Training in progress, step 4000, checkpoint

checkpoint-4000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7540034aa0a9e12d911b1764112258ac582fcc73f095efb17bab75391e341bd5
+oid sha256:a29f91b13918406626f95a285145596c4d7cbc2bd101aabf060146916d85e144
 size 1711848436

checkpoint-4000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52e9c31d4a49a8bf12f80795e49d6637db98cc2e0c9dc41aedcdd0c667d547c5
+oid sha256:b514ffd2da9541e3a38fbb11366bf0009607dceaa30c6cc84d839ef79066f9f8
 size 3424043887

checkpoint-4000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44c9fa5b803fb1f28770699a359bea30c519543808cc61771c36f033b35e9034
+oid sha256:b28f4c61991590f45a2b30389764764a767dc541e740fd34fe04b268dde61e01
 size 623

checkpoint-4000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 1.0920671224594116,
+  "best_metric": 1.169384241104126,
   "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l58-l/checkpoint-4000",
-  "epoch": 0.7369196757553427,
+  "epoch": 0.42662116040955633,
   "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": 20.04043960571289,
-      "learning_rate": 4.8848563006632274e-08,
-      "loss": 0.3758,
+      "epoch": 0.05332764505119454,
+      "grad_norm": 217.55360412597656,
+      "learning_rate": 4.933340443686007e-08,
+      "loss": 0.4667,
       "step": 500
     },
     {
-      "epoch": 0.09211495946941783,
-      "eval_loss": 1.4184927940368652,
-      "eval_runtime": 74.3562,
-      "eval_samples_per_second": 16.233,
-      "eval_steps_per_second": 2.031,
+      "epoch": 0.05332764505119454,
+      "eval_loss": 1.4426143169403076,
+      "eval_runtime": 61.9901,
+      "eval_samples_per_second": 15.922,
+      "eval_steps_per_second": 2.0,
       "step": 500
     },
     {
-      "epoch": 0.18422991893883567,
-      "grad_norm": 74.25025939941406,
-      "learning_rate": 4.7697126013264556e-08,
-      "loss": 0.4103,
+      "epoch": 0.10665529010238908,
+      "grad_norm": 73.73340606689453,
+      "learning_rate": 4.8666808873720136e-08,
+      "loss": 0.4532,
       "step": 1000
     },
     {
-      "epoch": 0.18422991893883567,
-      "eval_loss": 1.3500770330429077,
-      "eval_runtime": 75.7013,
-      "eval_samples_per_second": 15.944,
-      "eval_steps_per_second": 1.995,
+      "epoch": 0.10665529010238908,
+      "eval_loss": 1.3815597295761108,
+      "eval_runtime": 62.4624,
+      "eval_samples_per_second": 15.801,
+      "eval_steps_per_second": 1.985,
       "step": 1000
     },
     {
-      "epoch": 0.2763448784082535,
-      "grad_norm": 0.5102410912513733,
-      "learning_rate": 4.6545689019896826e-08,
-      "loss": 0.433,
+      "epoch": 0.1599829351535836,
+      "grad_norm": 471.4011535644531,
+      "learning_rate": 4.80002133105802e-08,
+      "loss": 0.3749,
       "step": 1500
     },
     {
-      "epoch": 0.2763448784082535,
-      "eval_loss": 1.2885024547576904,
-      "eval_runtime": 75.197,
-      "eval_samples_per_second": 16.051,
-      "eval_steps_per_second": 2.008,
+      "epoch": 0.1599829351535836,
+      "eval_loss": 1.3310539722442627,
+      "eval_runtime": 63.3918,
+      "eval_samples_per_second": 15.57,
+      "eval_steps_per_second": 1.956,
       "step": 1500
     },
     {
-      "epoch": 0.36845983787767134,
-      "grad_norm": 0.1667678952217102,
-      "learning_rate": 4.539425202652911e-08,
-      "loss": 0.3424,
+      "epoch": 0.21331058020477817,
+      "grad_norm": 11.962693214416504,
+      "learning_rate": 4.733361774744027e-08,
+      "loss": 0.336,
       "step": 2000
     },
     {
-      "epoch": 0.36845983787767134,
-      "eval_loss": 1.239119052886963,
-      "eval_runtime": 75.387,
-      "eval_samples_per_second": 16.011,
-      "eval_steps_per_second": 2.003,
+      "epoch": 0.21331058020477817,
+      "eval_loss": 1.2890639305114746,
+      "eval_runtime": 63.177,
+      "eval_samples_per_second": 15.623,
+      "eval_steps_per_second": 1.963,
       "step": 2000
     },
     {
-      "epoch": 0.46057479734708917,
-      "grad_norm": 343.8175048828125,
-      "learning_rate": 4.4242815033161385e-08,
-      "loss": 0.3645,
+      "epoch": 0.2666382252559727,
+      "grad_norm": 375.1609191894531,
+      "learning_rate": 4.666702218430034e-08,
+      "loss": 0.3585,
       "step": 2500
     },
     {
-      "epoch": 0.46057479734708917,
-      "eval_loss": 1.1902339458465576,
-      "eval_runtime": 75.9899,
-      "eval_samples_per_second": 15.884,
-      "eval_steps_per_second": 1.987,
+      "epoch": 0.2666382252559727,
+      "eval_loss": 1.2536433935165405,
+      "eval_runtime": 63.1512,
+      "eval_samples_per_second": 15.629,
+      "eval_steps_per_second": 1.964,
       "step": 2500
     },
     {
-      "epoch": 0.552689756816507,
-      "grad_norm": 305.5626525878906,
-      "learning_rate": 4.309137803979366e-08,
-      "loss": 0.3172,
+      "epoch": 0.3199658703071672,
+      "grad_norm": 0.0022936267778277397,
+      "learning_rate": 4.600042662116041e-08,
+      "loss": 0.303,
       "step": 3000
     },
     {
-      "epoch": 0.552689756816507,
-      "eval_loss": 1.1506118774414062,
-      "eval_runtime": 74.8655,
-      "eval_samples_per_second": 16.122,
-      "eval_steps_per_second": 2.017,
+      "epoch": 0.3199658703071672,
+      "eval_loss": 1.2202869653701782,
+      "eval_runtime": 63.5857,
+      "eval_samples_per_second": 15.522,
+      "eval_steps_per_second": 1.95,
       "step": 3000
     },
     {
-      "epoch": 0.6448047162859248,
-      "grad_norm": 65.59048461914062,
-      "learning_rate": 4.193994104642594e-08,
-      "loss": 0.2751,
+      "epoch": 0.37329351535836175,
+      "grad_norm": 15.805087089538574,
+      "learning_rate": 4.5333831058020476e-08,
+      "loss": 0.3242,
       "step": 3500
     },
     {
-      "epoch": 0.6448047162859248,
-      "eval_loss": 1.1169357299804688,
-      "eval_runtime": 76.2936,
-      "eval_samples_per_second": 15.82,
-      "eval_steps_per_second": 1.979,
+      "epoch": 0.37329351535836175,
+      "eval_loss": 1.195627212524414,
+      "eval_runtime": 63.4024,
+      "eval_samples_per_second": 15.567,
+      "eval_steps_per_second": 1.956,
       "step": 3500
     },
     {
-      "epoch": 0.7369196757553427,
-      "grad_norm": 18.106201171875,
-      "learning_rate": 4.078850405305821e-08,
-      "loss": 0.2919,
+      "epoch": 0.42662116040955633,
+      "grad_norm": 5.377908229827881,
+      "learning_rate": 4.4667235494880546e-08,
+      "loss": 0.2427,
       "step": 4000
     },
     {
-      "epoch": 0.7369196757553427,
-      "eval_loss": 1.0920671224594116,
-      "eval_runtime": 74.9101,
-      "eval_samples_per_second": 16.113,
-      "eval_steps_per_second": 2.016,
+      "epoch": 0.42662116040955633,
+      "eval_loss": 1.169384241104126,
+      "eval_runtime": 63.5082,
+      "eval_samples_per_second": 15.541,
+      "eval_steps_per_second": 1.953,
       "step": 4000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 21712,
+  "max_steps": 37504,
   "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 sha256:9a29c2374e59ce3f57b34db9cb20d2d1bdc870c3dad128e9a4d462636c37e8b0
+oid sha256:87e35dca972e13467c750106c19301b098fbd2918fda1087487e227bac3aca37
 size 4847