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:bd7ec14e27a16222acf0b78d41404007df36b059dd2bbc769aaafca07bdfc55d
+size 1711848436

checkpoint-6000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85fbdde82096ee138fe1737b43bed7668337af2ba039ee79823014ac67e5589b
+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:b414376c5fc3e40d23d8bc86633f1317793cc2e9aa4019b618112c2f042ca79f
+size 623

checkpoint-6000/trainer_state.json

@@ -0,0 +1,213 @@
+{
+  "best_metric": 0.6995787620544434,
+  "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l37-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": 8.437943458557129,
+      "learning_rate": 2.907885040530582e-07,
+      "loss": 0.3017,
+      "step": 500
+    },
+    {
+      "epoch": 0.09211495946941783,
+      "eval_loss": 1.164457082748413,
+      "eval_runtime": 75.2697,
+      "eval_samples_per_second": 16.036,
+      "eval_steps_per_second": 2.006,
+      "step": 500
+    },
+    {
+      "epoch": 0.18422991893883567,
+      "grad_norm": 14.740598678588867,
+      "learning_rate": 2.815770081061164e-07,
+      "loss": 0.2899,
+      "step": 1000
+    },
+    {
+      "epoch": 0.18422991893883567,
+      "eval_loss": 1.01627779006958,
+      "eval_runtime": 76.4934,
+      "eval_samples_per_second": 15.779,
+      "eval_steps_per_second": 1.974,
+      "step": 1000
+    },
+    {
+      "epoch": 0.2763448784082535,
+      "grad_norm": 0.17031528055667877,
+      "learning_rate": 2.723655121591746e-07,
+      "loss": 0.273,
+      "step": 1500
+    },
+    {
+      "epoch": 0.2763448784082535,
+      "eval_loss": 0.9395213723182678,
+      "eval_runtime": 76.6256,
+      "eval_samples_per_second": 15.752,
+      "eval_steps_per_second": 1.971,
+      "step": 1500
+    },
+    {
+      "epoch": 0.36845983787767134,
+      "grad_norm": 0.0030106802005320787,
+      "learning_rate": 2.6315401621223287e-07,
+      "loss": 0.1818,
+      "step": 2000
+    },
+    {
+      "epoch": 0.36845983787767134,
+      "eval_loss": 0.8942967057228088,
+      "eval_runtime": 76.8089,
+      "eval_samples_per_second": 15.714,
+      "eval_steps_per_second": 1.966,
+      "step": 2000
+    },
+    {
+      "epoch": 0.46057479734708917,
+      "grad_norm": 474.4410705566406,
+      "learning_rate": 2.539425202652911e-07,
+      "loss": 0.2094,
+      "step": 2500
+    },
+    {
+      "epoch": 0.46057479734708917,
+      "eval_loss": 0.8330363631248474,
+      "eval_runtime": 76.9939,
+      "eval_samples_per_second": 15.677,
+      "eval_steps_per_second": 1.961,
+      "step": 2500
+    },
+    {
+      "epoch": 0.552689756816507,
+      "grad_norm": 314.3533935546875,
+      "learning_rate": 2.447310243183493e-07,
+      "loss": 0.1949,
+      "step": 3000
+    },
+    {
+      "epoch": 0.552689756816507,
+      "eval_loss": 0.7885717749595642,
+      "eval_runtime": 76.7534,
+      "eval_samples_per_second": 15.726,
+      "eval_steps_per_second": 1.967,
+      "step": 3000
+    },
+    {
+      "epoch": 0.6448047162859248,
+      "grad_norm": 7.678028106689453,
+      "learning_rate": 2.3551952837140753e-07,
+      "loss": 0.1335,
+      "step": 3500
+    },
+    {
+      "epoch": 0.6448047162859248,
+      "eval_loss": 0.7657251358032227,
+      "eval_runtime": 76.791,
+      "eval_samples_per_second": 15.718,
+      "eval_steps_per_second": 1.966,
+      "step": 3500
+    },
+    {
+      "epoch": 0.7369196757553427,
+      "grad_norm": 30.121461868286133,
+      "learning_rate": 2.263080324244657e-07,
+      "loss": 0.1541,
+      "step": 4000
+    },
+    {
+      "epoch": 0.7369196757553427,
+      "eval_loss": 0.7473344802856445,
+      "eval_runtime": 76.929,
+      "eval_samples_per_second": 15.69,
+      "eval_steps_per_second": 1.963,
+      "step": 4000
+    },
+    {
+      "epoch": 0.8290346352247605,
+      "grad_norm": 654.2782592773438,
+      "learning_rate": 2.1709653647752394e-07,
+      "loss": 0.1267,
+      "step": 4500
+    },
+    {
+      "epoch": 0.8290346352247605,
+      "eval_loss": 0.7376419305801392,
+      "eval_runtime": 75.2838,
+      "eval_samples_per_second": 16.033,
+      "eval_steps_per_second": 2.006,
+      "step": 4500
+    },
+    {
+      "epoch": 0.9211495946941783,
+      "grad_norm": 0.00021513362298719585,
+      "learning_rate": 2.0788504053058218e-07,
+      "loss": 0.1404,
+      "step": 5000
+    },
+    {
+      "epoch": 0.9211495946941783,
+      "eval_loss": 0.724165678024292,
+      "eval_runtime": 76.8393,
+      "eval_samples_per_second": 15.708,
+      "eval_steps_per_second": 1.965,
+      "step": 5000
+    },
+    {
+      "epoch": 1.013264554163596,
+      "grad_norm": 0.0016925306990742683,
+      "learning_rate": 1.9867354458364036e-07,
+      "loss": 0.1197,
+      "step": 5500
+    },
+    {
+      "epoch": 1.013264554163596,
+      "eval_loss": 0.7046866416931152,
+      "eval_runtime": 76.6635,
+      "eval_samples_per_second": 15.744,
+      "eval_steps_per_second": 1.97,
+      "step": 5500
+    },
+    {
+      "epoch": 1.105379513633014,
+      "grad_norm": 0.38834691047668457,
+      "learning_rate": 1.894620486366986e-07,
+      "loss": 0.0834,
+      "step": 6000
+    },
+    {
+      "epoch": 1.105379513633014,
+      "eval_loss": 0.6995787620544434,
+      "eval_runtime": 77.9523,
+      "eval_samples_per_second": 15.484,
+      "eval_steps_per_second": 1.937,
+      "step": 6000
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 16284,
+  "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": 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 sha256:76f7dfc201ccc0039551250d06ef1282a727ca0a51016f874c2eff3c580dbf8e
+size 4847