Upload 5 files

config.json

@@ -0,0 +1,48 @@
+{
+  "architectures": [
+    "SiglipForImageClassification"
+  ],
+  "id2label": {
+    "0": "graphically_violent",
+    "1": "nudity_pornography",
+    "2": "safe_normal"
+  },
+  "initializer_factor": 1.0,
+  "label2id": {
+    "graphically_violent": 0,
+    "nudity_pornography": 1,
+    "safe_normal": 2
+  },
+  "model_type": "siglip",
+  "problem_type": "single_label_classification",
+  "text_config": {
+    "attention_dropout": 0.0,
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 768,
+    "intermediate_size": 3072,
+    "layer_norm_eps": 1e-06,
+    "max_position_embeddings": 64,
+    "model_type": "siglip_text_model",
+    "num_attention_heads": 12,
+    "num_hidden_layers": 12,
+    "projection_size": 768,
+    "torch_dtype": "float32",
+    "vocab_size": 256000
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.51.3",
+  "vision_config": {
+    "attention_dropout": 0.0,
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 768,
+    "image_size": 224,
+    "intermediate_size": 3072,
+    "layer_norm_eps": 1e-06,
+    "model_type": "siglip_vision_model",
+    "num_attention_heads": 12,
+    "num_channels": 3,
+    "num_hidden_layers": 12,
+    "patch_size": 16,
+    "torch_dtype": "float32"
+  }
+}

preprocessor_config.json

@@ -0,0 +1,24 @@
+{
+  "do_convert_rgb": null,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "SiglipImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "processor_class": "SiglipProcessor",
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}

trainer_state.json

@@ -0,0 +1,271 @@
+{
+  "best_global_step": 306,
+  "best_metric": 0.9845288326300985,
+  "best_model_checkpoint": "siglip2-safety-classifier-gpu/checkpoint-306",
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 306,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.09803921568627451,
+      "grad_norm": 17.626625061035156,
+      "learning_rate": 2.5e-06,
+      "loss": 1.1335,
+      "step": 10
+    },
+    {
+      "epoch": 0.19607843137254902,
+      "grad_norm": 10.657720565795898,
+      "learning_rate": 5.2777777777777785e-06,
+      "loss": 0.7086,
+      "step": 20
+    },
+    {
+      "epoch": 0.29411764705882354,
+      "grad_norm": 6.507628440856934,
+      "learning_rate": 8.055555555555557e-06,
+      "loss": 0.2973,
+      "step": 30
+    },
+    {
+      "epoch": 0.39215686274509803,
+      "grad_norm": 19.603195190429688,
+      "learning_rate": 1.0833333333333334e-05,
+      "loss": 0.1598,
+      "step": 40
+    },
+    {
+      "epoch": 0.49019607843137253,
+      "grad_norm": 4.1803460121154785,
+      "learning_rate": 1.3611111111111113e-05,
+      "loss": 0.1714,
+      "step": 50
+    },
+    {
+      "epoch": 0.5882352941176471,
+      "grad_norm": 25.094463348388672,
+      "learning_rate": 1.638888888888889e-05,
+      "loss": 0.1481,
+      "step": 60
+    },
+    {
+      "epoch": 0.6862745098039216,
+      "grad_norm": 7.0236897468566895,
+      "learning_rate": 1.916666666666667e-05,
+      "loss": 0.1073,
+      "step": 70
+    },
+    {
+      "epoch": 0.7843137254901961,
+      "grad_norm": 11.48787784576416,
+      "learning_rate": 1.9781931464174458e-05,
+      "loss": 0.0816,
+      "step": 80
+    },
+    {
+      "epoch": 0.8823529411764706,
+      "grad_norm": 13.502949714660645,
+      "learning_rate": 1.947040498442368e-05,
+      "loss": 0.1205,
+      "step": 90
+    },
+    {
+      "epoch": 0.9803921568627451,
+      "grad_norm": 13.85872745513916,
+      "learning_rate": 1.91588785046729e-05,
+      "loss": 0.1086,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9704641350210971,
+      "eval_loss": 0.08171376585960388,
+      "eval_runtime": 483.9596,
+      "eval_samples_per_second": 1.469,
+      "eval_steps_per_second": 0.012,
+      "step": 102
+    },
+    {
+      "epoch": 1.0784313725490196,
+      "grad_norm": 9.396234512329102,
+      "learning_rate": 1.8847352024922122e-05,
+      "loss": 0.0546,
+      "step": 110
+    },
+    {
+      "epoch": 1.1764705882352942,
+      "grad_norm": 3.7309114933013916,
+      "learning_rate": 1.853582554517134e-05,
+      "loss": 0.0574,
+      "step": 120
+    },
+    {
+      "epoch": 1.2745098039215685,
+      "grad_norm": 17.068557739257812,
+      "learning_rate": 1.8224299065420564e-05,
+      "loss": 0.0554,
+      "step": 130
+    },
+    {
+      "epoch": 1.3725490196078431,
+      "grad_norm": 3.3367063999176025,
+      "learning_rate": 1.7912772585669783e-05,
+      "loss": 0.0645,
+      "step": 140
+    },
+    {
+      "epoch": 1.4705882352941178,
+      "grad_norm": 10.730944633483887,
+      "learning_rate": 1.7601246105919006e-05,
+      "loss": 0.0565,
+      "step": 150
+    },
+    {
+      "epoch": 1.5686274509803921,
+      "grad_norm": 14.649934768676758,
+      "learning_rate": 1.7289719626168225e-05,
+      "loss": 0.083,
+      "step": 160
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 14.909839630126953,
+      "learning_rate": 1.6978193146417448e-05,
+      "loss": 0.0385,
+      "step": 170
+    },
+    {
+      "epoch": 1.7647058823529411,
+      "grad_norm": 3.265110731124878,
+      "learning_rate": 1.6666666666666667e-05,
+      "loss": 0.1152,
+      "step": 180
+    },
+    {
+      "epoch": 1.8627450980392157,
+      "grad_norm": 7.800755023956299,
+      "learning_rate": 1.635514018691589e-05,
+      "loss": 0.0867,
+      "step": 190
+    },
+    {
+      "epoch": 1.9607843137254903,
+      "grad_norm": 12.775182723999023,
+      "learning_rate": 1.604361370716511e-05,
+      "loss": 0.0415,
+      "step": 200
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9549929676511955,
+      "eval_loss": 0.12330988794565201,
+      "eval_runtime": 14.4147,
+      "eval_samples_per_second": 49.325,
+      "eval_steps_per_second": 0.416,
+      "step": 204
+    },
+    {
+      "epoch": 2.0588235294117645,
+      "grad_norm": 18.324607849121094,
+      "learning_rate": 1.5732087227414332e-05,
+      "loss": 0.0649,
+      "step": 210
+    },
+    {
+      "epoch": 2.156862745098039,
+      "grad_norm": 1.119167685508728,
+      "learning_rate": 1.5420560747663555e-05,
+      "loss": 0.0288,
+      "step": 220
+    },
+    {
+      "epoch": 2.2549019607843137,
+      "grad_norm": 13.742692947387695,
+      "learning_rate": 1.5109034267912774e-05,
+      "loss": 0.0569,
+      "step": 230
+    },
+    {
+      "epoch": 2.3529411764705883,
+      "grad_norm": 1.8996661901474,
+      "learning_rate": 1.4797507788161995e-05,
+      "loss": 0.0071,
+      "step": 240
+    },
+    {
+      "epoch": 2.450980392156863,
+      "grad_norm": 13.215072631835938,
+      "learning_rate": 1.4485981308411216e-05,
+      "loss": 0.0294,
+      "step": 250
+    },
+    {
+      "epoch": 2.549019607843137,
+      "grad_norm": 32.19113540649414,
+      "learning_rate": 1.4174454828660437e-05,
+      "loss": 0.0217,
+      "step": 260
+    },
+    {
+      "epoch": 2.6470588235294117,
+      "grad_norm": 7.469207286834717,
+      "learning_rate": 1.3862928348909658e-05,
+      "loss": 0.0194,
+      "step": 270
+    },
+    {
+      "epoch": 2.7450980392156863,
+      "grad_norm": 2.948065996170044,
+      "learning_rate": 1.355140186915888e-05,
+      "loss": 0.017,
+      "step": 280
+    },
+    {
+      "epoch": 2.843137254901961,
+      "grad_norm": 0.15147125720977783,
+      "learning_rate": 1.3239875389408102e-05,
+      "loss": 0.0186,
+      "step": 290
+    },
+    {
+      "epoch": 2.9411764705882355,
+      "grad_norm": 0.9559997320175171,
+      "learning_rate": 1.2928348909657323e-05,
+      "loss": 0.0302,
+      "step": 300
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9845288326300985,
+      "eval_loss": 0.05156884714961052,
+      "eval_runtime": 14.9504,
+      "eval_samples_per_second": 47.557,
+      "eval_steps_per_second": 0.401,
+      "step": 306
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 714,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 7,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.6395274262481408e+18,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": null
+}