Training in progress, step 4000, checkpoint

checkpoint-4000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d32326140983fe06df8a4feebe9254db82fddb9e5b81e5fc2852cf4950ae893e
+oid sha256:79156abefde4e4882460600714ae0567406deac99590a0ff234ba1048e2f2339
 size 1711848436

checkpoint-4000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:537ddc2ec13ff07e50acd17af98e66d697c093e9931cb5f6975b703e1c8246e7
+oid sha256:cc99807d330457613b7b4fbcc2866880b518f725723ad1f96fed56361fa06c7f
 size 3424043887

checkpoint-4000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:90a1b90840e0ce94ab1e7091ab856cfbbbce2e497912ae668cbbf9a86e863b92
+oid sha256:3fede1d7806deae69687da3edc51ede9ae3128bb8735c897888c46d7a052e8dd
 size 623

checkpoint-4000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.6952972412109375,
+  "best_metric": 0.7490401864051819,
   "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l57-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": 3.0293657779693604,
-      "learning_rate": 4.884856300663227e-07,
-      "loss": 0.271,
+      "epoch": 0.056287290329843524,
+      "grad_norm": 150.1642303466797,
+      "learning_rate": 4.929640887087696e-07,
+      "loss": 0.3486,
       "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,
+      "epoch": 0.056287290329843524,
+      "eval_loss": 1.1265727281570435,
+      "eval_runtime": 123.7806,
+      "eval_samples_per_second": 15.948,
+      "eval_steps_per_second": 1.995,
       "step": 500
     },
     {
-      "epoch": 0.18422991893883567,
-      "grad_norm": 7.3040852546691895,
-      "learning_rate": 4.769712601326456e-07,
-      "loss": 0.2493,
+      "epoch": 0.11257458065968705,
+      "grad_norm": 366.6913146972656,
+      "learning_rate": 4.859281774175391e-07,
+      "loss": 0.2733,
       "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,
+      "epoch": 0.11257458065968705,
+      "eval_loss": 0.9742079377174377,
+      "eval_runtime": 125.8047,
+      "eval_samples_per_second": 15.691,
+      "eval_steps_per_second": 1.963,
       "step": 1000
     },
     {
-      "epoch": 0.2763448784082535,
-      "grad_norm": 0.020736657083034515,
-      "learning_rate": 4.654568901989683e-07,
-      "loss": 0.2348,
+      "epoch": 0.16886187098953057,
+      "grad_norm": 555.5016479492188,
+      "learning_rate": 4.788922661263087e-07,
+      "loss": 0.1851,
       "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,
+      "epoch": 0.16886187098953057,
+      "eval_loss": 0.9162159562110901,
+      "eval_runtime": 126.7606,
+      "eval_samples_per_second": 15.573,
+      "eval_steps_per_second": 1.949,
       "step": 1500
     },
     {
-      "epoch": 0.36845983787767134,
-      "grad_norm": 5.3738909628009424e-05,
-      "learning_rate": 4.5394252026529107e-07,
-      "loss": 0.1552,
+      "epoch": 0.2251491613193741,
+      "grad_norm": 0.423260897397995,
+      "learning_rate": 4.7185635483507824e-07,
+      "loss": 0.1973,
       "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,
+      "epoch": 0.2251491613193741,
+      "eval_loss": 0.8716973662376404,
+      "eval_runtime": 127.6955,
+      "eval_samples_per_second": 15.459,
+      "eval_steps_per_second": 1.934,
       "step": 2000
     },
     {
-      "epoch": 0.46057479734708917,
-      "grad_norm": 404.510009765625,
-      "learning_rate": 4.4242815033161386e-07,
-      "loss": 0.1753,
+      "epoch": 0.2814364516492176,
+      "grad_norm": 6.840139389038086,
+      "learning_rate": 4.6482044354384774e-07,
+      "loss": 0.1881,
       "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,
+      "epoch": 0.2814364516492176,
+      "eval_loss": 0.830047070980072,
+      "eval_runtime": 128.0771,
+      "eval_samples_per_second": 15.413,
+      "eval_steps_per_second": 1.929,
       "step": 2500
     },
     {
-      "epoch": 0.552689756816507,
-      "grad_norm": 370.7579650878906,
-      "learning_rate": 4.3091378039793665e-07,
-      "loss": 0.1659,
+      "epoch": 0.33772374197906113,
+      "grad_norm": 729.8377685546875,
+      "learning_rate": 4.577845322526173e-07,
+      "loss": 0.1695,
       "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,
+      "epoch": 0.33772374197906113,
+      "eval_loss": 0.8265627026557922,
+      "eval_runtime": 128.052,
+      "eval_samples_per_second": 15.416,
+      "eval_steps_per_second": 1.929,
       "step": 3000
     },
     {
-      "epoch": 0.6448047162859248,
-      "grad_norm": 24.97766876220703,
-      "learning_rate": 4.193994104642594e-07,
-      "loss": 0.105,
+      "epoch": 0.39401103230890466,
+      "grad_norm": 7.96999938756926e-06,
+      "learning_rate": 4.507486209613869e-07,
+      "loss": 0.155,
       "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,
+      "epoch": 0.39401103230890466,
+      "eval_loss": 0.7721038460731506,
+      "eval_runtime": 128.1871,
+      "eval_samples_per_second": 15.399,
+      "eval_steps_per_second": 1.927,
       "step": 3500
     },
     {
-      "epoch": 0.7369196757553427,
-      "grad_norm": 305.62060546875,
-      "learning_rate": 4.0788504053058217e-07,
-      "loss": 0.1336,
+      "epoch": 0.4502983226387482,
+      "grad_norm": 0.006555848754942417,
+      "learning_rate": 4.4371270967015645e-07,
+      "loss": 0.1216,
       "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,
+      "epoch": 0.4502983226387482,
+      "eval_loss": 0.7490401864051819,
+      "eval_runtime": 127.5516,
+      "eval_samples_per_second": 15.476,
+      "eval_steps_per_second": 1.936,
       "step": 4000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 21712,
+  "max_steps": 35532,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 4,
   "save_steps": 500,

checkpoint-4000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bf76416e4a857628175385db7f617e54451cd7ffb62c1f9590ee9f30268f8c5b
+oid sha256:319be2a769ef7117b6826908976f080f9082f159f4f0ad75da1a151ad29d9bff
 size 4847