Training in progress, step 3500, checkpoint

checkpoint-3500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c99719a65b3933ff2f362dce875cdb39b09cc43b0e7d48a3672fb6e6f0f96582
+oid sha256:7e1418f340b77a6578e47214167f8270be129a69245ef30a7334227311c3d154
 size 1711848436

checkpoint-3500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ff959322597cd03d232df8ffa0fa0fdbb08d0298e00a42eee9916d036060fe31
+oid sha256:5d6f3257ab9775175b20fe85ff80542e02bb4040022d5aa0660c19d06271f059
 size 3424043887

checkpoint-3500/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:212bf18dc9b5b9400c154e7e5a9d4502bea517192b87703ec9f4c9ddfcd462dd
+oid sha256:8e07a7c95ddde9c2f0c42405698dc83c683c724f799722e57a5f46ad26044562
 size 623

checkpoint-3500/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 1.1233855485916138,
+  "best_metric": 1.1988952159881592,
   "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l58-l/checkpoint-3500",
-  "epoch": 0.6448047162859248,
+  "epoch": 0.37329351535836175,
   "eval_steps": 500,
   "global_step": 3500,
   "is_hyper_param_search": false,
@@ -9,113 +9,113 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.09211495946941783,
-      "grad_norm": 19.808820724487305,
-      "learning_rate": 4.84647506755097e-08,
-      "loss": 0.3758,
+      "epoch": 0.05332764505119454,
+      "grad_norm": 217.40109252929688,
+      "learning_rate": 4.911120591581342e-08,
+      "loss": 0.4667,
       "step": 500
     },
     {
-      "epoch": 0.09211495946941783,
-      "eval_loss": 1.418845295906067,
-      "eval_runtime": 75.2377,
-      "eval_samples_per_second": 16.042,
-      "eval_steps_per_second": 2.007,
+      "epoch": 0.05332764505119454,
+      "eval_loss": 1.4427576065063477,
+      "eval_runtime": 60.328,
+      "eval_samples_per_second": 16.361,
+      "eval_steps_per_second": 2.055,
       "step": 500
     },
     {
-      "epoch": 0.18422991893883567,
-      "grad_norm": 76.37991333007812,
-      "learning_rate": 4.69295013510194e-08,
-      "loss": 0.4108,
+      "epoch": 0.10665529010238908,
+      "grad_norm": 70.38941955566406,
+      "learning_rate": 4.822241183162685e-08,
+      "loss": 0.4534,
       "step": 1000
     },
     {
-      "epoch": 0.18422991893883567,
-      "eval_loss": 1.3513683080673218,
-      "eval_runtime": 74.9686,
-      "eval_samples_per_second": 16.1,
-      "eval_steps_per_second": 2.014,
+      "epoch": 0.10665529010238908,
+      "eval_loss": 1.3823094367980957,
+      "eval_runtime": 61.4884,
+      "eval_samples_per_second": 16.052,
+      "eval_steps_per_second": 2.017,
       "step": 1000
     },
     {
-      "epoch": 0.2763448784082535,
-      "grad_norm": 0.5795987844467163,
-      "learning_rate": 4.539425202652911e-08,
-      "loss": 0.4335,
+      "epoch": 0.1599829351535836,
+      "grad_norm": 389.9967956542969,
+      "learning_rate": 4.733361774744027e-08,
+      "loss": 0.3753,
       "step": 1500
     },
     {
-      "epoch": 0.2763448784082535,
-      "eval_loss": 1.2895005941390991,
-      "eval_runtime": 76.713,
-      "eval_samples_per_second": 15.734,
-      "eval_steps_per_second": 1.968,
+      "epoch": 0.1599829351535836,
+      "eval_loss": 1.3328827619552612,
+      "eval_runtime": 63.0378,
+      "eval_samples_per_second": 15.657,
+      "eval_steps_per_second": 1.967,
       "step": 1500
     },
     {
-      "epoch": 0.36845983787767134,
-      "grad_norm": 0.1463615745306015,
-      "learning_rate": 4.385900270203881e-08,
-      "loss": 0.3436,
+      "epoch": 0.21331058020477817,
+      "grad_norm": 13.427445411682129,
+      "learning_rate": 4.6444823663253694e-08,
+      "loss": 0.3361,
       "step": 2000
     },
     {
-      "epoch": 0.36845983787767134,
-      "eval_loss": 1.2414860725402832,
-      "eval_runtime": 76.9749,
-      "eval_samples_per_second": 15.68,
+      "epoch": 0.21331058020477817,
+      "eval_loss": 1.2922736406326294,
+      "eval_runtime": 63.1958,
+      "eval_samples_per_second": 15.618,
       "eval_steps_per_second": 1.962,
       "step": 2000
     },
     {
-      "epoch": 0.46057479734708917,
-      "grad_norm": 323.8576354980469,
-      "learning_rate": 4.232375337754851e-08,
-      "loss": 0.366,
+      "epoch": 0.2666382252559727,
+      "grad_norm": 320.7197265625,
+      "learning_rate": 4.555602957906712e-08,
+      "loss": 0.3597,
       "step": 2500
     },
     {
-      "epoch": 0.46057479734708917,
-      "eval_loss": 1.1949334144592285,
-      "eval_runtime": 77.0381,
-      "eval_samples_per_second": 15.668,
-      "eval_steps_per_second": 1.96,
+      "epoch": 0.2666382252559727,
+      "eval_loss": 1.257211685180664,
+      "eval_runtime": 63.3241,
+      "eval_samples_per_second": 15.586,
+      "eval_steps_per_second": 1.958,
       "step": 2500
     },
     {
-      "epoch": 0.552689756816507,
-      "grad_norm": 339.28533935546875,
-      "learning_rate": 4.078850405305821e-08,
-      "loss": 0.3191,
+      "epoch": 0.3199658703071672,
+      "grad_norm": 0.0019184721168130636,
+      "learning_rate": 4.4667235494880546e-08,
+      "loss": 0.3042,
       "step": 3000
     },
     {
-      "epoch": 0.552689756816507,
-      "eval_loss": 1.1560280323028564,
-      "eval_runtime": 77.0026,
-      "eval_samples_per_second": 15.675,
-      "eval_steps_per_second": 1.961,
+      "epoch": 0.3199658703071672,
+      "eval_loss": 1.2240594625473022,
+      "eval_runtime": 63.5118,
+      "eval_samples_per_second": 15.54,
+      "eval_steps_per_second": 1.952,
       "step": 3000
     },
     {
-      "epoch": 0.6448047162859248,
-      "grad_norm": 60.27859878540039,
-      "learning_rate": 3.925325472856792e-08,
-      "loss": 0.2779,
+      "epoch": 0.37329351535836175,
+      "grad_norm": 11.738512992858887,
+      "learning_rate": 4.3778441410693965e-08,
+      "loss": 0.3252,
       "step": 3500
     },
     {
-      "epoch": 0.6448047162859248,
-      "eval_loss": 1.1233855485916138,
-      "eval_runtime": 76.9006,
-      "eval_samples_per_second": 15.696,
-      "eval_steps_per_second": 1.964,
+      "epoch": 0.37329351535836175,
+      "eval_loss": 1.1988952159881592,
+      "eval_runtime": 63.5384,
+      "eval_samples_per_second": 15.534,
+      "eval_steps_per_second": 1.952,
       "step": 3500
     }
   ],
   "logging_steps": 500,
-  "max_steps": 16284,
+  "max_steps": 28128,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 500,

checkpoint-3500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cbcece3d46fd39b0ba80eb3e495fdf46c9f891af5b4bddce9de77ad5a467b625
+oid sha256:590a3eed3f4b2ba9fcb0b831d5e4c3f2410a46be8f57155ebf9fdd97c6c96524
 size 4847