Training in progress, step 9000, checkpoint

checkpoint-9000/config.json

@@ -0,0 +1,31 @@
+{
+  "_name_or_path": "openai/clip-vit-large-patch14-336",
+  "architectures": [
+    "CLIPModel"
+  ],
+  "initializer_factor": 1.0,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "clip",
+  "projection_dim": 768,
+  "text_config": {
+    "dropout": 0.0,
+    "hidden_size": 768,
+    "intermediate_size": 3072,
+    "model_type": "clip_text_model",
+    "num_attention_heads": 12,
+    "projection_dim": 768
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.45.0.dev0",
+  "vision_config": {
+    "dropout": 0.0,
+    "hidden_size": 1024,
+    "image_size": 336,
+    "intermediate_size": 4096,
+    "model_type": "clip_vision_model",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 24,
+    "patch_size": 14,
+    "projection_dim": 768
+  }
+}

checkpoint-9000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b68bebe93da141b209edaa58eb102e39f26ca50bab6ff4a7a372f36008365b87
+size 1711848436

checkpoint-9000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15267a61d1eddcfd8dcf3d66aaa87739dae109df30cc9fbc2c2e34c092868dda
+size 3424043887

checkpoint-9000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75e7a66179ff03d27874bf0d0676cd79fe7137c908da603458cd4b183abc4a42
+size 14503

checkpoint-9000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8765ee4c25b6df1b35540345711ef246e43000d00ea66561f06aaf565727e0c2
+size 623

checkpoint-9000/trainer_state.json

@@ -0,0 +1,303 @@
+{
+  "best_metric": 1.9636509418487549,
+  "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l55-l/checkpoint-7000",
+  "epoch": 1.6577638607478358,
+  "eval_steps": 500,
+  "global_step": 9000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.09209799226376865,
+      "grad_norm": 0.009597355499863625,
+      "learning_rate": 4.846503346227052e-05,
+      "loss": 0.7336,
+      "step": 500
+    },
+    {
+      "epoch": 0.09209799226376865,
+      "eval_loss": 2.0780556201934814,
+      "eval_runtime": 77.1404,
+      "eval_samples_per_second": 15.634,
+      "eval_steps_per_second": 1.957,
+      "step": 500
+    },
+    {
+      "epoch": 0.1841959845275373,
+      "grad_norm": 0.10049372166395187,
+      "learning_rate": 4.693006692454105e-05,
+      "loss": 0.6953,
+      "step": 1000
+    },
+    {
+      "epoch": 0.1841959845275373,
+      "eval_loss": 2.0782744884490967,
+      "eval_runtime": 77.8965,
+      "eval_samples_per_second": 15.482,
+      "eval_steps_per_second": 1.938,
+      "step": 1000
+    },
+    {
+      "epoch": 0.27629397679130596,
+      "grad_norm": 0.0041180383414030075,
+      "learning_rate": 4.539510038681157e-05,
+      "loss": 0.6953,
+      "step": 1500
+    },
+    {
+      "epoch": 0.27629397679130596,
+      "eval_loss": 2.077699899673462,
+      "eval_runtime": 78.9586,
+      "eval_samples_per_second": 15.274,
+      "eval_steps_per_second": 1.912,
+      "step": 1500
+    },
+    {
+      "epoch": 0.3683919690550746,
+      "grad_norm": 0.29397013783454895,
+      "learning_rate": 4.386013384908209e-05,
+      "loss": 0.6932,
+      "step": 2000
+    },
+    {
+      "epoch": 0.3683919690550746,
+      "eval_loss": 2.0721774101257324,
+      "eval_runtime": 77.349,
+      "eval_samples_per_second": 15.592,
+      "eval_steps_per_second": 1.952,
+      "step": 2000
+    },
+    {
+      "epoch": 0.46048996131884323,
+      "grad_norm": 0.9136261343955994,
+      "learning_rate": 4.232516731135262e-05,
+      "loss": 0.6907,
+      "step": 2500
+    },
+    {
+      "epoch": 0.46048996131884323,
+      "eval_loss": 2.079211711883545,
+      "eval_runtime": 79.3204,
+      "eval_samples_per_second": 15.204,
+      "eval_steps_per_second": 1.904,
+      "step": 2500
+    },
+    {
+      "epoch": 0.5525879535826119,
+      "grad_norm": 4.484213829040527,
+      "learning_rate": 4.079020077362314e-05,
+      "loss": 0.7043,
+      "step": 3000
+    },
+    {
+      "epoch": 0.5525879535826119,
+      "eval_loss": 2.0204577445983887,
+      "eval_runtime": 78.612,
+      "eval_samples_per_second": 15.341,
+      "eval_steps_per_second": 1.921,
+      "step": 3000
+    },
+    {
+      "epoch": 0.6446859458463805,
+      "grad_norm": 1.1021015644073486,
+      "learning_rate": 3.9255234235893664e-05,
+      "loss": 0.7078,
+      "step": 3500
+    },
+    {
+      "epoch": 0.6446859458463805,
+      "eval_loss": 2.1256344318389893,
+      "eval_runtime": 78.63,
+      "eval_samples_per_second": 15.338,
+      "eval_steps_per_second": 1.92,
+      "step": 3500
+    },
+    {
+      "epoch": 0.7367839381101492,
+      "grad_norm": 16.864044189453125,
+      "learning_rate": 3.772026769816418e-05,
+      "loss": 0.7034,
+      "step": 4000
+    },
+    {
+      "epoch": 0.7367839381101492,
+      "eval_loss": 2.0000314712524414,
+      "eval_runtime": 78.6543,
+      "eval_samples_per_second": 15.333,
+      "eval_steps_per_second": 1.92,
+      "step": 4000
+    },
+    {
+      "epoch": 0.8288819303739179,
+      "grad_norm": 20.071943283081055,
+      "learning_rate": 3.6185301160434705e-05,
+      "loss": 0.6671,
+      "step": 4500
+    },
+    {
+      "epoch": 0.8288819303739179,
+      "eval_loss": 2.1607539653778076,
+      "eval_runtime": 78.6404,
+      "eval_samples_per_second": 15.336,
+      "eval_steps_per_second": 1.92,
+      "step": 4500
+    },
+    {
+      "epoch": 0.9209799226376865,
+      "grad_norm": 17.817607879638672,
+      "learning_rate": 3.4650334622705225e-05,
+      "loss": 0.6899,
+      "step": 5000
+    },
+    {
+      "epoch": 0.9209799226376865,
+      "eval_loss": 2.093959331512451,
+      "eval_runtime": 78.539,
+      "eval_samples_per_second": 15.355,
+      "eval_steps_per_second": 1.923,
+      "step": 5000
+    },
+    {
+      "epoch": 1.013077914901455,
+      "grad_norm": 41.0746955871582,
+      "learning_rate": 3.311536808497575e-05,
+      "loss": 0.6649,
+      "step": 5500
+    },
+    {
+      "epoch": 1.013077914901455,
+      "eval_loss": 1.9776026010513306,
+      "eval_runtime": 78.5095,
+      "eval_samples_per_second": 15.361,
+      "eval_steps_per_second": 1.923,
+      "step": 5500
+    },
+    {
+      "epoch": 1.1051759071652238,
+      "grad_norm": 35.426021575927734,
+      "learning_rate": 3.158040154724627e-05,
+      "loss": 0.6748,
+      "step": 6000
+    },
+    {
+      "epoch": 1.1051759071652238,
+      "eval_loss": 1.9851746559143066,
+      "eval_runtime": 78.5944,
+      "eval_samples_per_second": 15.345,
+      "eval_steps_per_second": 1.921,
+      "step": 6000
+    },
+    {
+      "epoch": 1.1972738994289924,
+      "grad_norm": 0.8773899078369141,
+      "learning_rate": 3.0045435009516792e-05,
+      "loss": 0.6456,
+      "step": 6500
+    },
+    {
+      "epoch": 1.1972738994289924,
+      "eval_loss": 1.9981426000595093,
+      "eval_runtime": 78.4741,
+      "eval_samples_per_second": 15.368,
+      "eval_steps_per_second": 1.924,
+      "step": 6500
+    },
+    {
+      "epoch": 1.289371891692761,
+      "grad_norm": 19.7215518951416,
+      "learning_rate": 2.8510468471787316e-05,
+      "loss": 0.6514,
+      "step": 7000
+    },
+    {
+      "epoch": 1.289371891692761,
+      "eval_loss": 1.9636509418487549,
+      "eval_runtime": 78.4704,
+      "eval_samples_per_second": 15.369,
+      "eval_steps_per_second": 1.924,
+      "step": 7000
+    },
+    {
+      "epoch": 1.3814698839565298,
+      "grad_norm": 5.96924352645874,
+      "learning_rate": 2.697550193405784e-05,
+      "loss": 0.6179,
+      "step": 7500
+    },
+    {
+      "epoch": 1.3814698839565298,
+      "eval_loss": 2.0034029483795166,
+      "eval_runtime": 78.4771,
+      "eval_samples_per_second": 15.368,
+      "eval_steps_per_second": 1.924,
+      "step": 7500
+    },
+    {
+      "epoch": 1.4735678762202984,
+      "grad_norm": 1.2145048379898071,
+      "learning_rate": 2.5440535396328363e-05,
+      "loss": 0.6561,
+      "step": 8000
+    },
+    {
+      "epoch": 1.4735678762202984,
+      "eval_loss": 2.2864511013031006,
+      "eval_runtime": 78.3606,
+      "eval_samples_per_second": 15.39,
+      "eval_steps_per_second": 1.927,
+      "step": 8000
+    },
+    {
+      "epoch": 1.565665868484067,
+      "grad_norm": 15.816754341125488,
+      "learning_rate": 2.3905568858598883e-05,
+      "loss": 0.6328,
+      "step": 8500
+    },
+    {
+      "epoch": 1.565665868484067,
+      "eval_loss": 2.4807703495025635,
+      "eval_runtime": 77.0504,
+      "eval_samples_per_second": 15.652,
+      "eval_steps_per_second": 1.96,
+      "step": 8500
+    },
+    {
+      "epoch": 1.6577638607478358,
+      "grad_norm": 18.753801345825195,
+      "learning_rate": 2.2370602320869404e-05,
+      "loss": 0.7053,
+      "step": 9000
+    },
+    {
+      "epoch": 1.6577638607478358,
+      "eval_loss": 2.1609902381896973,
+      "eval_runtime": 76.5781,
+      "eval_samples_per_second": 15.749,
+      "eval_steps_per_second": 1.972,
+      "step": 9000
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 16287,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "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": 3238364798866890.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-9000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9885fd590ddfbffc4f8db1738c1993002dfa8c1603f7992a1cda9b93bbe7d34
+size 4847