Upload folder using huggingface_hub

checkpoint-424/config.json

@@ -0,0 +1,76 @@
+{
+  "architectures": [
+    "SiglipForImageClassification"
+  ],
+  "id2label": {
+    "0": "\"rotate anticlockwise\"",
+    "1": "\"increase\"",
+    "2": "\"release\"",
+    "3": "\"switch\"",
+    "4": "\"look up\"",
+    "5": "\"Terminate\"",
+    "6": "\"decrease\"",
+    "7": "\"move backward\"",
+    "8": "\"point\"",
+    "9": "\"rotate clockwise\"",
+    "10": "\"grasp\"",
+    "11": "\"pause\"",
+    "12": "\"move forward\"",
+    "13": "\"Confirm\"",
+    "14": "\"look down\"",
+    "15": "\"move left\"",
+    "16": "\"move right\""
+  },
+  "initializer_factor": 1.0,
+  "label2id": {
+    "\"Confirm\"": 13,
+    "\"Terminate\"": 5,
+    "\"decrease\"": 6,
+    "\"grasp\"": 10,
+    "\"increase\"": 1,
+    "\"look down\"": 14,
+    "\"look up\"": 4,
+    "\"move backward\"": 7,
+    "\"move forward\"": 12,
+    "\"move left\"": 15,
+    "\"move right\"": 16,
+    "\"pause\"": 11,
+    "\"point\"": 8,
+    "\"release\"": 2,
+    "\"rotate anticlockwise\"": 0,
+    "\"rotate clockwise\"": 9,
+    "\"switch\"": 3
+  },
+  "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": "4.50.2",
+  "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"
+  }
+}

checkpoint-424/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4729a0076ada1761a70c391852d57b178b7285fcb2f9c9a745abad551a010165
+size 371614140

checkpoint-424/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c21385e88025dc9a3f53bd17d22c39e0ce03894667f427cd84d0b47ceed4f0c
+size 686648058

checkpoint-424/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
+  }
+}

checkpoint-424/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b071a4e93eac6df7b9d7f06a6295288a51375a6d4ad3ba4db6708d90637ec5b
+size 14244

checkpoint-424/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dfe9450ff9b209da1d2083358044a50b1ecdea59cb17053db1e56ba52e96fb58
+size 1064

checkpoint-424/trainer_state.json

@@ -0,0 +1,44 @@
+{
+  "best_global_step": 424,
+  "best_metric": 0.13456620275974274,
+  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-424",
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 424,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9571280991735537,
+      "eval_loss": 0.13456620275974274,
+      "eval_model_preparation_time": 0.0024,
+      "eval_runtime": 175.2748,
+      "eval_samples_per_second": 77.319,
+      "eval_steps_per_second": 9.665,
+      "step": 424
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 848,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "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": 1.13519566468753e+18,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-424/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:035d5f50831a3a5d7cd09a166c9fa2c85bfdb9d0d1888ae17611b61aaa610ef2
+size 5304

checkpoint-848/config.json

@@ -0,0 +1,76 @@
+{
+  "architectures": [
+    "SiglipForImageClassification"
+  ],
+  "id2label": {
+    "0": "\"rotate anticlockwise\"",
+    "1": "\"increase\"",
+    "2": "\"release\"",
+    "3": "\"switch\"",
+    "4": "\"look up\"",
+    "5": "\"Terminate\"",
+    "6": "\"decrease\"",
+    "7": "\"move backward\"",
+    "8": "\"point\"",
+    "9": "\"rotate clockwise\"",
+    "10": "\"grasp\"",
+    "11": "\"pause\"",
+    "12": "\"move forward\"",
+    "13": "\"Confirm\"",
+    "14": "\"look down\"",
+    "15": "\"move left\"",
+    "16": "\"move right\""
+  },
+  "initializer_factor": 1.0,
+  "label2id": {
+    "\"Confirm\"": 13,
+    "\"Terminate\"": 5,
+    "\"decrease\"": 6,
+    "\"grasp\"": 10,
+    "\"increase\"": 1,
+    "\"look down\"": 14,
+    "\"look up\"": 4,
+    "\"move backward\"": 7,
+    "\"move forward\"": 12,
+    "\"move left\"": 15,
+    "\"move right\"": 16,
+    "\"pause\"": 11,
+    "\"point\"": 8,
+    "\"release\"": 2,
+    "\"rotate anticlockwise\"": 0,
+    "\"rotate clockwise\"": 9,
+    "\"switch\"": 3
+  },
+  "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": "4.50.2",
+  "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"
+  }
+}

checkpoint-848/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd7d84c5c9e7b32a364ac4afa1f6da0d52a782fa4acda3c01864313659c66085
+size 371614140

checkpoint-848/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:739aa88a80644fd55af301b01bfd40d158ff89ec67f09017a3c3d2c9961bad9c
+size 686648058

checkpoint-848/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
+  }
+}

checkpoint-848/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3bffe80938eb5850f672dbee35db9b3a3ce42dad73daf9d86b718f1f872daaf7
+size 14244

checkpoint-848/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2680d5ac4589c4d9a77dae4657939c0c0e5f79f4344a07077c81721689ff07aa
+size 1064

checkpoint-848/trainer_state.json

@@ -0,0 +1,61 @@
+{
+  "best_global_step": 848,
+  "best_metric": 0.008921321481466293,
+  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-848",
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 848,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9571280991735537,
+      "eval_loss": 0.13456620275974274,
+      "eval_model_preparation_time": 0.0024,
+      "eval_runtime": 175.2748,
+      "eval_samples_per_second": 77.319,
+      "eval_steps_per_second": 9.665,
+      "step": 424
+    },
+    {
+      "epoch": 1.179245283018868,
+      "grad_norm": 5.478814125061035,
+      "learning_rate": 8.721804511278195e-05,
+      "loss": 0.5777,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9971959858323495,
+      "eval_loss": 0.008921321481466293,
+      "eval_model_preparation_time": 0.0024,
+      "eval_runtime": 175.3207,
+      "eval_samples_per_second": 77.298,
+      "eval_steps_per_second": 9.662,
+      "step": 848
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 848,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.27039132937506e+18,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-848/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:035d5f50831a3a5d7cd09a166c9fa2c85bfdb9d0d1888ae17611b61aaa610ef2
+size 5304

config.json

@@ -0,0 +1,76 @@
+{
+  "architectures": [
+    "SiglipForImageClassification"
+  ],
+  "id2label": {
+    "0": "\"rotate anticlockwise\"",
+    "1": "\"increase\"",
+    "2": "\"release\"",
+    "3": "\"switch\"",
+    "4": "\"look up\"",
+    "5": "\"Terminate\"",
+    "6": "\"decrease\"",
+    "7": "\"move backward\"",
+    "8": "\"point\"",
+    "9": "\"rotate clockwise\"",
+    "10": "\"grasp\"",
+    "11": "\"pause\"",
+    "12": "\"move forward\"",
+    "13": "\"Confirm\"",
+    "14": "\"look down\"",
+    "15": "\"move left\"",
+    "16": "\"move right\""
+  },
+  "initializer_factor": 1.0,
+  "label2id": {
+    "\"Confirm\"": 13,
+    "\"Terminate\"": 5,
+    "\"decrease\"": 6,
+    "\"grasp\"": 10,
+    "\"increase\"": 1,
+    "\"look down\"": 14,
+    "\"look up\"": 4,
+    "\"move backward\"": 7,
+    "\"move forward\"": 12,
+    "\"move left\"": 15,
+    "\"move right\"": 16,
+    "\"pause\"": 11,
+    "\"point\"": 8,
+    "\"release\"": 2,
+    "\"rotate anticlockwise\"": 0,
+    "\"rotate clockwise\"": 9,
+    "\"switch\"": 3
+  },
+  "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": "4.50.2",
+  "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"
+  }
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd7d84c5c9e7b32a364ac4afa1f6da0d52a782fa4acda3c01864313659c66085
+size 371614140

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
+  }
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:035d5f50831a3a5d7cd09a166c9fa2c85bfdb9d0d1888ae17611b61aaa610ef2
+size 5304