Training in progress, step 2000, checkpoint

checkpoint-2000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:658ef593b5afe2803ad151bb23ad9fc23bf079302e1efe6f225e871d6c01269d
+oid sha256:ba4265b3765397f8aac0db504336d3d24e72b2fca800f06f1b8872260af35079
 size 1711848436

checkpoint-2000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b7ba47c1dadf8a574d11678bc2e0fc387eb9dba9ebd5c6b2e041c52486ea6348
+oid sha256:d92bca3e6b652f0244bbce2c24e5a1729353a782bca5392072c9987827ff7c9d
 size 3424043887

checkpoint-2000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0d804d7128be80d6c8b6b31acdc5c1b6f28216f78ebede9eee58f2e93b67c2fe
+oid sha256:06ba4f7e9a44701e1b8b7414dd6412ad68dfe04c3ac0c42c64fd1c080f355092
 size 623

checkpoint-2000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.8942967057228088,
+  "best_metric": 0.9504629373550415,
   "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l37-l/checkpoint-2000",
-  "epoch": 0.36845983787767134,
+  "epoch": 0.2251491613193741,
   "eval_steps": 500,
   "global_step": 2000,
   "is_hyper_param_search": false,
@@ -9,68 +9,68 @@
   "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
     }
   ],
   "logging_steps": 500,
-  "max_steps": 16284,
+  "max_steps": 26649,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 500,

checkpoint-2000/training_args.bin

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