Training in progress, step 2000, checkpoint

checkpoint-2000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7625bb0a9d44bfddb8e3e3fa8362b17b52a5b13db844caa6760ff0e56a2a88c1
+oid sha256:cc6ef7b14a301cca48a0ab3f744e2893273d9083a95b9acffdc93ba93f5f00e2
 size 1711848436

checkpoint-2000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1388be8cc0ed38f1ce073a8f5d7fa64ab54a13049645d7e0dfbff1854d0f4b40
+oid sha256:7a33d2013e7ecd9f7dfc7f136b8edf21ee4b7a2bc97a5866b0b46ebd94ca25cb
 size 3424043887

checkpoint-2000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c8a05793b13a11f248845b723ced182d25b409b522da96d1b80e0c13d9347bc9
+oid sha256:af1620fbfd02fde52b24927b0ab8a20aabb919a41f2de9ed57869a532f8fbed4
 size 623

checkpoint-2000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.8716973662376404,
+  "best_metric": 0.8570474982261658,
   "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l57-l/checkpoint-2000",
-  "epoch": 0.2251491613193741,
+  "epoch": 0.21331058020477817,
   "eval_steps": 500,
   "global_step": 2000,
   "is_hyper_param_search": false,
@@ -9,70 +9,70 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.056287290329843524,
-      "grad_norm": 150.1642303466797,
-      "learning_rate": 4.929640887087696e-07,
-      "loss": 0.3486,
+      "epoch": 0.05332764505119454,
+      "grad_norm": 414.2552795410156,
+      "learning_rate": 4.911120591581342e-07,
+      "loss": 0.3809,
       "step": 500
     },
     {
-      "epoch": 0.056287290329843524,
-      "eval_loss": 1.1265727281570435,
-      "eval_runtime": 123.7806,
-      "eval_samples_per_second": 15.948,
-      "eval_steps_per_second": 1.995,
+      "epoch": 0.05332764505119454,
+      "eval_loss": 1.1109352111816406,
+      "eval_runtime": 63.6734,
+      "eval_samples_per_second": 15.501,
+      "eval_steps_per_second": 1.947,
       "step": 500
     },
     {
-      "epoch": 0.11257458065968705,
-      "grad_norm": 366.6913146972656,
-      "learning_rate": 4.859281774175391e-07,
-      "loss": 0.2733,
+      "epoch": 0.10665529010238908,
+      "grad_norm": 45.880027770996094,
+      "learning_rate": 4.822241183162685e-07,
+      "loss": 0.2649,
       "step": 1000
     },
     {
-      "epoch": 0.11257458065968705,
-      "eval_loss": 0.9742079377174377,
-      "eval_runtime": 125.8047,
-      "eval_samples_per_second": 15.691,
-      "eval_steps_per_second": 1.963,
+      "epoch": 0.10665529010238908,
+      "eval_loss": 0.9588962197303772,
+      "eval_runtime": 62.6825,
+      "eval_samples_per_second": 15.746,
+      "eval_steps_per_second": 1.978,
       "step": 1000
     },
     {
-      "epoch": 0.16886187098953057,
-      "grad_norm": 555.5016479492188,
-      "learning_rate": 4.788922661263087e-07,
-      "loss": 0.1851,
+      "epoch": 0.1599829351535836,
+      "grad_norm": 432.9915466308594,
+      "learning_rate": 4.733361774744027e-07,
+      "loss": 0.2104,
       "step": 1500
     },
     {
-      "epoch": 0.16886187098953057,
-      "eval_loss": 0.9162159562110901,
-      "eval_runtime": 126.7606,
-      "eval_samples_per_second": 15.573,
-      "eval_steps_per_second": 1.949,
+      "epoch": 0.1599829351535836,
+      "eval_loss": 0.90110844373703,
+      "eval_runtime": 62.6757,
+      "eval_samples_per_second": 15.748,
+      "eval_steps_per_second": 1.978,
       "step": 1500
     },
     {
-      "epoch": 0.2251491613193741,
-      "grad_norm": 0.423260897397995,
-      "learning_rate": 4.7185635483507824e-07,
-      "loss": 0.1973,
+      "epoch": 0.21331058020477817,
+      "grad_norm": 0.014208819717168808,
+      "learning_rate": 4.6444823663253695e-07,
+      "loss": 0.1849,
       "step": 2000
     },
     {
-      "epoch": 0.2251491613193741,
-      "eval_loss": 0.8716973662376404,
-      "eval_runtime": 127.6955,
-      "eval_samples_per_second": 15.459,
-      "eval_steps_per_second": 1.934,
+      "epoch": 0.21331058020477817,
+      "eval_loss": 0.8570474982261658,
+      "eval_runtime": 63.7296,
+      "eval_samples_per_second": 15.487,
+      "eval_steps_per_second": 1.946,
       "step": 2000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 35532,
+  "max_steps": 28128,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 3,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {

checkpoint-2000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:319be2a769ef7117b6826908976f080f9082f159f4f0ad75da1a151ad29d9bff
+oid sha256:057f6bbf1ff3e34abda0863071bbf5fdcc7edb1102f04dba05c7a15087bc3396
 size 4847