Training in progress, step 6000, checkpoint

checkpoint-6000/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-6000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dfbe709c229ca2e1b8bc491502aa4ac9ed1e3b035d78ef2a870cb220da04455b
+size 1711848436

checkpoint-6000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c2744cf996c1f52ed2d7e5da0fdcb1f10cf00c8ada9818c771c104698ccaf8f
+size 3424043887

checkpoint-6000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2540cc73259311480694ca6863e114df4e2c18e837dc0f7839bdb7cbd64a19a0
+size 14503

checkpoint-6000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a77ac90b8cb52bbcf9e62fe287a067dcbc81a2751f20693a2e0376bfff0e56ad
+size 623

checkpoint-6000/trainer_state.json

@@ -0,0 +1,213 @@
+{
+  "best_metric": 0.6259325742721558,
+  "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l57-l/checkpoint-6000",
+  "epoch": 1.105379513633014,
+  "eval_steps": 500,
+  "global_step": 6000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.09211495946941783,
+      "grad_norm": 3.0293657779693604,
+      "learning_rate": 4.884856300663227e-07,
+      "loss": 0.271,
+      "step": 500
+    },
+    {
+      "epoch": 0.09211495946941783,
+      "eval_loss": 1.069305419921875,
+      "eval_runtime": 75.1409,
+      "eval_samples_per_second": 16.063,
+      "eval_steps_per_second": 2.01,
+      "step": 500
+    },
+    {
+      "epoch": 0.18422991893883567,
+      "grad_norm": 7.3040852546691895,
+      "learning_rate": 4.769712601326456e-07,
+      "loss": 0.2493,
+      "step": 1000
+    },
+    {
+      "epoch": 0.18422991893883567,
+      "eval_loss": 0.9427077770233154,
+      "eval_runtime": 74.8926,
+      "eval_samples_per_second": 16.116,
+      "eval_steps_per_second": 2.016,
+      "step": 1000
+    },
+    {
+      "epoch": 0.2763448784082535,
+      "grad_norm": 0.020736657083034515,
+      "learning_rate": 4.654568901989683e-07,
+      "loss": 0.2348,
+      "step": 1500
+    },
+    {
+      "epoch": 0.2763448784082535,
+      "eval_loss": 0.8726764917373657,
+      "eval_runtime": 76.6621,
+      "eval_samples_per_second": 15.744,
+      "eval_steps_per_second": 1.97,
+      "step": 1500
+    },
+    {
+      "epoch": 0.36845983787767134,
+      "grad_norm": 5.3738909628009424e-05,
+      "learning_rate": 4.5394252026529107e-07,
+      "loss": 0.1552,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36845983787767134,
+      "eval_loss": 0.8325753808021545,
+      "eval_runtime": 76.6127,
+      "eval_samples_per_second": 15.755,
+      "eval_steps_per_second": 1.971,
+      "step": 2000
+    },
+    {
+      "epoch": 0.46057479734708917,
+      "grad_norm": 404.510009765625,
+      "learning_rate": 4.4242815033161386e-07,
+      "loss": 0.1753,
+      "step": 2500
+    },
+    {
+      "epoch": 0.46057479734708917,
+      "eval_loss": 0.7550467848777771,
+      "eval_runtime": 76.8024,
+      "eval_samples_per_second": 15.716,
+      "eval_steps_per_second": 1.966,
+      "step": 2500
+    },
+    {
+      "epoch": 0.552689756816507,
+      "grad_norm": 370.7579650878906,
+      "learning_rate": 4.3091378039793665e-07,
+      "loss": 0.1659,
+      "step": 3000
+    },
+    {
+      "epoch": 0.552689756816507,
+      "eval_loss": 0.719167172908783,
+      "eval_runtime": 76.6327,
+      "eval_samples_per_second": 15.75,
+      "eval_steps_per_second": 1.97,
+      "step": 3000
+    },
+    {
+      "epoch": 0.6448047162859248,
+      "grad_norm": 24.97766876220703,
+      "learning_rate": 4.193994104642594e-07,
+      "loss": 0.105,
+      "step": 3500
+    },
+    {
+      "epoch": 0.6448047162859248,
+      "eval_loss": 0.7117515802383423,
+      "eval_runtime": 77.1286,
+      "eval_samples_per_second": 15.649,
+      "eval_steps_per_second": 1.958,
+      "step": 3500
+    },
+    {
+      "epoch": 0.7369196757553427,
+      "grad_norm": 305.62060546875,
+      "learning_rate": 4.0788504053058217e-07,
+      "loss": 0.1336,
+      "step": 4000
+    },
+    {
+      "epoch": 0.7369196757553427,
+      "eval_loss": 0.6952972412109375,
+      "eval_runtime": 76.7253,
+      "eval_samples_per_second": 15.731,
+      "eval_steps_per_second": 1.968,
+      "step": 4000
+    },
+    {
+      "epoch": 0.8290346352247605,
+      "grad_norm": 506.12237548828125,
+      "learning_rate": 3.9637067059690496e-07,
+      "loss": 0.1154,
+      "step": 4500
+    },
+    {
+      "epoch": 0.8290346352247605,
+      "eval_loss": 0.6744509935379028,
+      "eval_runtime": 76.9493,
+      "eval_samples_per_second": 15.686,
+      "eval_steps_per_second": 1.962,
+      "step": 4500
+    },
+    {
+      "epoch": 0.9211495946941783,
+      "grad_norm": 8.823301322991028e-05,
+      "learning_rate": 3.8485630066322774e-07,
+      "loss": 0.108,
+      "step": 5000
+    },
+    {
+      "epoch": 0.9211495946941783,
+      "eval_loss": 0.6559990048408508,
+      "eval_runtime": 75.2391,
+      "eval_samples_per_second": 16.042,
+      "eval_steps_per_second": 2.007,
+      "step": 5000
+    },
+    {
+      "epoch": 1.013264554163596,
+      "grad_norm": 0.0008560372516512871,
+      "learning_rate": 3.733419307295504e-07,
+      "loss": 0.1106,
+      "step": 5500
+    },
+    {
+      "epoch": 1.013264554163596,
+      "eval_loss": 0.6367496848106384,
+      "eval_runtime": 77.0031,
+      "eval_samples_per_second": 15.675,
+      "eval_steps_per_second": 1.961,
+      "step": 5500
+    },
+    {
+      "epoch": 1.105379513633014,
+      "grad_norm": 0.04510360211133957,
+      "learning_rate": 3.618275607958732e-07,
+      "loss": 0.0591,
+      "step": 6000
+    },
+    {
+      "epoch": 1.105379513633014,
+      "eval_loss": 0.6259325742721558,
+      "eval_runtime": 75.2787,
+      "eval_samples_per_second": 16.034,
+      "eval_steps_per_second": 2.006,
+      "step": 6000
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 21712,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "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": 2159269624262520.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-6000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf76416e4a857628175385db7f617e54451cd7ffb62c1f9590ee9f30268f8c5b
+size 4847