Hugging Face's logo

jnmrr
/
ds3-img-classification

Image Classification
Transformers
Safetensors
beit
document-classification
vision
Model card Files
xet
Community
jnmrr commited on
Commit
d94ca4a
·
verified ·
1 Parent(s): d3a7216

Upload model with metrics - 2024-12-12 15:17

Browse files
Files changed (23) hide show
  1. README.md +21 -21
  2. checkpoint-1562/config.json +169 -0
  3. checkpoint-1562/model.safetensors +3 -0
  4. checkpoint-1562/preprocessor_config.json +28 -0
  5. checkpoint-1562/rng_state.pth +3 -0
  6. checkpoint-1562/trainer_state.json +324 -0
  7. checkpoint-1562/training_args.bin +3 -0
  8. checkpoint-1704/config.json +169 -0
  9. checkpoint-1704/model.safetensors +3 -0
  10. checkpoint-1704/preprocessor_config.json +28 -0
  11. checkpoint-1704/rng_state.pth +3 -0
  12. checkpoint-1704/trainer_state.json +349 -0
  13. checkpoint-1704/training_args.bin +3 -0
  14. checkpoint-284/config.json +169 -0
  15. checkpoint-284/model.safetensors +3 -0
  16. checkpoint-284/preprocessor_config.json +28 -0
  17. checkpoint-284/rng_state.pth +3 -0
  18. checkpoint-284/trainer_state.json +99 -0
  19. checkpoint-284/training_args.bin +3 -0
  20. metrics.json +19 -19
  21. model.safetensors +1 -1
  22. trainer_state.json +329 -45
  23. training_args.bin +1 -1
README.md CHANGED
@@ -22,8 +22,8 @@ This model is trained for document classification using vision transformers (DiT
22
 
23
  ## Training Parameters
24
  * Batch Size: 256
25
- * Learning Rate: 0.002
26
- * Number of Epochs: 1
27
  * Mixed Precision: BF16
28
  * Gradient Accumulation Steps: 2
29
  * Weight Decay: 0.01
@@ -32,27 +32,27 @@ This model is trained for document classification using vision transformers (DiT
32
 
33
  ## Training and Evaluation Metrics
34
  ### Training Metrics
35
- * Loss: 5.2299
36
- * Grad Norm: 5.5809
37
- * Learning Rate: 0.0010
38
- * Epoch: 0.5426
39
- * Step: 35.0000
40
 
41
  ### Evaluation Metrics
42
- * Loss: 5.0066
43
- * Accuracy: 0.1351
44
- * Weighted F1: 0.0662
45
- * Micro F1: 0.1351
46
- * Macro F1: 0.0380
47
- * Weighted Recall: 0.1351
48
- * Micro Recall: 0.1351
49
- * Macro Recall: 0.0646
50
- * Weighted Precision: 0.0548
51
- * Micro Precision: 0.1351
52
- * Macro Precision: 0.0333
53
- * Runtime: 8.2574
54
- * Samples Per Second: 441.1820
55
- * Steps Per Second: 3.5120
56
 
57
  ## Usage
58
 
 
22
 
23
  ## Training Parameters
24
  * Batch Size: 256
25
+ * Learning Rate: 0.001
26
+ * Number of Epochs: 90
27
  * Mixed Precision: BF16
28
  * Gradient Accumulation Steps: 2
29
  * Weight Decay: 0.01
 
32
 
33
  ## Training and Evaluation Metrics
34
  ### Training Metrics
35
+ * Loss: 0.1915
36
+ * Grad Norm: 1.3002
37
+ * Learning Rate: 0.0009
38
+ * Epoch: 26.4186
39
+ * Step: 1704.0000
40
 
41
  ### Evaluation Metrics
42
+ * Loss: 0.9457
43
+ * Accuracy: 0.7757
44
+ * Weighted F1: 0.7689
45
+ * Micro F1: 0.7757
46
+ * Macro F1: 0.7518
47
+ * Weighted Recall: 0.7757
48
+ * Micro Recall: 0.7757
49
+ * Macro Recall: 0.7603
50
+ * Weighted Precision: 0.8023
51
+ * Micro Precision: 0.7757
52
+ * Macro Precision: 0.7941
53
+ * Runtime: 8.4106
54
+ * Samples Per Second: 433.1450
55
+ * Steps Per Second: 3.4480
56
 
57
  ## Usage
58
 
checkpoint-1562/config.json ADDED
@@ -0,0 +1,169 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/dit-large",
3
+ "add_fpn": false,
4
+ "architectures": [
5
+ "BeitForImageClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.0,
8
+ "auxiliary_channels": 256,
9
+ "auxiliary_concat_input": false,
10
+ "auxiliary_loss_weight": 0.4,
11
+ "auxiliary_num_convs": 1,
12
+ "drop_path_rate": 0.1,
13
+ "hidden_act": "gelu",
14
+ "hidden_dropout_prob": 0.0,
15
+ "hidden_size": 1024,
16
+ "id2label": {
17
+ "0": "informe_social",
18
+ "1": "cargo_ingreso_mpe",
19
+ "2": "acta_registro_personal",
20
+ "3": "citacion_pnp",
21
+ "4": "acta_lectura_derechos_ley30364",
22
+ "5": "acta_intervencion_minpu",
23
+ "6": "denuncia_policial",
24
+ "7": "providencia_fiscal",
25
+ "8": "consulta_medidas_proteccion",
26
+ "9": "oficio_medicina_legal",
27
+ "10": "oficio_evaluacion_psicologica",
28
+ "11": "resolucion_judicial_audiencia",
29
+ "12": "certificado_medico_legal",
30
+ "13": "ficha_datos_sidpol",
31
+ "14": "consulta_sucamec",
32
+ "15": "informe_mimp",
33
+ "16": "ficha_valoracion_riesgo_manual",
34
+ "17": "oficio_atencion_integral",
35
+ "18": "escrito_minpu",
36
+ "19": "informe_psicologico",
37
+ "20": "ficha_valoracion_riesgo_digital",
38
+ "21": "informe_medico",
39
+ "22": "croquis_domicilio",
40
+ "23": "ficha_datos_reniec",
41
+ "24": "escrito_pj",
42
+ "25": "evidencia_chat",
43
+ "26": "declaracion_minpu",
44
+ "27": "notificacion_resolucion_judicial_audiencia",
45
+ "28": "consulta_antecedentes_penales",
46
+ "29": "notificacion_detencion_pnp",
47
+ "30": "consulta_personas_sidpol",
48
+ "31": "escrito_pnp",
49
+ "32": "constancia_buen_trato",
50
+ "33": "declaracion_pnp",
51
+ "34": "escrito_mimp",
52
+ "35": "acta_intervencion_pnp",
53
+ "36": "constancia_notificacion",
54
+ "37": "cargo_ingreso_sij",
55
+ "38": "anexo_dni",
56
+ "39": "hoja_blanco",
57
+ "40": "consulta_requisitorias"
58
+ },
59
+ "image_size": 224,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "label2id": {
63
+ "acta_intervencion_minpu": 5,
64
+ "acta_intervencion_pnp": 35,
65
+ "acta_lectura_derechos_ley30364": 4,
66
+ "acta_registro_personal": 2,
67
+ "anexo_dni": 38,
68
+ "cargo_ingreso_mpe": 1,
69
+ "cargo_ingreso_sij": 37,
70
+ "certificado_medico_legal": 12,
71
+ "citacion_pnp": 3,
72
+ "constancia_buen_trato": 32,
73
+ "constancia_notificacion": 36,
74
+ "consulta_antecedentes_penales": 28,
75
+ "consulta_medidas_proteccion": 8,
76
+ "consulta_personas_sidpol": 30,
77
+ "consulta_requisitorias": 40,
78
+ "consulta_sucamec": 14,
79
+ "croquis_domicilio": 22,
80
+ "declaracion_minpu": 26,
81
+ "declaracion_pnp": 33,
82
+ "denuncia_policial": 6,
83
+ "escrito_mimp": 34,
84
+ "escrito_minpu": 18,
85
+ "escrito_pj": 24,
86
+ "escrito_pnp": 31,
87
+ "evidencia_chat": 25,
88
+ "ficha_datos_reniec": 23,
89
+ "ficha_datos_sidpol": 13,
90
+ "ficha_valoracion_riesgo_digital": 20,
91
+ "ficha_valoracion_riesgo_manual": 16,
92
+ "hoja_blanco": 39,
93
+ "informe_medico": 21,
94
+ "informe_mimp": 15,
95
+ "informe_psicologico": 19,
96
+ "informe_social": 0,
97
+ "notificacion_detencion_pnp": 29,
98
+ "notificacion_resolucion_judicial_audiencia": 27,
99
+ "oficio_atencion_integral": 17,
100
+ "oficio_evaluacion_psicologica": 10,
101
+ "oficio_medicina_legal": 9,
102
+ "providencia_fiscal": 7,
103
+ "resolucion_judicial_audiencia": 11
104
+ },
105
+ "layer_norm_eps": 1e-12,
106
+ "layer_scale_init_value": 0.1,
107
+ "model_type": "beit",
108
+ "num_attention_heads": 16,
109
+ "num_channels": 3,
110
+ "num_hidden_layers": 24,
111
+ "out_features": [
112
+ "stage3",
113
+ "stage5",
114
+ "stage7",
115
+ "stage11"
116
+ ],
117
+ "out_indices": [
118
+ 3,
119
+ 5,
120
+ 7,
121
+ 11
122
+ ],
123
+ "patch_size": 16,
124
+ "pool_scales": [
125
+ 1,
126
+ 2,
127
+ 3,
128
+ 6
129
+ ],
130
+ "problem_type": "single_label_classification",
131
+ "reshape_hidden_states": true,
132
+ "semantic_loss_ignore_index": 255,
133
+ "stage_names": [
134
+ "stem",
135
+ "stage1",
136
+ "stage2",
137
+ "stage3",
138
+ "stage4",
139
+ "stage5",
140
+ "stage6",
141
+ "stage7",
142
+ "stage8",
143
+ "stage9",
144
+ "stage10",
145
+ "stage11",
146
+ "stage12",
147
+ "stage13",
148
+ "stage14",
149
+ "stage15",
150
+ "stage16",
151
+ "stage17",
152
+ "stage18",
153
+ "stage19",
154
+ "stage20",
155
+ "stage21",
156
+ "stage22",
157
+ "stage23",
158
+ "stage24"
159
+ ],
160
+ "torch_dtype": "float32",
161
+ "transformers_version": "4.46.3",
162
+ "use_absolute_position_embeddings": true,
163
+ "use_auxiliary_head": true,
164
+ "use_mask_token": true,
165
+ "use_mean_pooling": true,
166
+ "use_relative_position_bias": false,
167
+ "use_shared_relative_position_bias": false,
168
+ "vocab_size": 8192
169
+ }
checkpoint-1562/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd0469d06ae797f51abf7a6b882a9e3cde9574cf8b5726b9c776c625dc4123a2
3
+ size 1213526036
checkpoint-1562/preprocessor_config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": {
3
+ "height": 224,
4
+ "width": 224
5
+ },
6
+ "do_center_crop": false,
7
+ "do_normalize": true,
8
+ "do_reduce_labels": false,
9
+ "do_rescale": true,
10
+ "do_resize": true,
11
+ "image_mean": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "image_processor_type": "BeitImageProcessor",
17
+ "image_std": [
18
+ 0.5,
19
+ 0.5,
20
+ 0.5
21
+ ],
22
+ "resample": 2,
23
+ "rescale_factor": 0.00392156862745098,
24
+ "size": {
25
+ "height": 224,
26
+ "width": 224
27
+ }
28
+ }
checkpoint-1562/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a685a91f3eecbc53340df781a56f0950dc253e3a8455515dcb9ae876605b352a
3
+ size 14244
checkpoint-1562/trainer_state.json ADDED
@@ -0,0 +1,324 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8528685149601977,
3
+ "best_model_checkpoint": "ds3-img-classification/checkpoint-284",
4
+ "epoch": 24.217054263565892,
5
+ "eval_steps": 142,
6
+ "global_step": 1562,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.015503875968992248,
13
+ "grad_norm": 14.292332649230957,
14
+ "learning_rate": 1.736111111111111e-06,
15
+ "loss": 5.8979,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 2.201550387596899,
20
+ "grad_norm": 3.2709274291992188,
21
+ "learning_rate": 0.0002465277777777778,
22
+ "loss": 2.2063,
23
+ "step": 142
24
+ },
25
+ {
26
+ "epoch": 2.201550387596899,
27
+ "eval_accuracy": 0.8300850947021685,
28
+ "eval_loss": 0.6956288814544678,
29
+ "eval_macro_f1": 0.8020115300522307,
30
+ "eval_macro_precision": 0.8326277180719005,
31
+ "eval_macro_recall": 0.8130180643302005,
32
+ "eval_micro_f1": 0.8300850947021685,
33
+ "eval_micro_precision": 0.8300850947021685,
34
+ "eval_micro_recall": 0.8300850947021685,
35
+ "eval_runtime": 8.1254,
36
+ "eval_samples_per_second": 448.347,
37
+ "eval_steps_per_second": 3.569,
38
+ "eval_weighted_f1": 0.8217784359221563,
39
+ "eval_weighted_precision": 0.8439879212180366,
40
+ "eval_weighted_recall": 0.8300850947021685,
41
+ "step": 142
42
+ },
43
+ {
44
+ "epoch": 4.403100775193798,
45
+ "grad_norm": 3.7781729698181152,
46
+ "learning_rate": 0.0004930555555555556,
47
+ "loss": 0.4906,
48
+ "step": 284
49
+ },
50
+ {
51
+ "epoch": 4.403100775193798,
52
+ "eval_accuracy": 0.8528685149601977,
53
+ "eval_loss": 0.5704612731933594,
54
+ "eval_macro_f1": 0.8447646928874578,
55
+ "eval_macro_precision": 0.8434580548053667,
56
+ "eval_macro_recall": 0.8590562294756982,
57
+ "eval_micro_f1": 0.8528685149601977,
58
+ "eval_micro_precision": 0.8528685149601977,
59
+ "eval_micro_recall": 0.8528685149601977,
60
+ "eval_runtime": 8.3026,
61
+ "eval_samples_per_second": 438.779,
62
+ "eval_steps_per_second": 3.493,
63
+ "eval_weighted_f1": 0.8542422373852568,
64
+ "eval_weighted_precision": 0.8661150819486806,
65
+ "eval_weighted_recall": 0.8528685149601977,
66
+ "step": 284
67
+ },
68
+ {
69
+ "epoch": 6.604651162790698,
70
+ "grad_norm": 2.7355144023895264,
71
+ "learning_rate": 0.0007395833333333334,
72
+ "loss": 0.4125,
73
+ "step": 426
74
+ },
75
+ {
76
+ "epoch": 6.604651162790698,
77
+ "eval_accuracy": 0.7400494098270656,
78
+ "eval_loss": 1.1121388673782349,
79
+ "eval_macro_f1": 0.7104399369512154,
80
+ "eval_macro_precision": 0.7984066441843661,
81
+ "eval_macro_recall": 0.7246907362358344,
82
+ "eval_micro_f1": 0.7400494098270657,
83
+ "eval_micro_precision": 0.7400494098270656,
84
+ "eval_micro_recall": 0.7400494098270656,
85
+ "eval_runtime": 8.3145,
86
+ "eval_samples_per_second": 438.152,
87
+ "eval_steps_per_second": 3.488,
88
+ "eval_weighted_f1": 0.7343329270391191,
89
+ "eval_weighted_precision": 0.8109139038669717,
90
+ "eval_weighted_recall": 0.7400494098270656,
91
+ "step": 426
92
+ },
93
+ {
94
+ "epoch": 8.806201550387597,
95
+ "grad_norm": 2.7910587787628174,
96
+ "learning_rate": 0.0009861111111111112,
97
+ "loss": 0.3948,
98
+ "step": 568
99
+ },
100
+ {
101
+ "epoch": 8.806201550387597,
102
+ "eval_accuracy": 0.769146307987922,
103
+ "eval_loss": 0.9053287506103516,
104
+ "eval_macro_f1": 0.7380470144030622,
105
+ "eval_macro_precision": 0.7929019396341611,
106
+ "eval_macro_recall": 0.7416184656713186,
107
+ "eval_micro_f1": 0.769146307987922,
108
+ "eval_micro_precision": 0.769146307987922,
109
+ "eval_micro_recall": 0.769146307987922,
110
+ "eval_runtime": 8.4314,
111
+ "eval_samples_per_second": 432.076,
112
+ "eval_steps_per_second": 3.44,
113
+ "eval_weighted_f1": 0.7647284160437106,
114
+ "eval_weighted_precision": 0.804728448308173,
115
+ "eval_weighted_recall": 0.769146307987922,
116
+ "step": 568
117
+ },
118
+ {
119
+ "epoch": 11.007751937984496,
120
+ "grad_norm": 6.627900123596191,
121
+ "learning_rate": 0.0009983522903302263,
122
+ "loss": 0.3706,
123
+ "step": 710
124
+ },
125
+ {
126
+ "epoch": 11.007751937984496,
127
+ "eval_accuracy": 0.7672248147131485,
128
+ "eval_loss": 0.9641202688217163,
129
+ "eval_macro_f1": 0.7528371715147637,
130
+ "eval_macro_precision": 0.787293122179915,
131
+ "eval_macro_recall": 0.7589348598973276,
132
+ "eval_micro_f1": 0.7672248147131485,
133
+ "eval_micro_precision": 0.7672248147131485,
134
+ "eval_micro_recall": 0.7672248147131485,
135
+ "eval_runtime": 8.3804,
136
+ "eval_samples_per_second": 434.705,
137
+ "eval_steps_per_second": 3.46,
138
+ "eval_weighted_f1": 0.7623056450515923,
139
+ "eval_weighted_precision": 0.7946085579995245,
140
+ "eval_weighted_recall": 0.7672248147131485,
141
+ "step": 710
142
+ },
143
+ {
144
+ "epoch": 13.209302325581396,
145
+ "grad_norm": 3.9615182876586914,
146
+ "learning_rate": 0.0009930222532081596,
147
+ "loss": 0.3302,
148
+ "step": 852
149
+ },
150
+ {
151
+ "epoch": 13.209302325581396,
152
+ "eval_accuracy": 0.7669503156738952,
153
+ "eval_loss": 0.9978815317153931,
154
+ "eval_macro_f1": 0.731377351970227,
155
+ "eval_macro_precision": 0.7752862718850362,
156
+ "eval_macro_recall": 0.7408605973033737,
157
+ "eval_micro_f1": 0.766950315673895,
158
+ "eval_micro_precision": 0.7669503156738952,
159
+ "eval_micro_recall": 0.7669503156738952,
160
+ "eval_runtime": 8.3024,
161
+ "eval_samples_per_second": 438.791,
162
+ "eval_steps_per_second": 3.493,
163
+ "eval_weighted_f1": 0.7591614272562255,
164
+ "eval_weighted_precision": 0.7900342900911238,
165
+ "eval_weighted_recall": 0.7669503156738952,
166
+ "step": 852
167
+ },
168
+ {
169
+ "epoch": 15.410852713178295,
170
+ "grad_norm": 2.4717087745666504,
171
+ "learning_rate": 0.0009840434590541415,
172
+ "loss": 0.2924,
173
+ "step": 994
174
+ },
175
+ {
176
+ "epoch": 15.410852713178295,
177
+ "eval_accuracy": 0.7233049684326105,
178
+ "eval_loss": 1.1638933420181274,
179
+ "eval_macro_f1": 0.7113593382146706,
180
+ "eval_macro_precision": 0.759489526605669,
181
+ "eval_macro_recall": 0.7371465474721075,
182
+ "eval_micro_f1": 0.7233049684326105,
183
+ "eval_micro_precision": 0.7233049684326105,
184
+ "eval_micro_recall": 0.7233049684326105,
185
+ "eval_runtime": 8.2739,
186
+ "eval_samples_per_second": 440.3,
187
+ "eval_steps_per_second": 3.505,
188
+ "eval_weighted_f1": 0.7249947526290721,
189
+ "eval_weighted_precision": 0.7928301153657854,
190
+ "eval_weighted_recall": 0.7233049684326105,
191
+ "step": 994
192
+ },
193
+ {
194
+ "epoch": 17.612403100775193,
195
+ "grad_norm": 1.886505126953125,
196
+ "learning_rate": 0.0009714823580904379,
197
+ "loss": 0.271,
198
+ "step": 1136
199
+ },
200
+ {
201
+ "epoch": 17.612403100775193,
202
+ "eval_accuracy": 0.7631073291243481,
203
+ "eval_loss": 0.9723167419433594,
204
+ "eval_macro_f1": 0.7630758622320426,
205
+ "eval_macro_precision": 0.8404983162678952,
206
+ "eval_macro_recall": 0.7684491516518803,
207
+ "eval_micro_f1": 0.7631073291243481,
208
+ "eval_micro_precision": 0.7631073291243481,
209
+ "eval_micro_recall": 0.7631073291243481,
210
+ "eval_runtime": 8.2536,
211
+ "eval_samples_per_second": 441.381,
212
+ "eval_steps_per_second": 3.514,
213
+ "eval_weighted_f1": 0.7641468306327522,
214
+ "eval_weighted_precision": 0.8384853206385829,
215
+ "eval_weighted_recall": 0.7631073291243481,
216
+ "step": 1136
217
+ },
218
+ {
219
+ "epoch": 19.813953488372093,
220
+ "grad_norm": 1.751386284828186,
221
+ "learning_rate": 0.000955431912460588,
222
+ "loss": 0.2493,
223
+ "step": 1278
224
+ },
225
+ {
226
+ "epoch": 19.813953488372093,
227
+ "eval_accuracy": 0.7556958550645073,
228
+ "eval_loss": 0.9924145340919495,
229
+ "eval_macro_f1": 0.7186954972930405,
230
+ "eval_macro_precision": 0.7780354194554568,
231
+ "eval_macro_recall": 0.7246815062488177,
232
+ "eval_micro_f1": 0.7556958550645072,
233
+ "eval_micro_precision": 0.7556958550645073,
234
+ "eval_micro_recall": 0.7556958550645073,
235
+ "eval_runtime": 8.2332,
236
+ "eval_samples_per_second": 442.478,
237
+ "eval_steps_per_second": 3.522,
238
+ "eval_weighted_f1": 0.7424053023538434,
239
+ "eval_weighted_precision": 0.7815999680084916,
240
+ "eval_weighted_recall": 0.7556958550645073,
241
+ "step": 1278
242
+ },
243
+ {
244
+ "epoch": 22.015503875968992,
245
+ "grad_norm": 1.887330174446106,
246
+ "learning_rate": 0.0009360109082355581,
247
+ "loss": 0.2276,
248
+ "step": 1420
249
+ },
250
+ {
251
+ "epoch": 22.015503875968992,
252
+ "eval_accuracy": 0.7488333790831732,
253
+ "eval_loss": 1.0906082391738892,
254
+ "eval_macro_f1": 0.7232412414342934,
255
+ "eval_macro_precision": 0.7888326922300277,
256
+ "eval_macro_recall": 0.7295020066680354,
257
+ "eval_micro_f1": 0.7488333790831732,
258
+ "eval_micro_precision": 0.7488333790831732,
259
+ "eval_micro_recall": 0.7488333790831732,
260
+ "eval_runtime": 8.3573,
261
+ "eval_samples_per_second": 435.905,
262
+ "eval_steps_per_second": 3.47,
263
+ "eval_weighted_f1": 0.7411123038164863,
264
+ "eval_weighted_precision": 0.78924044171482,
265
+ "eval_weighted_recall": 0.7488333790831732,
266
+ "step": 1420
267
+ },
268
+ {
269
+ "epoch": 24.217054263565892,
270
+ "grad_norm": 1.4166276454925537,
271
+ "learning_rate": 0.0009133630763022861,
272
+ "loss": 0.2147,
273
+ "step": 1562
274
+ },
275
+ {
276
+ "epoch": 24.217054263565892,
277
+ "eval_accuracy": 0.7595388416140544,
278
+ "eval_loss": 1.046807050704956,
279
+ "eval_macro_f1": 0.7218058945395823,
280
+ "eval_macro_precision": 0.7806200306355751,
281
+ "eval_macro_recall": 0.7363816501236903,
282
+ "eval_micro_f1": 0.7595388416140544,
283
+ "eval_micro_precision": 0.7595388416140544,
284
+ "eval_micro_recall": 0.7595388416140544,
285
+ "eval_runtime": 8.3705,
286
+ "eval_samples_per_second": 435.218,
287
+ "eval_steps_per_second": 3.465,
288
+ "eval_weighted_f1": 0.7500356793933665,
289
+ "eval_weighted_precision": 0.7963937500158152,
290
+ "eval_weighted_recall": 0.7595388416140544,
291
+ "step": 1562
292
+ }
293
+ ],
294
+ "logging_steps": 142,
295
+ "max_steps": 5760,
296
+ "num_input_tokens_seen": 0,
297
+ "num_train_epochs": 90,
298
+ "save_steps": 142,
299
+ "stateful_callbacks": {
300
+ "EarlyStoppingCallback": {
301
+ "args": {
302
+ "early_stopping_patience": 10,
303
+ "early_stopping_threshold": 0.01
304
+ },
305
+ "attributes": {
306
+ "early_stopping_patience_counter": 9
307
+ }
308
+ },
309
+ "TrainerControl": {
310
+ "args": {
311
+ "should_epoch_stop": false,
312
+ "should_evaluate": false,
313
+ "should_log": false,
314
+ "should_save": true,
315
+ "should_training_stop": false
316
+ },
317
+ "attributes": {}
318
+ }
319
+ },
320
+ "total_flos": 2.175595314432671e+20,
321
+ "train_batch_size": 256,
322
+ "trial_name": null,
323
+ "trial_params": null
324
+ }
checkpoint-1562/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5303c2eee1a2a0f03943849b01f3f9eb8ddea13fa808e8ed7ade34eba227eba4
3
+ size 5368
checkpoint-1704/config.json ADDED
@@ -0,0 +1,169 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/dit-large",
3
+ "add_fpn": false,
4
+ "architectures": [
5
+ "BeitForImageClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.0,
8
+ "auxiliary_channels": 256,
9
+ "auxiliary_concat_input": false,
10
+ "auxiliary_loss_weight": 0.4,
11
+ "auxiliary_num_convs": 1,
12
+ "drop_path_rate": 0.1,
13
+ "hidden_act": "gelu",
14
+ "hidden_dropout_prob": 0.0,
15
+ "hidden_size": 1024,
16
+ "id2label": {
17
+ "0": "informe_social",
18
+ "1": "cargo_ingreso_mpe",
19
+ "2": "acta_registro_personal",
20
+ "3": "citacion_pnp",
21
+ "4": "acta_lectura_derechos_ley30364",
22
+ "5": "acta_intervencion_minpu",
23
+ "6": "denuncia_policial",
24
+ "7": "providencia_fiscal",
25
+ "8": "consulta_medidas_proteccion",
26
+ "9": "oficio_medicina_legal",
27
+ "10": "oficio_evaluacion_psicologica",
28
+ "11": "resolucion_judicial_audiencia",
29
+ "12": "certificado_medico_legal",
30
+ "13": "ficha_datos_sidpol",
31
+ "14": "consulta_sucamec",
32
+ "15": "informe_mimp",
33
+ "16": "ficha_valoracion_riesgo_manual",
34
+ "17": "oficio_atencion_integral",
35
+ "18": "escrito_minpu",
36
+ "19": "informe_psicologico",
37
+ "20": "ficha_valoracion_riesgo_digital",
38
+ "21": "informe_medico",
39
+ "22": "croquis_domicilio",
40
+ "23": "ficha_datos_reniec",
41
+ "24": "escrito_pj",
42
+ "25": "evidencia_chat",
43
+ "26": "declaracion_minpu",
44
+ "27": "notificacion_resolucion_judicial_audiencia",
45
+ "28": "consulta_antecedentes_penales",
46
+ "29": "notificacion_detencion_pnp",
47
+ "30": "consulta_personas_sidpol",
48
+ "31": "escrito_pnp",
49
+ "32": "constancia_buen_trato",
50
+ "33": "declaracion_pnp",
51
+ "34": "escrito_mimp",
52
+ "35": "acta_intervencion_pnp",
53
+ "36": "constancia_notificacion",
54
+ "37": "cargo_ingreso_sij",
55
+ "38": "anexo_dni",
56
+ "39": "hoja_blanco",
57
+ "40": "consulta_requisitorias"
58
+ },
59
+ "image_size": 224,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "label2id": {
63
+ "acta_intervencion_minpu": 5,
64
+ "acta_intervencion_pnp": 35,
65
+ "acta_lectura_derechos_ley30364": 4,
66
+ "acta_registro_personal": 2,
67
+ "anexo_dni": 38,
68
+ "cargo_ingreso_mpe": 1,
69
+ "cargo_ingreso_sij": 37,
70
+ "certificado_medico_legal": 12,
71
+ "citacion_pnp": 3,
72
+ "constancia_buen_trato": 32,
73
+ "constancia_notificacion": 36,
74
+ "consulta_antecedentes_penales": 28,
75
+ "consulta_medidas_proteccion": 8,
76
+ "consulta_personas_sidpol": 30,
77
+ "consulta_requisitorias": 40,
78
+ "consulta_sucamec": 14,
79
+ "croquis_domicilio": 22,
80
+ "declaracion_minpu": 26,
81
+ "declaracion_pnp": 33,
82
+ "denuncia_policial": 6,
83
+ "escrito_mimp": 34,
84
+ "escrito_minpu": 18,
85
+ "escrito_pj": 24,
86
+ "escrito_pnp": 31,
87
+ "evidencia_chat": 25,
88
+ "ficha_datos_reniec": 23,
89
+ "ficha_datos_sidpol": 13,
90
+ "ficha_valoracion_riesgo_digital": 20,
91
+ "ficha_valoracion_riesgo_manual": 16,
92
+ "hoja_blanco": 39,
93
+ "informe_medico": 21,
94
+ "informe_mimp": 15,
95
+ "informe_psicologico": 19,
96
+ "informe_social": 0,
97
+ "notificacion_detencion_pnp": 29,
98
+ "notificacion_resolucion_judicial_audiencia": 27,
99
+ "oficio_atencion_integral": 17,
100
+ "oficio_evaluacion_psicologica": 10,
101
+ "oficio_medicina_legal": 9,
102
+ "providencia_fiscal": 7,
103
+ "resolucion_judicial_audiencia": 11
104
+ },
105
+ "layer_norm_eps": 1e-12,
106
+ "layer_scale_init_value": 0.1,
107
+ "model_type": "beit",
108
+ "num_attention_heads": 16,
109
+ "num_channels": 3,
110
+ "num_hidden_layers": 24,
111
+ "out_features": [
112
+ "stage3",
113
+ "stage5",
114
+ "stage7",
115
+ "stage11"
116
+ ],
117
+ "out_indices": [
118
+ 3,
119
+ 5,
120
+ 7,
121
+ 11
122
+ ],
123
+ "patch_size": 16,
124
+ "pool_scales": [
125
+ 1,
126
+ 2,
127
+ 3,
128
+ 6
129
+ ],
130
+ "problem_type": "single_label_classification",
131
+ "reshape_hidden_states": true,
132
+ "semantic_loss_ignore_index": 255,
133
+ "stage_names": [
134
+ "stem",
135
+ "stage1",
136
+ "stage2",
137
+ "stage3",
138
+ "stage4",
139
+ "stage5",
140
+ "stage6",
141
+ "stage7",
142
+ "stage8",
143
+ "stage9",
144
+ "stage10",
145
+ "stage11",
146
+ "stage12",
147
+ "stage13",
148
+ "stage14",
149
+ "stage15",
150
+ "stage16",
151
+ "stage17",
152
+ "stage18",
153
+ "stage19",
154
+ "stage20",
155
+ "stage21",
156
+ "stage22",
157
+ "stage23",
158
+ "stage24"
159
+ ],
160
+ "torch_dtype": "float32",
161
+ "transformers_version": "4.46.3",
162
+ "use_absolute_position_embeddings": true,
163
+ "use_auxiliary_head": true,
164
+ "use_mask_token": true,
165
+ "use_mean_pooling": true,
166
+ "use_relative_position_bias": false,
167
+ "use_shared_relative_position_bias": false,
168
+ "vocab_size": 8192
169
+ }
checkpoint-1704/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c6632d203c629fa9af481ebf308a317edf1c38fece5afbda9cdf207ad2242bea
3
+ size 1213526036
checkpoint-1704/preprocessor_config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": {
3
+ "height": 224,
4
+ "width": 224
5
+ },
6
+ "do_center_crop": false,
7
+ "do_normalize": true,
8
+ "do_reduce_labels": false,
9
+ "do_rescale": true,
10
+ "do_resize": true,
11
+ "image_mean": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "image_processor_type": "BeitImageProcessor",
17
+ "image_std": [
18
+ 0.5,
19
+ 0.5,
20
+ 0.5
21
+ ],
22
+ "resample": 2,
23
+ "rescale_factor": 0.00392156862745098,
24
+ "size": {
25
+ "height": 224,
26
+ "width": 224
27
+ }
28
+ }
checkpoint-1704/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8840807761b431e17dae94107ab83fda91f2983bdb562dae75718a2af123594
3
+ size 14244
checkpoint-1704/trainer_state.json ADDED
@@ -0,0 +1,349 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8528685149601977,
3
+ "best_model_checkpoint": "ds3-img-classification/checkpoint-284",
4
+ "epoch": 26.41860465116279,
5
+ "eval_steps": 142,
6
+ "global_step": 1704,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.015503875968992248,
13
+ "grad_norm": 14.292332649230957,
14
+ "learning_rate": 1.736111111111111e-06,
15
+ "loss": 5.8979,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 2.201550387596899,
20
+ "grad_norm": 3.2709274291992188,
21
+ "learning_rate": 0.0002465277777777778,
22
+ "loss": 2.2063,
23
+ "step": 142
24
+ },
25
+ {
26
+ "epoch": 2.201550387596899,
27
+ "eval_accuracy": 0.8300850947021685,
28
+ "eval_loss": 0.6956288814544678,
29
+ "eval_macro_f1": 0.8020115300522307,
30
+ "eval_macro_precision": 0.8326277180719005,
31
+ "eval_macro_recall": 0.8130180643302005,
32
+ "eval_micro_f1": 0.8300850947021685,
33
+ "eval_micro_precision": 0.8300850947021685,
34
+ "eval_micro_recall": 0.8300850947021685,
35
+ "eval_runtime": 8.1254,
36
+ "eval_samples_per_second": 448.347,
37
+ "eval_steps_per_second": 3.569,
38
+ "eval_weighted_f1": 0.8217784359221563,
39
+ "eval_weighted_precision": 0.8439879212180366,
40
+ "eval_weighted_recall": 0.8300850947021685,
41
+ "step": 142
42
+ },
43
+ {
44
+ "epoch": 4.403100775193798,
45
+ "grad_norm": 3.7781729698181152,
46
+ "learning_rate": 0.0004930555555555556,
47
+ "loss": 0.4906,
48
+ "step": 284
49
+ },
50
+ {
51
+ "epoch": 4.403100775193798,
52
+ "eval_accuracy": 0.8528685149601977,
53
+ "eval_loss": 0.5704612731933594,
54
+ "eval_macro_f1": 0.8447646928874578,
55
+ "eval_macro_precision": 0.8434580548053667,
56
+ "eval_macro_recall": 0.8590562294756982,
57
+ "eval_micro_f1": 0.8528685149601977,
58
+ "eval_micro_precision": 0.8528685149601977,
59
+ "eval_micro_recall": 0.8528685149601977,
60
+ "eval_runtime": 8.3026,
61
+ "eval_samples_per_second": 438.779,
62
+ "eval_steps_per_second": 3.493,
63
+ "eval_weighted_f1": 0.8542422373852568,
64
+ "eval_weighted_precision": 0.8661150819486806,
65
+ "eval_weighted_recall": 0.8528685149601977,
66
+ "step": 284
67
+ },
68
+ {
69
+ "epoch": 6.604651162790698,
70
+ "grad_norm": 2.7355144023895264,
71
+ "learning_rate": 0.0007395833333333334,
72
+ "loss": 0.4125,
73
+ "step": 426
74
+ },
75
+ {
76
+ "epoch": 6.604651162790698,
77
+ "eval_accuracy": 0.7400494098270656,
78
+ "eval_loss": 1.1121388673782349,
79
+ "eval_macro_f1": 0.7104399369512154,
80
+ "eval_macro_precision": 0.7984066441843661,
81
+ "eval_macro_recall": 0.7246907362358344,
82
+ "eval_micro_f1": 0.7400494098270657,
83
+ "eval_micro_precision": 0.7400494098270656,
84
+ "eval_micro_recall": 0.7400494098270656,
85
+ "eval_runtime": 8.3145,
86
+ "eval_samples_per_second": 438.152,
87
+ "eval_steps_per_second": 3.488,
88
+ "eval_weighted_f1": 0.7343329270391191,
89
+ "eval_weighted_precision": 0.8109139038669717,
90
+ "eval_weighted_recall": 0.7400494098270656,
91
+ "step": 426
92
+ },
93
+ {
94
+ "epoch": 8.806201550387597,
95
+ "grad_norm": 2.7910587787628174,
96
+ "learning_rate": 0.0009861111111111112,
97
+ "loss": 0.3948,
98
+ "step": 568
99
+ },
100
+ {
101
+ "epoch": 8.806201550387597,
102
+ "eval_accuracy": 0.769146307987922,
103
+ "eval_loss": 0.9053287506103516,
104
+ "eval_macro_f1": 0.7380470144030622,
105
+ "eval_macro_precision": 0.7929019396341611,
106
+ "eval_macro_recall": 0.7416184656713186,
107
+ "eval_micro_f1": 0.769146307987922,
108
+ "eval_micro_precision": 0.769146307987922,
109
+ "eval_micro_recall": 0.769146307987922,
110
+ "eval_runtime": 8.4314,
111
+ "eval_samples_per_second": 432.076,
112
+ "eval_steps_per_second": 3.44,
113
+ "eval_weighted_f1": 0.7647284160437106,
114
+ "eval_weighted_precision": 0.804728448308173,
115
+ "eval_weighted_recall": 0.769146307987922,
116
+ "step": 568
117
+ },
118
+ {
119
+ "epoch": 11.007751937984496,
120
+ "grad_norm": 6.627900123596191,
121
+ "learning_rate": 0.0009983522903302263,
122
+ "loss": 0.3706,
123
+ "step": 710
124
+ },
125
+ {
126
+ "epoch": 11.007751937984496,
127
+ "eval_accuracy": 0.7672248147131485,
128
+ "eval_loss": 0.9641202688217163,
129
+ "eval_macro_f1": 0.7528371715147637,
130
+ "eval_macro_precision": 0.787293122179915,
131
+ "eval_macro_recall": 0.7589348598973276,
132
+ "eval_micro_f1": 0.7672248147131485,
133
+ "eval_micro_precision": 0.7672248147131485,
134
+ "eval_micro_recall": 0.7672248147131485,
135
+ "eval_runtime": 8.3804,
136
+ "eval_samples_per_second": 434.705,
137
+ "eval_steps_per_second": 3.46,
138
+ "eval_weighted_f1": 0.7623056450515923,
139
+ "eval_weighted_precision": 0.7946085579995245,
140
+ "eval_weighted_recall": 0.7672248147131485,
141
+ "step": 710
142
+ },
143
+ {
144
+ "epoch": 13.209302325581396,
145
+ "grad_norm": 3.9615182876586914,
146
+ "learning_rate": 0.0009930222532081596,
147
+ "loss": 0.3302,
148
+ "step": 852
149
+ },
150
+ {
151
+ "epoch": 13.209302325581396,
152
+ "eval_accuracy": 0.7669503156738952,
153
+ "eval_loss": 0.9978815317153931,
154
+ "eval_macro_f1": 0.731377351970227,
155
+ "eval_macro_precision": 0.7752862718850362,
156
+ "eval_macro_recall": 0.7408605973033737,
157
+ "eval_micro_f1": 0.766950315673895,
158
+ "eval_micro_precision": 0.7669503156738952,
159
+ "eval_micro_recall": 0.7669503156738952,
160
+ "eval_runtime": 8.3024,
161
+ "eval_samples_per_second": 438.791,
162
+ "eval_steps_per_second": 3.493,
163
+ "eval_weighted_f1": 0.7591614272562255,
164
+ "eval_weighted_precision": 0.7900342900911238,
165
+ "eval_weighted_recall": 0.7669503156738952,
166
+ "step": 852
167
+ },
168
+ {
169
+ "epoch": 15.410852713178295,
170
+ "grad_norm": 2.4717087745666504,
171
+ "learning_rate": 0.0009840434590541415,
172
+ "loss": 0.2924,
173
+ "step": 994
174
+ },
175
+ {
176
+ "epoch": 15.410852713178295,
177
+ "eval_accuracy": 0.7233049684326105,
178
+ "eval_loss": 1.1638933420181274,
179
+ "eval_macro_f1": 0.7113593382146706,
180
+ "eval_macro_precision": 0.759489526605669,
181
+ "eval_macro_recall": 0.7371465474721075,
182
+ "eval_micro_f1": 0.7233049684326105,
183
+ "eval_micro_precision": 0.7233049684326105,
184
+ "eval_micro_recall": 0.7233049684326105,
185
+ "eval_runtime": 8.2739,
186
+ "eval_samples_per_second": 440.3,
187
+ "eval_steps_per_second": 3.505,
188
+ "eval_weighted_f1": 0.7249947526290721,
189
+ "eval_weighted_precision": 0.7928301153657854,
190
+ "eval_weighted_recall": 0.7233049684326105,
191
+ "step": 994
192
+ },
193
+ {
194
+ "epoch": 17.612403100775193,
195
+ "grad_norm": 1.886505126953125,
196
+ "learning_rate": 0.0009714823580904379,
197
+ "loss": 0.271,
198
+ "step": 1136
199
+ },
200
+ {
201
+ "epoch": 17.612403100775193,
202
+ "eval_accuracy": 0.7631073291243481,
203
+ "eval_loss": 0.9723167419433594,
204
+ "eval_macro_f1": 0.7630758622320426,
205
+ "eval_macro_precision": 0.8404983162678952,
206
+ "eval_macro_recall": 0.7684491516518803,
207
+ "eval_micro_f1": 0.7631073291243481,
208
+ "eval_micro_precision": 0.7631073291243481,
209
+ "eval_micro_recall": 0.7631073291243481,
210
+ "eval_runtime": 8.2536,
211
+ "eval_samples_per_second": 441.381,
212
+ "eval_steps_per_second": 3.514,
213
+ "eval_weighted_f1": 0.7641468306327522,
214
+ "eval_weighted_precision": 0.8384853206385829,
215
+ "eval_weighted_recall": 0.7631073291243481,
216
+ "step": 1136
217
+ },
218
+ {
219
+ "epoch": 19.813953488372093,
220
+ "grad_norm": 1.751386284828186,
221
+ "learning_rate": 0.000955431912460588,
222
+ "loss": 0.2493,
223
+ "step": 1278
224
+ },
225
+ {
226
+ "epoch": 19.813953488372093,
227
+ "eval_accuracy": 0.7556958550645073,
228
+ "eval_loss": 0.9924145340919495,
229
+ "eval_macro_f1": 0.7186954972930405,
230
+ "eval_macro_precision": 0.7780354194554568,
231
+ "eval_macro_recall": 0.7246815062488177,
232
+ "eval_micro_f1": 0.7556958550645072,
233
+ "eval_micro_precision": 0.7556958550645073,
234
+ "eval_micro_recall": 0.7556958550645073,
235
+ "eval_runtime": 8.2332,
236
+ "eval_samples_per_second": 442.478,
237
+ "eval_steps_per_second": 3.522,
238
+ "eval_weighted_f1": 0.7424053023538434,
239
+ "eval_weighted_precision": 0.7815999680084916,
240
+ "eval_weighted_recall": 0.7556958550645073,
241
+ "step": 1278
242
+ },
243
+ {
244
+ "epoch": 22.015503875968992,
245
+ "grad_norm": 1.887330174446106,
246
+ "learning_rate": 0.0009360109082355581,
247
+ "loss": 0.2276,
248
+ "step": 1420
249
+ },
250
+ {
251
+ "epoch": 22.015503875968992,
252
+ "eval_accuracy": 0.7488333790831732,
253
+ "eval_loss": 1.0906082391738892,
254
+ "eval_macro_f1": 0.7232412414342934,
255
+ "eval_macro_precision": 0.7888326922300277,
256
+ "eval_macro_recall": 0.7295020066680354,
257
+ "eval_micro_f1": 0.7488333790831732,
258
+ "eval_micro_precision": 0.7488333790831732,
259
+ "eval_micro_recall": 0.7488333790831732,
260
+ "eval_runtime": 8.3573,
261
+ "eval_samples_per_second": 435.905,
262
+ "eval_steps_per_second": 3.47,
263
+ "eval_weighted_f1": 0.7411123038164863,
264
+ "eval_weighted_precision": 0.78924044171482,
265
+ "eval_weighted_recall": 0.7488333790831732,
266
+ "step": 1420
267
+ },
268
+ {
269
+ "epoch": 24.217054263565892,
270
+ "grad_norm": 1.4166276454925537,
271
+ "learning_rate": 0.0009133630763022861,
272
+ "loss": 0.2147,
273
+ "step": 1562
274
+ },
275
+ {
276
+ "epoch": 24.217054263565892,
277
+ "eval_accuracy": 0.7595388416140544,
278
+ "eval_loss": 1.046807050704956,
279
+ "eval_macro_f1": 0.7218058945395823,
280
+ "eval_macro_precision": 0.7806200306355751,
281
+ "eval_macro_recall": 0.7363816501236903,
282
+ "eval_micro_f1": 0.7595388416140544,
283
+ "eval_micro_precision": 0.7595388416140544,
284
+ "eval_micro_recall": 0.7595388416140544,
285
+ "eval_runtime": 8.3705,
286
+ "eval_samples_per_second": 435.218,
287
+ "eval_steps_per_second": 3.465,
288
+ "eval_weighted_f1": 0.7500356793933665,
289
+ "eval_weighted_precision": 0.7963937500158152,
290
+ "eval_weighted_recall": 0.7595388416140544,
291
+ "step": 1562
292
+ },
293
+ {
294
+ "epoch": 26.41860465116279,
295
+ "grad_norm": 1.3001500368118286,
296
+ "learning_rate": 0.0008876560286407329,
297
+ "loss": 0.1915,
298
+ "step": 1704
299
+ },
300
+ {
301
+ "epoch": 26.41860465116279,
302
+ "eval_accuracy": 0.7757342849300027,
303
+ "eval_loss": 0.9456538558006287,
304
+ "eval_macro_f1": 0.7518200808982276,
305
+ "eval_macro_precision": 0.7941242669868954,
306
+ "eval_macro_recall": 0.7603347162166142,
307
+ "eval_micro_f1": 0.7757342849300026,
308
+ "eval_micro_precision": 0.7757342849300027,
309
+ "eval_micro_recall": 0.7757342849300027,
310
+ "eval_runtime": 8.4106,
311
+ "eval_samples_per_second": 433.145,
312
+ "eval_steps_per_second": 3.448,
313
+ "eval_weighted_f1": 0.7688586665917762,
314
+ "eval_weighted_precision": 0.8022618721710044,
315
+ "eval_weighted_recall": 0.7757342849300027,
316
+ "step": 1704
317
+ }
318
+ ],
319
+ "logging_steps": 142,
320
+ "max_steps": 5760,
321
+ "num_input_tokens_seen": 0,
322
+ "num_train_epochs": 90,
323
+ "save_steps": 142,
324
+ "stateful_callbacks": {
325
+ "EarlyStoppingCallback": {
326
+ "args": {
327
+ "early_stopping_patience": 10,
328
+ "early_stopping_threshold": 0.01
329
+ },
330
+ "attributes": {
331
+ "early_stopping_patience_counter": 10
332
+ }
333
+ },
334
+ "TrainerControl": {
335
+ "args": {
336
+ "should_epoch_stop": false,
337
+ "should_evaluate": false,
338
+ "should_log": false,
339
+ "should_save": true,
340
+ "should_training_stop": true
341
+ },
342
+ "attributes": {}
343
+ }
344
+ },
345
+ "total_flos": 2.3734952710727598e+20,
346
+ "train_batch_size": 256,
347
+ "trial_name": null,
348
+ "trial_params": null
349
+ }
checkpoint-1704/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5303c2eee1a2a0f03943849b01f3f9eb8ddea13fa808e8ed7ade34eba227eba4
3
+ size 5368
checkpoint-284/config.json ADDED
@@ -0,0 +1,169 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/dit-large",
3
+ "add_fpn": false,
4
+ "architectures": [
5
+ "BeitForImageClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.0,
8
+ "auxiliary_channels": 256,
9
+ "auxiliary_concat_input": false,
10
+ "auxiliary_loss_weight": 0.4,
11
+ "auxiliary_num_convs": 1,
12
+ "drop_path_rate": 0.1,
13
+ "hidden_act": "gelu",
14
+ "hidden_dropout_prob": 0.0,
15
+ "hidden_size": 1024,
16
+ "id2label": {
17
+ "0": "informe_social",
18
+ "1": "cargo_ingreso_mpe",
19
+ "2": "acta_registro_personal",
20
+ "3": "citacion_pnp",
21
+ "4": "acta_lectura_derechos_ley30364",
22
+ "5": "acta_intervencion_minpu",
23
+ "6": "denuncia_policial",
24
+ "7": "providencia_fiscal",
25
+ "8": "consulta_medidas_proteccion",
26
+ "9": "oficio_medicina_legal",
27
+ "10": "oficio_evaluacion_psicologica",
28
+ "11": "resolucion_judicial_audiencia",
29
+ "12": "certificado_medico_legal",
30
+ "13": "ficha_datos_sidpol",
31
+ "14": "consulta_sucamec",
32
+ "15": "informe_mimp",
33
+ "16": "ficha_valoracion_riesgo_manual",
34
+ "17": "oficio_atencion_integral",
35
+ "18": "escrito_minpu",
36
+ "19": "informe_psicologico",
37
+ "20": "ficha_valoracion_riesgo_digital",
38
+ "21": "informe_medico",
39
+ "22": "croquis_domicilio",
40
+ "23": "ficha_datos_reniec",
41
+ "24": "escrito_pj",
42
+ "25": "evidencia_chat",
43
+ "26": "declaracion_minpu",
44
+ "27": "notificacion_resolucion_judicial_audiencia",
45
+ "28": "consulta_antecedentes_penales",
46
+ "29": "notificacion_detencion_pnp",
47
+ "30": "consulta_personas_sidpol",
48
+ "31": "escrito_pnp",
49
+ "32": "constancia_buen_trato",
50
+ "33": "declaracion_pnp",
51
+ "34": "escrito_mimp",
52
+ "35": "acta_intervencion_pnp",
53
+ "36": "constancia_notificacion",
54
+ "37": "cargo_ingreso_sij",
55
+ "38": "anexo_dni",
56
+ "39": "hoja_blanco",
57
+ "40": "consulta_requisitorias"
58
+ },
59
+ "image_size": 224,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "label2id": {
63
+ "acta_intervencion_minpu": 5,
64
+ "acta_intervencion_pnp": 35,
65
+ "acta_lectura_derechos_ley30364": 4,
66
+ "acta_registro_personal": 2,
67
+ "anexo_dni": 38,
68
+ "cargo_ingreso_mpe": 1,
69
+ "cargo_ingreso_sij": 37,
70
+ "certificado_medico_legal": 12,
71
+ "citacion_pnp": 3,
72
+ "constancia_buen_trato": 32,
73
+ "constancia_notificacion": 36,
74
+ "consulta_antecedentes_penales": 28,
75
+ "consulta_medidas_proteccion": 8,
76
+ "consulta_personas_sidpol": 30,
77
+ "consulta_requisitorias": 40,
78
+ "consulta_sucamec": 14,
79
+ "croquis_domicilio": 22,
80
+ "declaracion_minpu": 26,
81
+ "declaracion_pnp": 33,
82
+ "denuncia_policial": 6,
83
+ "escrito_mimp": 34,
84
+ "escrito_minpu": 18,
85
+ "escrito_pj": 24,
86
+ "escrito_pnp": 31,
87
+ "evidencia_chat": 25,
88
+ "ficha_datos_reniec": 23,
89
+ "ficha_datos_sidpol": 13,
90
+ "ficha_valoracion_riesgo_digital": 20,
91
+ "ficha_valoracion_riesgo_manual": 16,
92
+ "hoja_blanco": 39,
93
+ "informe_medico": 21,
94
+ "informe_mimp": 15,
95
+ "informe_psicologico": 19,
96
+ "informe_social": 0,
97
+ "notificacion_detencion_pnp": 29,
98
+ "notificacion_resolucion_judicial_audiencia": 27,
99
+ "oficio_atencion_integral": 17,
100
+ "oficio_evaluacion_psicologica": 10,
101
+ "oficio_medicina_legal": 9,
102
+ "providencia_fiscal": 7,
103
+ "resolucion_judicial_audiencia": 11
104
+ },
105
+ "layer_norm_eps": 1e-12,
106
+ "layer_scale_init_value": 0.1,
107
+ "model_type": "beit",
108
+ "num_attention_heads": 16,
109
+ "num_channels": 3,
110
+ "num_hidden_layers": 24,
111
+ "out_features": [
112
+ "stage3",
113
+ "stage5",
114
+ "stage7",
115
+ "stage11"
116
+ ],
117
+ "out_indices": [
118
+ 3,
119
+ 5,
120
+ 7,
121
+ 11
122
+ ],
123
+ "patch_size": 16,
124
+ "pool_scales": [
125
+ 1,
126
+ 2,
127
+ 3,
128
+ 6
129
+ ],
130
+ "problem_type": "single_label_classification",
131
+ "reshape_hidden_states": true,
132
+ "semantic_loss_ignore_index": 255,
133
+ "stage_names": [
134
+ "stem",
135
+ "stage1",
136
+ "stage2",
137
+ "stage3",
138
+ "stage4",
139
+ "stage5",
140
+ "stage6",
141
+ "stage7",
142
+ "stage8",
143
+ "stage9",
144
+ "stage10",
145
+ "stage11",
146
+ "stage12",
147
+ "stage13",
148
+ "stage14",
149
+ "stage15",
150
+ "stage16",
151
+ "stage17",
152
+ "stage18",
153
+ "stage19",
154
+ "stage20",
155
+ "stage21",
156
+ "stage22",
157
+ "stage23",
158
+ "stage24"
159
+ ],
160
+ "torch_dtype": "float32",
161
+ "transformers_version": "4.46.3",
162
+ "use_absolute_position_embeddings": true,
163
+ "use_auxiliary_head": true,
164
+ "use_mask_token": true,
165
+ "use_mean_pooling": true,
166
+ "use_relative_position_bias": false,
167
+ "use_shared_relative_position_bias": false,
168
+ "vocab_size": 8192
169
+ }
checkpoint-284/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e398613a129119aa2e107753f0aff1a6b654ae42ae77939f2944f4830fd8fcde
3
+ size 1213526036
checkpoint-284/preprocessor_config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": {
3
+ "height": 224,
4
+ "width": 224
5
+ },
6
+ "do_center_crop": false,
7
+ "do_normalize": true,
8
+ "do_reduce_labels": false,
9
+ "do_rescale": true,
10
+ "do_resize": true,
11
+ "image_mean": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "image_processor_type": "BeitImageProcessor",
17
+ "image_std": [
18
+ 0.5,
19
+ 0.5,
20
+ 0.5
21
+ ],
22
+ "resample": 2,
23
+ "rescale_factor": 0.00392156862745098,
24
+ "size": {
25
+ "height": 224,
26
+ "width": 224
27
+ }
28
+ }
checkpoint-284/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:56d6b61c7133ff78254b85d2496da504539890a8a3c3cd3a2693e6b63c102671
3
+ size 14244
checkpoint-284/trainer_state.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8528685149601977,
3
+ "best_model_checkpoint": "ds3-img-classification/checkpoint-284",
4
+ "epoch": 4.403100775193798,
5
+ "eval_steps": 142,
6
+ "global_step": 284,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.015503875968992248,
13
+ "grad_norm": 14.292332649230957,
14
+ "learning_rate": 1.736111111111111e-06,
15
+ "loss": 5.8979,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 2.201550387596899,
20
+ "grad_norm": 3.2709274291992188,
21
+ "learning_rate": 0.0002465277777777778,
22
+ "loss": 2.2063,
23
+ "step": 142
24
+ },
25
+ {
26
+ "epoch": 2.201550387596899,
27
+ "eval_accuracy": 0.8300850947021685,
28
+ "eval_loss": 0.6956288814544678,
29
+ "eval_macro_f1": 0.8020115300522307,
30
+ "eval_macro_precision": 0.8326277180719005,
31
+ "eval_macro_recall": 0.8130180643302005,
32
+ "eval_micro_f1": 0.8300850947021685,
33
+ "eval_micro_precision": 0.8300850947021685,
34
+ "eval_micro_recall": 0.8300850947021685,
35
+ "eval_runtime": 8.1254,
36
+ "eval_samples_per_second": 448.347,
37
+ "eval_steps_per_second": 3.569,
38
+ "eval_weighted_f1": 0.8217784359221563,
39
+ "eval_weighted_precision": 0.8439879212180366,
40
+ "eval_weighted_recall": 0.8300850947021685,
41
+ "step": 142
42
+ },
43
+ {
44
+ "epoch": 4.403100775193798,
45
+ "grad_norm": 3.7781729698181152,
46
+ "learning_rate": 0.0004930555555555556,
47
+ "loss": 0.4906,
48
+ "step": 284
49
+ },
50
+ {
51
+ "epoch": 4.403100775193798,
52
+ "eval_accuracy": 0.8528685149601977,
53
+ "eval_loss": 0.5704612731933594,
54
+ "eval_macro_f1": 0.8447646928874578,
55
+ "eval_macro_precision": 0.8434580548053667,
56
+ "eval_macro_recall": 0.8590562294756982,
57
+ "eval_micro_f1": 0.8528685149601977,
58
+ "eval_micro_precision": 0.8528685149601977,
59
+ "eval_micro_recall": 0.8528685149601977,
60
+ "eval_runtime": 8.3026,
61
+ "eval_samples_per_second": 438.779,
62
+ "eval_steps_per_second": 3.493,
63
+ "eval_weighted_f1": 0.8542422373852568,
64
+ "eval_weighted_precision": 0.8661150819486806,
65
+ "eval_weighted_recall": 0.8528685149601977,
66
+ "step": 284
67
+ }
68
+ ],
69
+ "logging_steps": 142,
70
+ "max_steps": 5760,
71
+ "num_input_tokens_seen": 0,
72
+ "num_train_epochs": 90,
73
+ "save_steps": 142,
74
+ "stateful_callbacks": {
75
+ "EarlyStoppingCallback": {
76
+ "args": {
77
+ "early_stopping_patience": 10,
78
+ "early_stopping_threshold": 0.01
79
+ },
80
+ "attributes": {
81
+ "early_stopping_patience_counter": 0
82
+ }
83
+ },
84
+ "TrainerControl": {
85
+ "args": {
86
+ "should_epoch_stop": false,
87
+ "should_evaluate": false,
88
+ "should_log": false,
89
+ "should_save": true,
90
+ "should_training_stop": false
91
+ },
92
+ "attributes": {}
93
+ }
94
+ },
95
+ "total_flos": 3.957999132801776e+19,
96
+ "train_batch_size": 256,
97
+ "trial_name": null,
98
+ "trial_params": null
99
+ }
checkpoint-284/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5303c2eee1a2a0f03943849b01f3f9eb8ddea13fa808e8ed7ade34eba227eba4
3
+ size 5368
metrics.json CHANGED
@@ -1,25 +1,25 @@
1
  {
2
  "train": {
3
- "loss": 5.2299,
4
- "grad_norm": 5.580934524536133,
5
- "learning_rate": 0.0010275543423681622,
6
- "epoch": 0.5426356589147286,
7
- "step": 35.0
8
  },
9
  "eval": {
10
- "eval_loss": 5.006561279296875,
11
- "eval_accuracy": 0.1350535273126544,
12
- "eval_weighted_f1": 0.0662231556079451,
13
- "eval_micro_f1": 0.1350535273126544,
14
- "eval_macro_f1": 0.038001266460012215,
15
- "eval_weighted_recall": 0.1350535273126544,
16
- "eval_micro_recall": 0.1350535273126544,
17
- "eval_macro_recall": 0.06460349403258135,
18
- "eval_weighted_precision": 0.054814565244902975,
19
- "eval_micro_precision": 0.1350535273126544,
20
- "eval_macro_precision": 0.033336047367919855,
21
- "eval_runtime": 8.2574,
22
- "eval_samples_per_second": 441.182,
23
- "eval_steps_per_second": 3.512
24
  }
25
  }
 
1
  {
2
  "train": {
3
+ "loss": 0.1915,
4
+ "grad_norm": 1.3001500368118286,
5
+ "learning_rate": 0.0008876560286407329,
6
+ "epoch": 26.41860465116279,
7
+ "step": 1704.0
8
  },
9
  "eval": {
10
+ "eval_loss": 0.9456538558006287,
11
+ "eval_accuracy": 0.7757342849300027,
12
+ "eval_weighted_f1": 0.7688586665917762,
13
+ "eval_micro_f1": 0.7757342849300026,
14
+ "eval_macro_f1": 0.7518200808982276,
15
+ "eval_weighted_recall": 0.7757342849300027,
16
+ "eval_micro_recall": 0.7757342849300027,
17
+ "eval_macro_recall": 0.7603347162166142,
18
+ "eval_weighted_precision": 0.8022618721710044,
19
+ "eval_micro_precision": 0.7757342849300027,
20
+ "eval_macro_precision": 0.7941242669868954,
21
+ "eval_runtime": 8.4106,
22
+ "eval_samples_per_second": 433.145,
23
+ "eval_steps_per_second": 3.448
24
  }
25
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e72a88f97c761c0a531cb086303001d50691997a2b12f53e5ba97ac052c09560
3
  size 1213526036
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e398613a129119aa2e107753f0aff1a6b654ae42ae77939f2944f4830fd8fcde
3
  size 1213526036
trainer_state.json CHANGED
@@ -1,61 +1,345 @@
1
  {
2
- "best_metric": 0.1350535273126544,
3
- "best_model_checkpoint": "ds3-img-classification/checkpoint-35",
4
- "epoch": 0.9922480620155039,
5
- "eval_steps": 35,
6
- "global_step": 64,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
  "epoch": 0.015503875968992248,
13
- "grad_norm": 16.251585006713867,
14
- "learning_rate": 0.0002857142857142857,
15
- "loss": 6.0665,
16
  "step": 1
17
  },
18
  {
19
- "epoch": 0.5426356589147286,
20
- "grad_norm": 5.580934524536133,
21
- "learning_rate": 0.0010275543423681622,
22
- "loss": 5.2299,
23
- "step": 35
24
- },
25
- {
26
- "epoch": 0.5426356589147286,
27
- "eval_accuracy": 0.1350535273126544,
28
- "eval_loss": 5.006561279296875,
29
- "eval_macro_f1": 0.038001266460012215,
30
- "eval_macro_precision": 0.033336047367919855,
31
- "eval_macro_recall": 0.06460349403258135,
32
- "eval_micro_f1": 0.1350535273126544,
33
- "eval_micro_precision": 0.1350535273126544,
34
- "eval_micro_recall": 0.1350535273126544,
35
- "eval_runtime": 8.2574,
36
- "eval_samples_per_second": 441.182,
37
- "eval_steps_per_second": 3.512,
38
- "eval_weighted_f1": 0.0662231556079451,
39
- "eval_weighted_precision": 0.054814565244902975,
40
- "eval_weighted_recall": 0.1350535273126544,
41
- "step": 35
42
- },
43
- {
44
- "epoch": 0.9922480620155039,
45
- "step": 64,
46
- "total_flos": 8.978215898519175e+18,
47
- "train_loss": 4.950647212564945,
48
- "train_runtime": 163.8995,
49
- "train_samples_per_second": 200.037,
50
- "train_steps_per_second": 0.39
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
51
  }
52
  ],
53
- "logging_steps": 35,
54
- "max_steps": 64,
55
  "num_input_tokens_seen": 0,
56
- "num_train_epochs": 1,
57
- "save_steps": 35,
58
  "stateful_callbacks": {
 
 
 
 
 
 
 
 
 
59
  "TrainerControl": {
60
  "args": {
61
  "should_epoch_stop": false,
@@ -67,7 +351,7 @@
67
  "attributes": {}
68
  }
69
  },
70
- "total_flos": 8.978215898519175e+18,
71
  "train_batch_size": 256,
72
  "trial_name": null,
73
  "trial_params": null
 
1
  {
2
+ "best_metric": 0.8528685149601977,
3
+ "best_model_checkpoint": "ds3-img-classification/checkpoint-284",
4
+ "epoch": 26.41860465116279,
5
+ "eval_steps": 142,
6
+ "global_step": 1704,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
  "epoch": 0.015503875968992248,
13
+ "grad_norm": 14.292332649230957,
14
+ "learning_rate": 1.736111111111111e-06,
15
+ "loss": 5.8979,
16
  "step": 1
17
  },
18
  {
19
+ "epoch": 2.201550387596899,
20
+ "grad_norm": 3.2709274291992188,
21
+ "learning_rate": 0.0002465277777777778,
22
+ "loss": 2.2063,
23
+ "step": 142
24
+ },
25
+ {
26
+ "epoch": 2.201550387596899,
27
+ "eval_accuracy": 0.8300850947021685,
28
+ "eval_loss": 0.6956288814544678,
29
+ "eval_macro_f1": 0.8020115300522307,
30
+ "eval_macro_precision": 0.8326277180719005,
31
+ "eval_macro_recall": 0.8130180643302005,
32
+ "eval_micro_f1": 0.8300850947021685,
33
+ "eval_micro_precision": 0.8300850947021685,
34
+ "eval_micro_recall": 0.8300850947021685,
35
+ "eval_runtime": 8.1254,
36
+ "eval_samples_per_second": 448.347,
37
+ "eval_steps_per_second": 3.569,
38
+ "eval_weighted_f1": 0.8217784359221563,
39
+ "eval_weighted_precision": 0.8439879212180366,
40
+ "eval_weighted_recall": 0.8300850947021685,
41
+ "step": 142
42
+ },
43
+ {
44
+ "epoch": 4.403100775193798,
45
+ "grad_norm": 3.7781729698181152,
46
+ "learning_rate": 0.0004930555555555556,
47
+ "loss": 0.4906,
48
+ "step": 284
49
+ },
50
+ {
51
+ "epoch": 4.403100775193798,
52
+ "eval_accuracy": 0.8528685149601977,
53
+ "eval_loss": 0.5704612731933594,
54
+ "eval_macro_f1": 0.8447646928874578,
55
+ "eval_macro_precision": 0.8434580548053667,
56
+ "eval_macro_recall": 0.8590562294756982,
57
+ "eval_micro_f1": 0.8528685149601977,
58
+ "eval_micro_precision": 0.8528685149601977,
59
+ "eval_micro_recall": 0.8528685149601977,
60
+ "eval_runtime": 8.3026,
61
+ "eval_samples_per_second": 438.779,
62
+ "eval_steps_per_second": 3.493,
63
+ "eval_weighted_f1": 0.8542422373852568,
64
+ "eval_weighted_precision": 0.8661150819486806,
65
+ "eval_weighted_recall": 0.8528685149601977,
66
+ "step": 284
67
+ },
68
+ {
69
+ "epoch": 6.604651162790698,
70
+ "grad_norm": 2.7355144023895264,
71
+ "learning_rate": 0.0007395833333333334,
72
+ "loss": 0.4125,
73
+ "step": 426
74
+ },
75
+ {
76
+ "epoch": 6.604651162790698,
77
+ "eval_accuracy": 0.7400494098270656,
78
+ "eval_loss": 1.1121388673782349,
79
+ "eval_macro_f1": 0.7104399369512154,
80
+ "eval_macro_precision": 0.7984066441843661,
81
+ "eval_macro_recall": 0.7246907362358344,
82
+ "eval_micro_f1": 0.7400494098270657,
83
+ "eval_micro_precision": 0.7400494098270656,
84
+ "eval_micro_recall": 0.7400494098270656,
85
+ "eval_runtime": 8.3145,
86
+ "eval_samples_per_second": 438.152,
87
+ "eval_steps_per_second": 3.488,
88
+ "eval_weighted_f1": 0.7343329270391191,
89
+ "eval_weighted_precision": 0.8109139038669717,
90
+ "eval_weighted_recall": 0.7400494098270656,
91
+ "step": 426
92
+ },
93
+ {
94
+ "epoch": 8.806201550387597,
95
+ "grad_norm": 2.7910587787628174,
96
+ "learning_rate": 0.0009861111111111112,
97
+ "loss": 0.3948,
98
+ "step": 568
99
+ },
100
+ {
101
+ "epoch": 8.806201550387597,
102
+ "eval_accuracy": 0.769146307987922,
103
+ "eval_loss": 0.9053287506103516,
104
+ "eval_macro_f1": 0.7380470144030622,
105
+ "eval_macro_precision": 0.7929019396341611,
106
+ "eval_macro_recall": 0.7416184656713186,
107
+ "eval_micro_f1": 0.769146307987922,
108
+ "eval_micro_precision": 0.769146307987922,
109
+ "eval_micro_recall": 0.769146307987922,
110
+ "eval_runtime": 8.4314,
111
+ "eval_samples_per_second": 432.076,
112
+ "eval_steps_per_second": 3.44,
113
+ "eval_weighted_f1": 0.7647284160437106,
114
+ "eval_weighted_precision": 0.804728448308173,
115
+ "eval_weighted_recall": 0.769146307987922,
116
+ "step": 568
117
+ },
118
+ {
119
+ "epoch": 11.007751937984496,
120
+ "grad_norm": 6.627900123596191,
121
+ "learning_rate": 0.0009983522903302263,
122
+ "loss": 0.3706,
123
+ "step": 710
124
+ },
125
+ {
126
+ "epoch": 11.007751937984496,
127
+ "eval_accuracy": 0.7672248147131485,
128
+ "eval_loss": 0.9641202688217163,
129
+ "eval_macro_f1": 0.7528371715147637,
130
+ "eval_macro_precision": 0.787293122179915,
131
+ "eval_macro_recall": 0.7589348598973276,
132
+ "eval_micro_f1": 0.7672248147131485,
133
+ "eval_micro_precision": 0.7672248147131485,
134
+ "eval_micro_recall": 0.7672248147131485,
135
+ "eval_runtime": 8.3804,
136
+ "eval_samples_per_second": 434.705,
137
+ "eval_steps_per_second": 3.46,
138
+ "eval_weighted_f1": 0.7623056450515923,
139
+ "eval_weighted_precision": 0.7946085579995245,
140
+ "eval_weighted_recall": 0.7672248147131485,
141
+ "step": 710
142
+ },
143
+ {
144
+ "epoch": 13.209302325581396,
145
+ "grad_norm": 3.9615182876586914,
146
+ "learning_rate": 0.0009930222532081596,
147
+ "loss": 0.3302,
148
+ "step": 852
149
+ },
150
+ {
151
+ "epoch": 13.209302325581396,
152
+ "eval_accuracy": 0.7669503156738952,
153
+ "eval_loss": 0.9978815317153931,
154
+ "eval_macro_f1": 0.731377351970227,
155
+ "eval_macro_precision": 0.7752862718850362,
156
+ "eval_macro_recall": 0.7408605973033737,
157
+ "eval_micro_f1": 0.766950315673895,
158
+ "eval_micro_precision": 0.7669503156738952,
159
+ "eval_micro_recall": 0.7669503156738952,
160
+ "eval_runtime": 8.3024,
161
+ "eval_samples_per_second": 438.791,
162
+ "eval_steps_per_second": 3.493,
163
+ "eval_weighted_f1": 0.7591614272562255,
164
+ "eval_weighted_precision": 0.7900342900911238,
165
+ "eval_weighted_recall": 0.7669503156738952,
166
+ "step": 852
167
+ },
168
+ {
169
+ "epoch": 15.410852713178295,
170
+ "grad_norm": 2.4717087745666504,
171
+ "learning_rate": 0.0009840434590541415,
172
+ "loss": 0.2924,
173
+ "step": 994
174
+ },
175
+ {
176
+ "epoch": 15.410852713178295,
177
+ "eval_accuracy": 0.7233049684326105,
178
+ "eval_loss": 1.1638933420181274,
179
+ "eval_macro_f1": 0.7113593382146706,
180
+ "eval_macro_precision": 0.759489526605669,
181
+ "eval_macro_recall": 0.7371465474721075,
182
+ "eval_micro_f1": 0.7233049684326105,
183
+ "eval_micro_precision": 0.7233049684326105,
184
+ "eval_micro_recall": 0.7233049684326105,
185
+ "eval_runtime": 8.2739,
186
+ "eval_samples_per_second": 440.3,
187
+ "eval_steps_per_second": 3.505,
188
+ "eval_weighted_f1": 0.7249947526290721,
189
+ "eval_weighted_precision": 0.7928301153657854,
190
+ "eval_weighted_recall": 0.7233049684326105,
191
+ "step": 994
192
+ },
193
+ {
194
+ "epoch": 17.612403100775193,
195
+ "grad_norm": 1.886505126953125,
196
+ "learning_rate": 0.0009714823580904379,
197
+ "loss": 0.271,
198
+ "step": 1136
199
+ },
200
+ {
201
+ "epoch": 17.612403100775193,
202
+ "eval_accuracy": 0.7631073291243481,
203
+ "eval_loss": 0.9723167419433594,
204
+ "eval_macro_f1": 0.7630758622320426,
205
+ "eval_macro_precision": 0.8404983162678952,
206
+ "eval_macro_recall": 0.7684491516518803,
207
+ "eval_micro_f1": 0.7631073291243481,
208
+ "eval_micro_precision": 0.7631073291243481,
209
+ "eval_micro_recall": 0.7631073291243481,
210
+ "eval_runtime": 8.2536,
211
+ "eval_samples_per_second": 441.381,
212
+ "eval_steps_per_second": 3.514,
213
+ "eval_weighted_f1": 0.7641468306327522,
214
+ "eval_weighted_precision": 0.8384853206385829,
215
+ "eval_weighted_recall": 0.7631073291243481,
216
+ "step": 1136
217
+ },
218
+ {
219
+ "epoch": 19.813953488372093,
220
+ "grad_norm": 1.751386284828186,
221
+ "learning_rate": 0.000955431912460588,
222
+ "loss": 0.2493,
223
+ "step": 1278
224
+ },
225
+ {
226
+ "epoch": 19.813953488372093,
227
+ "eval_accuracy": 0.7556958550645073,
228
+ "eval_loss": 0.9924145340919495,
229
+ "eval_macro_f1": 0.7186954972930405,
230
+ "eval_macro_precision": 0.7780354194554568,
231
+ "eval_macro_recall": 0.7246815062488177,
232
+ "eval_micro_f1": 0.7556958550645072,
233
+ "eval_micro_precision": 0.7556958550645073,
234
+ "eval_micro_recall": 0.7556958550645073,
235
+ "eval_runtime": 8.2332,
236
+ "eval_samples_per_second": 442.478,
237
+ "eval_steps_per_second": 3.522,
238
+ "eval_weighted_f1": 0.7424053023538434,
239
+ "eval_weighted_precision": 0.7815999680084916,
240
+ "eval_weighted_recall": 0.7556958550645073,
241
+ "step": 1278
242
+ },
243
+ {
244
+ "epoch": 22.015503875968992,
245
+ "grad_norm": 1.887330174446106,
246
+ "learning_rate": 0.0009360109082355581,
247
+ "loss": 0.2276,
248
+ "step": 1420
249
+ },
250
+ {
251
+ "epoch": 22.015503875968992,
252
+ "eval_accuracy": 0.7488333790831732,
253
+ "eval_loss": 1.0906082391738892,
254
+ "eval_macro_f1": 0.7232412414342934,
255
+ "eval_macro_precision": 0.7888326922300277,
256
+ "eval_macro_recall": 0.7295020066680354,
257
+ "eval_micro_f1": 0.7488333790831732,
258
+ "eval_micro_precision": 0.7488333790831732,
259
+ "eval_micro_recall": 0.7488333790831732,
260
+ "eval_runtime": 8.3573,
261
+ "eval_samples_per_second": 435.905,
262
+ "eval_steps_per_second": 3.47,
263
+ "eval_weighted_f1": 0.7411123038164863,
264
+ "eval_weighted_precision": 0.78924044171482,
265
+ "eval_weighted_recall": 0.7488333790831732,
266
+ "step": 1420
267
+ },
268
+ {
269
+ "epoch": 24.217054263565892,
270
+ "grad_norm": 1.4166276454925537,
271
+ "learning_rate": 0.0009133630763022861,
272
+ "loss": 0.2147,
273
+ "step": 1562
274
+ },
275
+ {
276
+ "epoch": 24.217054263565892,
277
+ "eval_accuracy": 0.7595388416140544,
278
+ "eval_loss": 1.046807050704956,
279
+ "eval_macro_f1": 0.7218058945395823,
280
+ "eval_macro_precision": 0.7806200306355751,
281
+ "eval_macro_recall": 0.7363816501236903,
282
+ "eval_micro_f1": 0.7595388416140544,
283
+ "eval_micro_precision": 0.7595388416140544,
284
+ "eval_micro_recall": 0.7595388416140544,
285
+ "eval_runtime": 8.3705,
286
+ "eval_samples_per_second": 435.218,
287
+ "eval_steps_per_second": 3.465,
288
+ "eval_weighted_f1": 0.7500356793933665,
289
+ "eval_weighted_precision": 0.7963937500158152,
290
+ "eval_weighted_recall": 0.7595388416140544,
291
+ "step": 1562
292
+ },
293
+ {
294
+ "epoch": 26.41860465116279,
295
+ "grad_norm": 1.3001500368118286,
296
+ "learning_rate": 0.0008876560286407329,
297
+ "loss": 0.1915,
298
+ "step": 1704
299
+ },
300
+ {
301
+ "epoch": 26.41860465116279,
302
+ "eval_accuracy": 0.7757342849300027,
303
+ "eval_loss": 0.9456538558006287,
304
+ "eval_macro_f1": 0.7518200808982276,
305
+ "eval_macro_precision": 0.7941242669868954,
306
+ "eval_macro_recall": 0.7603347162166142,
307
+ "eval_micro_f1": 0.7757342849300026,
308
+ "eval_micro_precision": 0.7757342849300027,
309
+ "eval_micro_recall": 0.7757342849300027,
310
+ "eval_runtime": 8.4106,
311
+ "eval_samples_per_second": 433.145,
312
+ "eval_steps_per_second": 3.448,
313
+ "eval_weighted_f1": 0.7688586665917762,
314
+ "eval_weighted_precision": 0.8022618721710044,
315
+ "eval_weighted_recall": 0.7757342849300027,
316
+ "step": 1704
317
+ },
318
+ {
319
+ "epoch": 26.41860465116279,
320
+ "step": 1704,
321
+ "total_flos": 2.3734952710727598e+20,
322
+ "train_loss": 0.4731367647368024,
323
+ "train_runtime": 2312.991,
324
+ "train_samples_per_second": 1275.725,
325
+ "train_steps_per_second": 2.49
326
  }
327
  ],
328
+ "logging_steps": 142,
329
+ "max_steps": 5760,
330
  "num_input_tokens_seen": 0,
331
+ "num_train_epochs": 90,
332
+ "save_steps": 142,
333
  "stateful_callbacks": {
334
+ "EarlyStoppingCallback": {
335
+ "args": {
336
+ "early_stopping_patience": 10,
337
+ "early_stopping_threshold": 0.01
338
+ },
339
+ "attributes": {
340
+ "early_stopping_patience_counter": 10
341
+ }
342
+ },
343
  "TrainerControl": {
344
  "args": {
345
  "should_epoch_stop": false,
 
351
  "attributes": {}
352
  }
353
  },
354
+ "total_flos": 2.3734952710727598e+20,
355
  "train_batch_size": 256,
356
  "trial_name": null,
357
  "trial_params": null
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c6cc8ffd94025c9e00fadf213825a92f261358a56e78b613f233178b10436ef6
3
  size 5368
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5303c2eee1a2a0f03943849b01f3f9eb8ddea13fa808e8ed7ade34eba227eba4
3
  size 5368