+ {
286
+ "epoch": 42.10526315789474,
287
+ "grad_norm": 28.07533836364746,
288
+ "learning_rate": 2.998125e-05,
289
+ "loss": 7.2446,
290
+ "step": 1600
291
+ },
292
+ {
293
+ "epoch": 43.1578947368421,
294
+ "grad_norm": 51.62834167480469,
295
+ "learning_rate": 2.91875e-05,
296
+ "loss": 7.2844,
297
+ "step": 1640
298
+ },
299
+ {
300
+ "epoch": 44.21052631578947,
301
+ "grad_norm": 94.51573944091797,
302
+ "learning_rate": 2.8354166666666667e-05,
303
+ "loss": 7.3347,
304
+ "step": 1680
305
+ },
306
+ {
307
+ "epoch": 45.26315789473684,
308
+ "grad_norm": 33.503990173339844,
309
+ "learning_rate": 2.7520833333333333e-05,
310
+ "loss": 7.1209,
311
+ "step": 1720
312
+ },
313
+ {
314
+ "epoch": 46.31578947368421,
315
+ "grad_norm": 74.56465148925781,
316
+ "learning_rate": 2.6687499999999998e-05,
317
+ "loss": 6.7968,
318
+ "step": 1760
319
+ },
320
+ {
321
+ "epoch": 47.36842105263158,
322
+ "grad_norm": 32.059810638427734,
323
+ "learning_rate": 2.5854166666666667e-05,
324
+ "loss": 6.8781,
325
+ "step": 1800
326
+ },
327
+ {
328
+ "epoch": 48.421052631578945,
329
+ "grad_norm": 97.38982391357422,
330
+ "learning_rate": 2.5020833333333336e-05,
331
+ "loss": 6.7608,
332
+ "step": 1840
333
+ },
334
+ {
335
+ "epoch": 49.473684210526315,
336
+ "grad_norm": 34.4022102355957,
337
+ "learning_rate": 2.4187500000000002e-05,
338
+ "loss": 6.4979,
339
+ "step": 1880
340
+ },
341
+ {
342
+ "epoch": 50.526315789473685,
343
+ "grad_norm": 35.79924011230469,
344
+ "learning_rate": 2.3354166666666667e-05,
345
+ "loss": 6.5747,
346
+ "step": 1920
347
+ },
348
+ {
349
+ "epoch": 51.578947368421055,
350
+ "grad_norm": 100.87942504882812,
351
+ "learning_rate": 2.2520833333333333e-05,
352
+ "loss": 6.1757,
353
+ "step": 1960
354
+ },
355
+ {
356
+ "epoch": 52.63157894736842,
357
+ "grad_norm": 159.83860778808594,
358
+ "learning_rate": 2.16875e-05,
359
+ "loss": 6.4399,
360
+ "step": 2000
361
+ },
362
+ {
363
+ "epoch": 53.68421052631579,
364
+ "grad_norm": 58.15259552001953,
365
+ "learning_rate": 2.0854166666666668e-05,
366
+ "loss": 6.3292,
367
+ "step": 2040
368
+ },
369
+ {
370
+ "epoch": 54.73684210526316,
371
+ "grad_norm": 23.384042739868164,
372
+ "learning_rate": 2.0020833333333333e-05,
373
+ "loss": 6.2464,
374
+ "step": 2080
375
+ },
376
+ {
377
+ "epoch": 55.78947368421053,
378
+ "grad_norm": 22.12693977355957,
379
+ "learning_rate": 1.91875e-05,
380
+ "loss": 6.1714,
381
+ "step": 2120
382
+ },
383
+ {
384
+ "epoch": 56.8421052631579,
385
+ "grad_norm": 17.758420944213867,
386
+ "learning_rate": 1.8354166666666668e-05,
387
+ "loss": 5.9527,
388
+ "step": 2160
389
+ },
390
+ {
391
+ "epoch": 57.89473684210526,
392
+ "grad_norm": 39.71998977661133,
393
+ "learning_rate": 1.7520833333333333e-05,
394
+ "loss": 5.9018,
395
+ "step": 2200
396
+ },
397
+ {
398
+ "epoch": 58.94736842105263,
399
+ "grad_norm": 28.227684020996094,
400
+ "learning_rate": 1.6687500000000002e-05,
401
+ "loss": 5.8687,
402
+ "step": 2240
403
+ },
404
+ {
405
+ "epoch": 60.0,
406
+ "grad_norm": 18.77845573425293,
407
+ "learning_rate": 1.5854166666666668e-05,
408
+ "loss": 5.7978,
409
+ "step": 2280
410
+ },
411
+ {
412
+ "epoch": 61.05263157894737,
413
+ "grad_norm": 25.35810089111328,
414
+ "learning_rate": 1.5020833333333334e-05,
415
+ "loss": 5.8052,
416
+ "step": 2320
417
+ },
418
+ {
419
+ "epoch": 62.10526315789474,
420
+ "grad_norm": 92.96112823486328,
421
+ "learning_rate": 1.41875e-05,
422
+ "loss": 5.7386,
423
+ "step": 2360
424
+ },
425
+ {
426
+ "epoch": 63.1578947368421,
427
+ "grad_norm": 35.39582824707031,
428
+ "learning_rate": 1.3354166666666667e-05,
429
+ "loss": 5.6511,
430
+ "step": 2400
431
+ },
432
+ {
433
+ "epoch": 64.21052631578948,
434
+ "grad_norm": 16.94559669494629,
435
+ "learning_rate": 1.2520833333333334e-05,
436
+ "loss": 5.5223,
437
+ "step": 2440
438
+ },
439
+ {
440
+ "epoch": 65.26315789473684,
441
+ "grad_norm": 25.390945434570312,
442
+ "learning_rate": 1.1687500000000001e-05,
443
+ "loss": 5.4784,
444
+ "step": 2480
445
+ },
446
+ {
447
+ "epoch": 66.3157894736842,
448
+ "grad_norm": 81.70599365234375,
449
+ "learning_rate": 1.0854166666666667e-05,
450
+ "loss": 5.4446,
451
+ "step": 2520
452
+ },
453
+ {
454
+ "epoch": 67.36842105263158,
455
+ "grad_norm": 30.602882385253906,
456
+ "learning_rate": 1.0020833333333332e-05,
457
+ "loss": 5.3098,
458
+ "step": 2560
459
+ },
460
+ {
461
+ "epoch": 68.42105263157895,
462
+ "grad_norm": 23.538394927978516,
463
+ "learning_rate": 9.187500000000001e-06,
464
+ "loss": 5.4437,
465
+ "step": 2600
466
+ },
467
+ {
468
+ "epoch": 69.47368421052632,
469
+ "grad_norm": 65.4513168334961,
470
+ "learning_rate": 8.354166666666667e-06,
471
+ "loss": 5.3217,
472
+ "step": 2640
473
+ },
474
+ {
475
+ "epoch": 70.52631578947368,
476
+ "grad_norm": 21.81036376953125,
477
+ "learning_rate": 7.5208333333333335e-06,
478
+ "loss": 5.3911,
479
+ "step": 2680
480
+ },
481
+ {
482
+ "epoch": 71.57894736842105,
483
+ "grad_norm": 21.587907791137695,
484
+ "learning_rate": 6.687500000000001e-06,
485
+ "loss": 5.3211,
486
+ "step": 2720
487
+ },
488
+ {
489
+ "epoch": 72.63157894736842,
490
+ "grad_norm": 69.81565856933594,
491
+ "learning_rate": 5.854166666666667e-06,
492
+ "loss": 5.116,
493
+ "step": 2760
494
+ },
495
+ {
496
+ "epoch": 73.6842105263158,
497
+ "grad_norm": 28.424556732177734,
498
+ "learning_rate": 5.020833333333333e-06,
499
+ "loss": 5.1371,
500
+ "step": 2800
501
+ },
502
+ {
503
+ "epoch": 74.73684210526316,
504
+ "grad_norm": 14.789178848266602,
505
+ "learning_rate": 4.1875e-06,
506
+ "loss": 5.1651,
507
+ "step": 2840
508
+ },
509
+ {
510
+ "epoch": 75.78947368421052,
511
+ "grad_norm": 17.214509963989258,
512
+ "learning_rate": 3.354166666666667e-06,
513
+ "loss": 5.0846,
514
+ "step": 2880
515
+ },
516
+ {
517
+ "epoch": 76.84210526315789,
518
+ "grad_norm": 144.82855224609375,
519
+ "learning_rate": 2.5208333333333335e-06,
520
+ "loss": 5.0452,
521
+ "step": 2920
522
+ },
523
+ {
524
+ "epoch": 77.89473684210526,
525
+ "grad_norm": 55.73275375366211,
526
+ "learning_rate": 1.6875000000000001e-06,
527
+ "loss": 4.9982,
528
+ "step": 2960
529
+ },
530
+ {
531
+ "epoch": 78.94736842105263,
532
+ "grad_norm": 22.088788986206055,
533
+ "learning_rate": 8.541666666666667e-07,
534
+ "loss": 4.9426,
535
+ "step": 3000
536
+ }
537
+ ],
538
+ "logging_steps": 40,
539
+ "max_steps": 3040,
540
+ "num_input_tokens_seen": 0,
541
+ "num_train_epochs": 80,
542
+ "save_steps": 200,
543
+ "stateful_callbacks": {
544
+ "TrainerControl": {
545
+ "args": {
546
+ "should_epoch_stop": false,
547
+ "should_evaluate": false,
548
+ "should_log": false,
549
+ "should_save": true,
550
+ "should_training_stop": false
551
+ },
552
+ "attributes": {}
553
+ }
554
+ },
555
+ "total_flos": 9.138328097705165e+19,
556
+ "train_batch_size": 32,
557
+ "trial_name": null,
558
+ "trial_params": null
559
+ }
checkpoint-3000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a669420e271b606473a6402f32de0daffc871134717e7055e89961f40d85f8d
3
+ size 5777
checkpoint-3040/config.json ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "silu",
4
+ "anchor_image_size": null,
5
+ "architectures": [
6
+ "RTDetrV2ForObjectDetection"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "auxiliary_loss": true,
10
+ "backbone": null,
11
+ "backbone_config": {
12
+ "depths": [
13
+ 3,
14
+ 4,
15
+ 23,
16
+ 3
17
+ ],
18
+ "downsample_in_bottleneck": false,
19
+ "downsample_in_first_stage": false,
20
+ "embedding_size": 64,
21
+ "hidden_act": "relu",
22
+ "hidden_sizes": [
23
+ 256,
24
+ 512,
25
+ 1024,
26
+ 2048
27
+ ],
28
+ "layer_type": "bottleneck",
29
+ "model_type": "rt_detr_resnet",
30
+ "num_channels": 3,
31
+ "out_features": [
32
+ "stage2",
33
+ "stage3",
34
+ "stage4"
35
+ ],
36
+ "out_indices": [
37
+ 2,
38
+ 3,
39
+ 4
40
+ ],
41
+ "stage_names": [
42
+ "stem",
43
+ "stage1",
44
+ "stage2",
45
+ "stage3",
46
+ "stage4"
47
+ ],
48
+ "torch_dtype": "float32"
49
+ },
50
+ "backbone_kwargs": null,
51
+ "batch_norm_eps": 1e-05,
52
+ "box_noise_scale": 1.0,
53
+ "d_model": 256,
54
+ "decoder_activation_function": "relu",
55
+ "decoder_attention_heads": 8,
56
+ "decoder_ffn_dim": 1024,
57
+ "decoder_in_channels": [
58
+ 384,
59
+ 384,
60
+ 384
61
+ ],
62
+ "decoder_layers": 6,
63
+ "decoder_method": "default",
64
+ "decoder_n_levels": 3,
65
+ "decoder_n_points": 4,
66
+ "decoder_offset_scale": 0.5,
67
+ "disable_custom_kernels": true,
68
+ "dropout": 0.0,
69
+ "encode_proj_layers": [
70
+ 2
71
+ ],
72
+ "encoder_activation_function": "gelu",
73
+ "encoder_attention_heads": 8,
74
+ "encoder_ffn_dim": 2048,
75
+ "encoder_hidden_dim": 384,
76
+ "encoder_in_channels": [
77
+ 512,
78
+ 1024,
79
+ 2048
80
+ ],
81
+ "encoder_layers": 1,
82
+ "eos_coefficient": 0.0001,
83
+ "eval_size": null,
84
+ "feat_strides": [
85
+ 8,
86
+ 16,
87
+ 32
88
+ ],
89
+ "focal_loss_alpha": 0.75,
90
+ "focal_loss_gamma": 2.0,
91
+ "freeze_backbone_batch_norms": true,
92
+ "hidden_expansion": 1.0,
93
+ "id2label": {
94
+ "0": "digital",
95
+ "1": "fisico",
96
+ "2": "tesoreria"
97
+ },
98
+ "initializer_bias_prior_prob": null,
99
+ "initializer_range": 0.01,
100
+ "is_encoder_decoder": true,
101
+ "label2id": {
102
+ "digital": 0,
103
+ "fisico": 1,
104
+ "tesoreria": 2
105
+ },
106
+ "label_noise_ratio": 0.5,
107
+ "layer_norm_eps": 1e-05,
108
+ "learn_initial_query": false,
109
+ "matcher_alpha": 0.25,
110
+ "matcher_bbox_cost": 5.0,
111
+ "matcher_class_cost": 2.0,
112
+ "matcher_gamma": 2.0,
113
+ "matcher_giou_cost": 2.0,
114
+ "model_type": "rt_detr_v2",
115
+ "normalize_before": false,
116
+ "num_denoising": 100,
117
+ "num_feature_levels": 3,
118
+ "num_queries": 300,
119
+ "positional_encoding_temperature": 10000,
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.55.2",
122
+ "use_focal_loss": true,
123
+ "use_pretrained_backbone": false,
124
+ "use_timm_backbone": false,
125
+ "weight_loss_bbox": 5.0,
126
+ "weight_loss_giou": 2.0,
127
+ "weight_loss_vfl": 1.0,
128
+ "with_box_refine": true
129
+ }
checkpoint-3040/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf4ffe63a6ce70e4187387c31c31d516665a6e309010d02c47c7feea75a693f5
3
+ size 306699044
checkpoint-3040/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7ff0a7fab1246ca8857b1427941999591af9780aab7660119b6e0b5f20fb279
3
+ size 611580898
checkpoint-3040/preprocessor_config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": false,
5
+ "device": null,
6
+ "disable_grouping": null,
7
+ "do_center_crop": null,
8
+ "do_convert_annotations": true,
9
+ "do_convert_rgb": null,
10
+ "do_normalize": false,
11
+ "do_pad": false,
12
+ "do_rescale": true,
13
+ "do_resize": true,
14
+ "format": "coco_detection",
15
+ "image_mean": [
16
+ 0.485,
17
+ 0.456,
18
+ 0.406
19
+ ],
20
+ "image_processor_type": "RTDetrImageProcessorFast",
21
+ "image_std": [
22
+ 0.229,
23
+ 0.224,
24
+ 0.225
25
+ ],
26
+ "input_data_format": null,
27
+ "pad_size": null,
28
+ "resample": 2,
29
+ "rescale_factor": 0.00392156862745098,
30
+ "return_segmentation_masks": null,
31
+ "return_tensors": null,
32
+ "size": {
33
+ "height": 832,
34
+ "width": 832
35
+ }
36
+ }
checkpoint-3040/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0750a3ba6372189fe6b060790cbb611abb9c0851f3e0b9a27cde89badd76dcaf
3
+ size 14645
checkpoint-3040/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a72d093c93f54dbba7db39183b657cfa4789e875470824184738dff4447c7ce4
3
+ size 1465
checkpoint-3040/trainer_state.json ADDED
@@ -0,0 +1,566 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 80.0,
6
+ "eval_steps": 500,
7
+ "global_step": 3040,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 1.0526315789473684,
14
+ "grad_norm": 8679.8642578125,
15
+ "learning_rate": 7.312500000000001e-07,
16
+ "loss": 1384.9168,
17
+ "step": 40
18
+ },
19
+ {
20
+ "epoch": 2.1052631578947367,
21
+ "grad_norm": 4251.486328125,
22
+ "learning_rate": 1.48125e-06,
23
+ "loss": 1290.1081,
24
+ "step": 80
25
+ },
26
+ {
27
+ "epoch": 3.1578947368421053,
28
+ "grad_norm": 4015.17138671875,
29
+ "learning_rate": 2.23125e-06,
30
+ "loss": 1097.6352,
31
+ "step": 120
32
+ },
33
+ {
34
+ "epoch": 4.2105263157894735,
35
+ "grad_norm": 2110.156982421875,
36
+ "learning_rate": 2.98125e-06,
37
+ "loss": 846.727,
38
+ "step": 160
39
+ },
40
+ {
41
+ "epoch": 5.2631578947368425,
42
+ "grad_norm": 1568.4326171875,
43
+ "learning_rate": 3.73125e-06,
44
+ "loss": 578.2698,
45
+ "step": 200
46
+ },
47
+ {
48
+ "epoch": 6.315789473684211,
49
+ "grad_norm": 1924.275146484375,
50
+ "learning_rate": 4.4812500000000005e-06,
51
+ "loss": 412.4179,
52
+ "step": 240
53
+ },
54
+ {
55
+ "epoch": 7.368421052631579,
56
+ "grad_norm": 862.5057373046875,
57
+ "learning_rate": 5.23125e-06,
58
+ "loss": 311.6504,
59
+ "step": 280
60
+ },
61
+ {
62
+ "epoch": 8.421052631578947,
63
+ "grad_norm": 1463.6671142578125,
64
+ "learning_rate": 5.98125e-06,
65
+ "loss": 251.8019,
66
+ "step": 320
67
+ },
68
+ {
69
+ "epoch": 9.473684210526315,
70
+ "grad_norm": 1956.26708984375,
71
+ "learning_rate": 6.73125e-06,
72
+ "loss": 197.8886,
73
+ "step": 360
74
+ },
75
+ {
76
+ "epoch": 10.526315789473685,
77
+ "grad_norm": 404.4883728027344,
78
+ "learning_rate": 7.481250000000001e-06,
79
+ "loss": 155.3046,
80
+ "step": 400
81
+ },
82
+ {
83
+ "epoch": 11.578947368421053,
84
+ "grad_norm": 319.11956787109375,
85
+ "learning_rate": 8.23125e-06,
86
+ "loss": 127.7263,
87
+ "step": 440
88
+ },
89
+ {
90
+ "epoch": 12.631578947368421,
91
+ "grad_norm": 289.92144775390625,
92
+ "learning_rate": 8.98125e-06,
93
+ "loss": 101.4958,
94
+ "step": 480
95
+ },
96
+ {
97
+ "epoch": 13.68421052631579,
98
+ "grad_norm": 238.7841033935547,
99
+ "learning_rate": 9.731250000000001e-06,
100
+ "loss": 80.8442,
101
+ "step": 520
102
+ },
103
+ {
104
+ "epoch": 14.736842105263158,
105
+ "grad_norm": 194.82733154296875,
106
+ "learning_rate": 1.048125e-05,
107
+ "loss": 62.7129,
108
+ "step": 560
109
+ },
110
+ {
111
+ "epoch": 15.789473684210526,
112
+ "grad_norm": 154.37205505371094,
113
+ "learning_rate": 1.123125e-05,
114
+ "loss": 50.3538,
115
+ "step": 600
116
+ },
117
+ {
118
+ "epoch": 16.842105263157894,
119
+ "grad_norm": 208.374267578125,
120
+ "learning_rate": 1.198125e-05,
121
+ "loss": 41.4226,
122
+ "step": 640
123
+ },
124
+ {
125
+ "epoch": 17.894736842105264,
126
+ "grad_norm": 101.1446304321289,
127
+ "learning_rate": 1.2731250000000001e-05,
128
+ "loss": 32.4618,
129
+ "step": 680
130
+ },
131
+ {
132
+ "epoch": 18.94736842105263,
133
+ "grad_norm": 147.55865478515625,
134
+ "learning_rate": 1.348125e-05,
135
+ "loss": 27.0828,
136
+ "step": 720
137
+ },
138
+ {
139
+ "epoch": 20.0,
140
+ "grad_norm": 112.81380462646484,
141
+ "learning_rate": 1.423125e-05,
142
+ "loss": 22.4964,
143
+ "step": 760
144
+ },
145
+ {
146
+ "epoch": 21.05263157894737,
147
+ "grad_norm": 73.30062866210938,
148
+ "learning_rate": 1.4981250000000002e-05,
149
+ "loss": 18.7236,
150
+ "step": 800
151
+ },
152
+ {
153
+ "epoch": 22.105263157894736,
154
+ "grad_norm": 64.0571517944336,
155
+ "learning_rate": 1.5731250000000003e-05,
156
+ "loss": 16.3388,
157
+ "step": 840
158
+ },
159
+ {
160
+ "epoch": 23.157894736842106,
161
+ "grad_norm": 76.62981414794922,
162
+ "learning_rate": 1.6481249999999997e-05,
163
+ "loss": 14.0038,
164
+ "step": 880
165
+ },
166
+ {
167
+ "epoch": 24.210526315789473,
168
+ "grad_norm": 83.39350891113281,
169
+ "learning_rate": 1.723125e-05,
170
+ "loss": 12.6697,
171
+ "step": 920
172
+ },
173
+ {
174
+ "epoch": 25.263157894736842,
175
+ "grad_norm": 49.30060577392578,
176
+ "learning_rate": 1.798125e-05,
177
+ "loss": 11.5521,
178
+ "step": 960
179
+ },
180
+ {
181
+ "epoch": 26.31578947368421,
182
+ "grad_norm": 49.78151321411133,
183
+ "learning_rate": 1.873125e-05,
184
+ "loss": 10.8067,
185
+ "step": 1000
186
+ },
187
+ {
188
+ "epoch": 27.36842105263158,
189
+ "grad_norm": 64.97428894042969,
190
+ "learning_rate": 1.9481250000000003e-05,
191
+ "loss": 9.9502,
192
+ "step": 1040
193
+ },
194
+ {
195
+ "epoch": 28.42105263157895,
196
+ "grad_norm": 96.64849853515625,
197
+ "learning_rate": 2.023125e-05,
198
+ "loss": 9.6148,
199
+ "step": 1080
200
+ },
201
+ {
202
+ "epoch": 29.473684210526315,
203
+ "grad_norm": 65.55039978027344,
204
+ "learning_rate": 2.098125e-05,
205
+ "loss": 9.462,
206
+ "step": 1120
207
+ },
208
+ {
209
+ "epoch": 30.526315789473685,
210
+ "grad_norm": 85.08692169189453,
211
+ "learning_rate": 2.173125e-05,
212
+ "loss": 8.8002,
213
+ "step": 1160
214
+ },
215
+ {
216
+ "epoch": 31.57894736842105,
217
+ "grad_norm": 124.7302017211914,
218
+ "learning_rate": 2.248125e-05,
219
+ "loss": 8.7508,
220
+ "step": 1200
221
+ },
222
+ {
223
+ "epoch": 32.63157894736842,
224
+ "grad_norm": 152.3477020263672,
225
+ "learning_rate": 2.3231250000000002e-05,
226
+ "loss": 8.5803,
227
+ "step": 1240
228
+ },
229
+ {
230
+ "epoch": 33.68421052631579,
231
+ "grad_norm": 162.75997924804688,
232
+ "learning_rate": 2.398125e-05,
233
+ "loss": 8.1324,
234
+ "step": 1280
235
+ },
236
+ {
237
+ "epoch": 34.73684210526316,
238
+ "grad_norm": 78.276611328125,
239
+ "learning_rate": 2.4731249999999998e-05,
240
+ "loss": 8.4976,
241
+ "step": 1320
242
+ },
243
+ {
244
+ "epoch": 35.78947368421053,
245
+ "grad_norm": 151.76458740234375,
246
+ "learning_rate": 2.548125e-05,
247
+ "loss": 8.9999,
248
+ "step": 1360
249
+ },
250
+ {
251
+ "epoch": 36.8421052631579,
252
+ "grad_norm": 67.1139144897461,
253
+ "learning_rate": 2.623125e-05,
254
+ "loss": 8.0035,
255
+ "step": 1400
256
+ },
257
+ {
258
+ "epoch": 37.89473684210526,
259
+ "grad_norm": 61.43220901489258,
260
+ "learning_rate": 2.6981250000000002e-05,
261
+ "loss": 7.8756,
262
+ "step": 1440
263
+ },
264
+ {
265
+ "epoch": 38.94736842105263,
266
+ "grad_norm": 68.46588134765625,
267
+ "learning_rate": 2.773125e-05,
268
+ "loss": 7.6236,
269
+ "step": 1480
270
+ },
271
+ {
272
+ "epoch": 40.0,
273
+ "grad_norm": 52.86519241333008,
274
+ "learning_rate": 2.848125e-05,
275
+ "loss": 7.3731,
276
+ "step": 1520
277
+ },
278
+ {
279
+ "epoch": 41.05263157894737,
280
+ "grad_norm": 32.2613525390625,
281
+ "learning_rate": 2.923125e-05,
282
+ "loss": 7.5533,
283
+ "step": 1560
284
+ },
285
+ {
286
+ "epoch": 42.10526315789474,
287
+ "grad_norm": 28.07533836364746,
288
+ "learning_rate": 2.998125e-05,
289
+ "loss": 7.2446,
290
+ "step": 1600
291
+ },
292
+ {
293
+ "epoch": 43.1578947368421,
294
+ "grad_norm": 51.62834167480469,
295
+ "learning_rate": 2.91875e-05,
296
+ "loss": 7.2844,
297
+ "step": 1640
298
+ },
299
+ {
300
+ "epoch": 44.21052631578947,
301
+ "grad_norm": 94.51573944091797,
302
+ "learning_rate": 2.8354166666666667e-05,
303
+ "loss": 7.3347,
304
+ "step": 1680
305
+ },
306
+ {
307
+ "epoch": 45.26315789473684,
308
+ "grad_norm": 33.503990173339844,
309
+ "learning_rate": 2.7520833333333333e-05,
310
+ "loss": 7.1209,
311
+ "step": 1720
312
+ },
313
+ {
314
+ "epoch": 46.31578947368421,
315
+ "grad_norm": 74.56465148925781,
316
+ "learning_rate": 2.6687499999999998e-05,
317
+ "loss": 6.7968,
318
+ "step": 1760
319
+ },
320
+ {
321
+ "epoch": 47.36842105263158,
322
+ "grad_norm": 32.059810638427734,
323
+ "learning_rate": 2.5854166666666667e-05,
324
+ "loss": 6.8781,
325
+ "step": 1800
326
+ },
327
+ {
328
+ "epoch": 48.421052631578945,
329
+ "grad_norm": 97.38982391357422,
330
+ "learning_rate": 2.5020833333333336e-05,
331
+ "loss": 6.7608,
332
+ "step": 1840
333
+ },
334
+ {
335
+ "epoch": 49.473684210526315,
336
+ "grad_norm": 34.4022102355957,
337
+ "learning_rate": 2.4187500000000002e-05,
338
+ "loss": 6.4979,
339
+ "step": 1880
340
+ },
341
+ {
342
+ "epoch": 50.526315789473685,
343
+ "grad_norm": 35.79924011230469,
344
+ "learning_rate": 2.3354166666666667e-05,
345
+ "loss": 6.5747,
346
+ "step": 1920
347
+ },
348
+ {
349
+ "epoch": 51.578947368421055,
350
+ "grad_norm": 100.87942504882812,
351
+ "learning_rate": 2.2520833333333333e-05,
352
+ "loss": 6.1757,
353
+ "step": 1960
354
+ },
355
+ {
356
+ "epoch": 52.63157894736842,
357
+ "grad_norm": 159.83860778808594,
358
+ "learning_rate": 2.16875e-05,
359
+ "loss": 6.4399,
360
+ "step": 2000
361
+ },
362
+ {
363
+ "epoch": 53.68421052631579,
364
+ "grad_norm": 58.15259552001953,
365
+ "learning_rate": 2.0854166666666668e-05,
366
+ "loss": 6.3292,
367
+ "step": 2040
368
+ },
369
+ {
370
+ "epoch": 54.73684210526316,
371
+ "grad_norm": 23.384042739868164,
372
+ "learning_rate": 2.0020833333333333e-05,
373
+ "loss": 6.2464,
374
+ "step": 2080
375
+ },
376
+ {
377
+ "epoch": 55.78947368421053,
378
+ "grad_norm": 22.12693977355957,
379
+ "learning_rate": 1.91875e-05,
380
+ "loss": 6.1714,
381
+ "step": 2120
382
+ },
383
+ {
384
+ "epoch": 56.8421052631579,
385
+ "grad_norm": 17.758420944213867,
386
+ "learning_rate": 1.8354166666666668e-05,
387
+ "loss": 5.9527,
388
+ "step": 2160
389
+ },
390
+ {
391
+ "epoch": 57.89473684210526,
392
+ "grad_norm": 39.71998977661133,
393
+ "learning_rate": 1.7520833333333333e-05,
394
+ "loss": 5.9018,
395
+ "step": 2200
396
+ },
397
+ {
398
+ "epoch": 58.94736842105263,
399
+ "grad_norm": 28.227684020996094,
400
+ "learning_rate": 1.6687500000000002e-05,
401
+ "loss": 5.8687,
402
+ "step": 2240
403
+ },
404
+ {
405
+ "epoch": 60.0,
406
+ "grad_norm": 18.77845573425293,
407
+ "learning_rate": 1.5854166666666668e-05,
408
+ "loss": 5.7978,
409
+ "step": 2280
410
+ },
411
+ {
412
+ "epoch": 61.05263157894737,
413
+ "grad_norm": 25.35810089111328,
414
+ "learning_rate": 1.5020833333333334e-05,
415
+ "loss": 5.8052,
416
+ "step": 2320
417
+ },
418
+ {
419
+ "epoch": 62.10526315789474,
420
+ "grad_norm": 92.96112823486328,
421
+ "learning_rate": 1.41875e-05,
422
+ "loss": 5.7386,
423
+ "step": 2360
424
+ },
425
+ {
426
+ "epoch": 63.1578947368421,
427
+ "grad_norm": 35.39582824707031,
428
+ "learning_rate": 1.3354166666666667e-05,
429
+ "loss": 5.6511,
430
+ "step": 2400
431
+ },
432
+ {
433
+ "epoch": 64.21052631578948,
434
+ "grad_norm": 16.94559669494629,
435
+ "learning_rate": 1.2520833333333334e-05,
436
+ "loss": 5.5223,
437
+ "step": 2440
438
+ },
439
+ {
440
+ "epoch": 65.26315789473684,
441
+ "grad_norm": 25.390945434570312,
442
+ "learning_rate": 1.1687500000000001e-05,
443
+ "loss": 5.4784,
444
+ "step": 2480
445
+ },
446
+ {
447
+ "epoch": 66.3157894736842,
448
+ "grad_norm": 81.70599365234375,
449
+ "learning_rate": 1.0854166666666667e-05,
450
+ "loss": 5.4446,
451
+ "step": 2520
452
+ },
453
+ {
454
+ "epoch": 67.36842105263158,
455
+ "grad_norm": 30.602882385253906,
456
+ "learning_rate": 1.0020833333333332e-05,
457
+ "loss": 5.3098,
458
+ "step": 2560
459
+ },
460
+ {
461
+ "epoch": 68.42105263157895,
462
+ "grad_norm": 23.538394927978516,
463
+ "learning_rate": 9.187500000000001e-06,
464
+ "loss": 5.4437,
465
+ "step": 2600
466
+ },
467
+ {
468
+ "epoch": 69.47368421052632,
469
+ "grad_norm": 65.4513168334961,
470
+ "learning_rate": 8.354166666666667e-06,
471
+ "loss": 5.3217,
472
+ "step": 2640
473
+ },
474
+ {
475
+ "epoch": 70.52631578947368,
476
+ "grad_norm": 21.81036376953125,
477
+ "learning_rate": 7.5208333333333335e-06,
478
+ "loss": 5.3911,
479
+ "step": 2680
480
+ },
481
+ {
482
+ "epoch": 71.57894736842105,
483
+ "grad_norm": 21.587907791137695,
484
+ "learning_rate": 6.687500000000001e-06,
485
+ "loss": 5.3211,
486
+ "step": 2720
487
+ },
488
+ {
489
+ "epoch": 72.63157894736842,
490
+ "grad_norm": 69.81565856933594,
491
+ "learning_rate": 5.854166666666667e-06,
492
+ "loss": 5.116,
493
+ "step": 2760
494
+ },
495
+ {
496
+ "epoch": 73.6842105263158,
497
+ "grad_norm": 28.424556732177734,
498
+ "learning_rate": 5.020833333333333e-06,
499
+ "loss": 5.1371,
500
+ "step": 2800
501
+ },
502
+ {
503
+ "epoch": 74.73684210526316,
504
+ "grad_norm": 14.789178848266602,
505
+ "learning_rate": 4.1875e-06,
506
+ "loss": 5.1651,
507
+ "step": 2840
508
+ },
509
+ {
510
+ "epoch": 75.78947368421052,
511
+ "grad_norm": 17.214509963989258,
512
+ "learning_rate": 3.354166666666667e-06,
513
+ "loss": 5.0846,
514
+ "step": 2880
515
+ },
516
+ {
517
+ "epoch": 76.84210526315789,
518
+ "grad_norm": 144.82855224609375,
519
+ "learning_rate": 2.5208333333333335e-06,
520
+ "loss": 5.0452,
521
+ "step": 2920
522
+ },
523
+ {
524
+ "epoch": 77.89473684210526,
525
+ "grad_norm": 55.73275375366211,
526
+ "learning_rate": 1.6875000000000001e-06,
527
+ "loss": 4.9982,
528
+ "step": 2960
529
+ },
530
+ {
531
+ "epoch": 78.94736842105263,
532
+ "grad_norm": 22.088788986206055,
533
+ "learning_rate": 8.541666666666667e-07,
534
+ "loss": 4.9426,
535
+ "step": 3000
536
+ },
537
+ {
538
+ "epoch": 80.0,
539
+ "grad_norm": 12.161073684692383,
540
+ "learning_rate": 2.0833333333333335e-08,
541
+ "loss": 4.9881,
542
+ "step": 3040
543
+ }
544
+ ],
545
+ "logging_steps": 40,
546
+ "max_steps": 3040,
547
+ "num_input_tokens_seen": 0,
548
+ "num_train_epochs": 80,
549
+ "save_steps": 200,
550
+ "stateful_callbacks": {
551
+ "TrainerControl": {
552
+ "args": {
553
+ "should_epoch_stop": false,
554
+ "should_evaluate": false,
555
+ "should_log": false,
556
+ "should_save": true,
557
+ "should_training_stop": true
558
+ },
559
+ "attributes": {}
560
+ }
561
+ },
562
+ "total_flos": 9.260172472341234e+19,
563
+ "train_batch_size": 32,
564
+ "trial_name": null,
565
+ "trial_params": null
566
+ }
checkpoint-3040/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a669420e271b606473a6402f32de0daffc871134717e7055e89961f40d85f8d
3
+ size 5777
config.json CHANGED
@@ -12,7 +12,7 @@
12
  "depths": [
13
  3,
14
  4,
15
- 6,
16
  3
17
  ],
18
  "downsample_in_bottleneck": false,
@@ -55,9 +55,9 @@
55
  "decoder_attention_heads": 8,
56
  "decoder_ffn_dim": 1024,
57
  "decoder_in_channels": [
58
- 256,
59
- 256,
60
- 256
61
  ],
62
  "decoder_layers": 6,
63
  "decoder_method": "default",
@@ -71,8 +71,8 @@
71
  ],
72
  "encoder_activation_function": "gelu",
73
  "encoder_attention_heads": 8,
74
- "encoder_ffn_dim": 1024,
75
- "encoder_hidden_dim": 256,
76
  "encoder_in_channels": [
77
  512,
78
  1024,
 
12
  "depths": [
13
  3,
14
  4,
15
+ 23,
16
  3
17
  ],
18
  "downsample_in_bottleneck": false,
 
55
  "decoder_attention_heads": 8,
56
  "decoder_ffn_dim": 1024,
57
  "decoder_in_channels": [
58
+ 384,
59
+ 384,
60
+ 384
61
  ],
62
  "decoder_layers": 6,
63
  "decoder_method": "default",
 
71
  ],
72
  "encoder_activation_function": "gelu",
73
  "encoder_attention_heads": 8,
74
+ "encoder_ffn_dim": 2048,
75
+ "encoder_hidden_dim": 384,
76
  "encoder_in_channels": [
77
  512,
78
  1024,
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6bbb6c0a0b18bc32d3fd25ca3b8133060b964e2b0a3fccdb4f480c6f60ca76bd
3
- size 171543900
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf4ffe63a6ce70e4187387c31c31d516665a6e309010d02c47c7feea75a693f5
3
+ size 306699044
preprocessor_config.json CHANGED
@@ -30,7 +30,7 @@
30
  "return_segmentation_masks": null,
31
  "return_tensors": null,
32
  "size": {
33
- "height": 665,
34
- "width": 665
35
  }
36
  }
 
30
  "return_segmentation_masks": null,
31
  "return_tensors": null,
32
  "size": {
33
+ "height": 832,
34
+ "width": 832
35
  }
36
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:da3a438d7d8497331b97a4ad4e3f87cc1475357011c441ae8213c62ec78bcc4d
3
  size 5777
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a669420e271b606473a6402f32de0daffc871134717e7055e89961f40d85f8d
3
  size 5777