Training in progress, step 4000, checkpoint

checkpoint-4000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:026d4481f383b2a87f0cb94045f8baa817fb0a55bc69f2bc32be28a8999436a4
+oid sha256:a878d08f72540c9f3479285cc88310f0cdf60138a32db2e5150d64c4a5eccf34
 size 1711848436

checkpoint-4000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1b88defc3a89c4d8c55cd6ca0e38e0203ec283b45b98ce7709c5e93dd07789b
+oid sha256:a10dc97051a53b281e997547623dcdb90737be23513b9e3e7c89eaf1b21ab058
 size 3424043887

checkpoint-4000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f8610ffd95364d971241bbea36f29abb2299a046b81442005be36cebcbe4f45b
+oid sha256:b0cc521a4ce9508ca76ecef33ab44295ea7b55ac389a0be2ed51baa6f934287c
 size 623

checkpoint-4000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.7473344802856445,
+  "best_metric": 0.813768208026886,
   "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l37-l/checkpoint-4000",
-  "epoch": 0.7369196757553427,
+  "epoch": 0.4502983226387482,
   "eval_steps": 500,
   "global_step": 4000,
   "is_hyper_param_search": false,
@@ -9,128 +9,128 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.09211495946941783,
-      "grad_norm": 8.437943458557129,
-      "learning_rate": 2.907885040530582e-07,
-      "loss": 0.3017,
+      "epoch": 0.056287290329843524,
+      "grad_norm": 96.92656707763672,
+      "learning_rate": 2.9437127096701565e-07,
+      "loss": 0.3748,
       "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,
+      "epoch": 0.056287290329843524,
+      "eval_loss": 1.2226382493972778,
+      "eval_runtime": 131.2779,
+      "eval_samples_per_second": 15.037,
+      "eval_steps_per_second": 1.882,
       "step": 500
     },
     {
-      "epoch": 0.18422991893883567,
-      "grad_norm": 14.740598678588867,
-      "learning_rate": 2.815770081061164e-07,
-      "loss": 0.2899,
+      "epoch": 0.11257458065968705,
+      "grad_norm": 402.31158447265625,
+      "learning_rate": 2.887425419340313e-07,
+      "loss": 0.3057,
       "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,
+      "epoch": 0.11257458065968705,
+      "eval_loss": 1.0702259540557861,
+      "eval_runtime": 131.0109,
+      "eval_samples_per_second": 15.067,
+      "eval_steps_per_second": 1.885,
       "step": 1000
     },
     {
-      "epoch": 0.2763448784082535,
-      "grad_norm": 0.17031528055667877,
-      "learning_rate": 2.723655121591746e-07,
-      "loss": 0.273,
+      "epoch": 0.16886187098953057,
+      "grad_norm": 515.0911865234375,
+      "learning_rate": 2.831138129010469e-07,
+      "loss": 0.2239,
       "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,
+      "epoch": 0.16886187098953057,
+      "eval_loss": 0.9957238435745239,
+      "eval_runtime": 134.8377,
+      "eval_samples_per_second": 14.64,
+      "eval_steps_per_second": 1.832,
       "step": 1500
     },
     {
-      "epoch": 0.36845983787767134,
-      "grad_norm": 0.0030106802005320787,
-      "learning_rate": 2.6315401621223287e-07,
-      "loss": 0.1818,
+      "epoch": 0.2251491613193741,
+      "grad_norm": 3.1967105865478516,
+      "learning_rate": 2.774850838680626e-07,
+      "loss": 0.2229,
       "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,
+      "epoch": 0.2251491613193741,
+      "eval_loss": 0.9504629373550415,
+      "eval_runtime": 129.0169,
+      "eval_samples_per_second": 15.3,
+      "eval_steps_per_second": 1.914,
       "step": 2000
     },
     {
-      "epoch": 0.46057479734708917,
-      "grad_norm": 474.4410705566406,
-      "learning_rate": 2.539425202652911e-07,
-      "loss": 0.2094,
+      "epoch": 0.2814364516492176,
+      "grad_norm": 276.22998046875,
+      "learning_rate": 2.718563548350782e-07,
+      "loss": 0.2098,
       "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,
+      "epoch": 0.2814364516492176,
+      "eval_loss": 0.9006705284118652,
+      "eval_runtime": 129.1816,
+      "eval_samples_per_second": 15.281,
+      "eval_steps_per_second": 1.912,
       "step": 2500
     },
     {
-      "epoch": 0.552689756816507,
-      "grad_norm": 314.3533935546875,
-      "learning_rate": 2.447310243183493e-07,
-      "loss": 0.1949,
+      "epoch": 0.33772374197906113,
+      "grad_norm": 1105.317138671875,
+      "learning_rate": 2.6622762580209386e-07,
+      "loss": 0.1938,
       "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,
+      "epoch": 0.33772374197906113,
+      "eval_loss": 0.8782906532287598,
+      "eval_runtime": 128.6387,
+      "eval_samples_per_second": 15.345,
+      "eval_steps_per_second": 1.92,
       "step": 3000
     },
     {
-      "epoch": 0.6448047162859248,
-      "grad_norm": 7.678028106689453,
-      "learning_rate": 2.3551952837140753e-07,
-      "loss": 0.1335,
+      "epoch": 0.39401103230890466,
+      "grad_norm": 0.00043465400813147426,
+      "learning_rate": 2.605988967691095e-07,
+      "loss": 0.1688,
       "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,
+      "epoch": 0.39401103230890466,
+      "eval_loss": 0.8405746221542358,
+      "eval_runtime": 128.6989,
+      "eval_samples_per_second": 15.338,
+      "eval_steps_per_second": 1.919,
       "step": 3500
     },
     {
-      "epoch": 0.7369196757553427,
-      "grad_norm": 30.121461868286133,
-      "learning_rate": 2.263080324244657e-07,
-      "loss": 0.1541,
+      "epoch": 0.4502983226387482,
+      "grad_norm": 0.027426382526755333,
+      "learning_rate": 2.549701677361252e-07,
+      "loss": 0.1457,
       "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,
+      "epoch": 0.4502983226387482,
+      "eval_loss": 0.813768208026886,
+      "eval_runtime": 128.7209,
+      "eval_samples_per_second": 15.336,
+      "eval_steps_per_second": 1.919,
       "step": 4000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 16284,
+  "max_steps": 26649,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 500,

checkpoint-4000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76f7dfc201ccc0039551250d06ef1282a727ca0a51016f874c2eff3c580dbf8e
+oid sha256:074eb2b3aae1f3310fb0b27d83d259baa7ad74a32c8433c81af819e28558f6b0
 size 4847