Publicacion del mejor modelo (Parte 9.2)

README.md

@@ -0,0 +1,50 @@
+---
+language: es
+license: apache-2.0
+library_name: transformers
+pipeline_tag: image-classification
+tags:
+- vision-transformer
+- image-classification
+- binary-classification
+- deepfake-detection
+datasets:
+- djramirezp/face-classification-dataset
+metrics:
+- accuracy
+- precision
+- recall
+- f1
+---
+
+# ViT fine-tuned para clasificacion FAKE/REAL
+
+## Resumen metodologico
+Se utilizo el modelo base `google/vit-base-patch16-224-in21k` y se realizo fine-tuning con `Trainer` de Hugging Face.
+Las imagenes se preprocesaron con `AutoImageProcessor` y el entrenamiento se ejecuto con early stopping.
+La seleccion del mejor checkpoint se hizo con base en la metrica F1 de validacion.
+
+## Hiperparametros principales
+- learning_rate: 2e-05
+- batch_size: 16
+- num_train_epochs: 8
+- weight_decay: 0.01
+- warmup_ratio: 0.1
+- early_stopping_patience: 2
+- early_stopping_threshold: 0.001
+
+## Resultados
+
+### Validacion
+- loss: 0.014208527281880379
+- accuracy: 0.9963436928702011
+- precision: 0.9963706766866571
+- recall: 0.9963436928702011
+- f1: 0.996343863976966
+
+### Test
+- loss: 0.03636797145009041
+- accuracy: 0.9920071047957372
+- precision: 0.9920799653809312
+- recall: 0.9920071047957372
+- f1: 0.9920051158081001

checkpoint-1215/config.json

@@ -0,0 +1,33 @@
+{
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "dtype": "float32",
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "FAKE",
+    "1": "REAL"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "FAKE": 0,
+    "REAL": 1
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "pooler_act": "tanh",
+  "pooler_output_size": 768,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "transformers_version": "4.57.6"
+}

checkpoint-1215/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5016fc6ba2598ee6c01c32dadd8a5ca9cd5796421ce14eb80bf7072cacc56f5e
+size 343223968

checkpoint-1215/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17899310ae9a7cfa3507f70914543335242233e27021154fe2758b3e83b67368
+size 686563322

checkpoint-1215/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:013d13e4b269a8648da58d2f40bf48e004388ea0631771b8c8e588ab75375092
+size 13990

checkpoint-1215/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2e8be170ec5af75ca844cd0e5320008abeb06a4e17a5390c4414ff2355d9503
+size 1064

checkpoint-1215/trainer_state.json

@@ -0,0 +1,138 @@
+{
+  "best_global_step": 1215,
+  "best_metric": 0.996343863976966,
+  "best_model_checkpoint": "vit-finetuned-face-classification/checkpoint-1215",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 1215,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.483053982257843,
+      "learning_rate": 1.9462550028587764e-05,
+      "loss": 0.2831,
+      "step": 243
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9872029250457038,
+      "eval_f1": 0.9872035238293116,
+      "eval_loss": 0.07077323645353317,
+      "eval_precision": 0.9875274885409214,
+      "eval_recall": 0.9872029250457038,
+      "eval_runtime": 6.4135,
+      "eval_samples_per_second": 85.289,
+      "eval_steps_per_second": 5.457,
+      "step": 243
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.16903339326381683,
+      "learning_rate": 1.668381932532876e-05,
+      "loss": 0.0492,
+      "step": 486
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9926873857404022,
+      "eval_f1": 0.9926868478468266,
+      "eval_loss": 0.029720129445195198,
+      "eval_precision": 0.9927128664254561,
+      "eval_recall": 0.9926873857404022,
+      "eval_runtime": 6.9345,
+      "eval_samples_per_second": 78.881,
+      "eval_steps_per_second": 5.047,
+      "step": 486
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.11401919275522232,
+      "learning_rate": 1.3905088622069755e-05,
+      "loss": 0.0261,
+      "step": 729
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9963436928702011,
+      "eval_f1": 0.9963434239234135,
+      "eval_loss": 0.01450187899172306,
+      "eval_precision": 0.9963698093496997,
+      "eval_recall": 0.9963436928702011,
+      "eval_runtime": 7.5417,
+      "eval_samples_per_second": 72.531,
+      "eval_steps_per_second": 4.641,
+      "step": 729
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.05247194692492485,
+      "learning_rate": 1.1126357918810751e-05,
+      "loss": 0.0138,
+      "step": 972
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9926873857404022,
+      "eval_f1": 0.9926878745776567,
+      "eval_loss": 0.019891683012247086,
+      "eval_precision": 0.9927945302716784,
+      "eval_recall": 0.9926873857404022,
+      "eval_runtime": 8.4221,
+      "eval_samples_per_second": 64.948,
+      "eval_steps_per_second": 4.156,
+      "step": 972
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.04118124768137932,
+      "learning_rate": 8.347627215551745e-06,
+      "loss": 0.0106,
+      "step": 1215
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9963436928702011,
+      "eval_f1": 0.996343863976966,
+      "eval_loss": 0.014208527281880379,
+      "eval_precision": 0.9963706766866571,
+      "eval_recall": 0.9963436928702011,
+      "eval_runtime": 7.8492,
+      "eval_samples_per_second": 69.689,
+      "eval_steps_per_second": 4.459,
+      "step": 1215
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 1944,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 8,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 2,
+        "early_stopping_threshold": 0.001
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 2
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.5048944383153766e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-1215/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:996910b20e8895000999ca655730e24555f2998d3c6e626c04f395fdf89b53f1
+size 5432

checkpoint-972/config.json

@@ -0,0 +1,33 @@
+{
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "dtype": "float32",
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "FAKE",
+    "1": "REAL"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "FAKE": 0,
+    "REAL": 1
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "pooler_act": "tanh",
+  "pooler_output_size": 768,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "transformers_version": "4.57.6"
+}

checkpoint-972/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:537ca22b31a69b2d5ae12de01562d6b50d5e8dfc442b31c94038425208be89ed
+size 343223968

checkpoint-972/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:207ed87ab0af44cd5b803e54f5e9f965054c4143f08ce0e53c677ce419df0846
+size 686563322

checkpoint-972/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fed77b14fe062f6db72d68cedd6fd95bae3305b7a735eef3c85da43fd15d476
+size 13990

checkpoint-972/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:184b0ab1538186d8e31eb0e2c09c13d416a1c5570e8a7e98c11e5684739b5949
+size 1064

checkpoint-972/trainer_state.json

@@ -0,0 +1,119 @@
+{
+  "best_global_step": 729,
+  "best_metric": 0.9963434239234135,
+  "best_model_checkpoint": "vit-finetuned-face-classification/checkpoint-729",
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 972,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.483053982257843,
+      "learning_rate": 1.9462550028587764e-05,
+      "loss": 0.2831,
+      "step": 243
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9872029250457038,
+      "eval_f1": 0.9872035238293116,
+      "eval_loss": 0.07077323645353317,
+      "eval_precision": 0.9875274885409214,
+      "eval_recall": 0.9872029250457038,
+      "eval_runtime": 6.4135,
+      "eval_samples_per_second": 85.289,
+      "eval_steps_per_second": 5.457,
+      "step": 243
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.16903339326381683,
+      "learning_rate": 1.668381932532876e-05,
+      "loss": 0.0492,
+      "step": 486
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9926873857404022,
+      "eval_f1": 0.9926868478468266,
+      "eval_loss": 0.029720129445195198,
+      "eval_precision": 0.9927128664254561,
+      "eval_recall": 0.9926873857404022,
+      "eval_runtime": 6.9345,
+      "eval_samples_per_second": 78.881,
+      "eval_steps_per_second": 5.047,
+      "step": 486
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.11401919275522232,
+      "learning_rate": 1.3905088622069755e-05,
+      "loss": 0.0261,
+      "step": 729
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9963436928702011,
+      "eval_f1": 0.9963434239234135,
+      "eval_loss": 0.01450187899172306,
+      "eval_precision": 0.9963698093496997,
+      "eval_recall": 0.9963436928702011,
+      "eval_runtime": 7.5417,
+      "eval_samples_per_second": 72.531,
+      "eval_steps_per_second": 4.641,
+      "step": 729
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.05247194692492485,
+      "learning_rate": 1.1126357918810751e-05,
+      "loss": 0.0138,
+      "step": 972
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9926873857404022,
+      "eval_f1": 0.9926878745776567,
+      "eval_loss": 0.019891683012247086,
+      "eval_precision": 0.9927945302716784,
+      "eval_recall": 0.9926873857404022,
+      "eval_runtime": 8.4221,
+      "eval_samples_per_second": 64.948,
+      "eval_steps_per_second": 4.156,
+      "step": 972
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 1944,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 8,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 2,
+        "early_stopping_threshold": 0.001
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 1
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.2039155506523013e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-972/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:996910b20e8895000999ca655730e24555f2998d3c6e626c04f395fdf89b53f1
+size 5432

config.json

@@ -0,0 +1,33 @@
+{
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "dtype": "float32",
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "FAKE",
+    "1": "REAL"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "FAKE": 0,
+    "REAL": 1
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "pooler_act": "tanh",
+  "pooler_output_size": 768,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "transformers_version": "4.57.6"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5016fc6ba2598ee6c01c32dadd8a5ca9cd5796421ce14eb80bf7072cacc56f5e
+size 343223968

preprocessor_config.json

@@ -0,0 +1,23 @@
+{
+  "do_convert_rgb": null,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:996910b20e8895000999ca655730e24555f2998d3c6e626c04f395fdf89b53f1
+size 5432