End of training

README.md

@@ -15,6 +15,8 @@ should probably proofread and complete it, then remove this comment. -->
 # clip-finetuned
 
 This model is a fine-tuned version of [openai/clip-vit-base-patch32](https://huggingface.co/openai/clip-vit-base-patch32) on the imagefolder dataset.
+It achieves the following results on the evaluation set:
+- Loss: 3.6758
 
 ## Model description
 

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 15.0,
-    "eval_loss": 5.025272369384766,
-    "eval_runtime": 1.359,
-    "eval_samples_per_second": 30.905,
-    "eval_steps_per_second": 0.736,
-    "test_loss": 4.871317386627197,
-    "test_runtime": 2.4966,
-    "test_samples_per_second": 33.646,
-    "test_steps_per_second": 0.801,
-    "train_loss": 1.6000991821289063,
-    "train_runtime": 149.3142,
-    "train_samples_per_second": 34.056,
-    "train_steps_per_second": 0.603
+    "epoch": 10.0,
+    "eval_loss": 3.675844192504883,
+    "eval_runtime": 1.6492,
+    "eval_samples_per_second": 25.467,
+    "eval_steps_per_second": 0.606,
+    "test_loss": 4.198638439178467,
+    "test_runtime": 3.9304,
+    "test_samples_per_second": 21.372,
+    "test_steps_per_second": 0.509,
+    "train_loss": 2.1348843892415363,
+    "train_runtime": 100.2817,
+    "train_samples_per_second": 33.805,
+    "train_steps_per_second": 0.598
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 15.0,
-    "eval_loss": 5.025272369384766,
-    "eval_runtime": 1.359,
-    "eval_samples_per_second": 30.905,
-    "eval_steps_per_second": 0.736
+    "epoch": 10.0,
+    "eval_loss": 3.675844192504883,
+    "eval_runtime": 1.6492,
+    "eval_samples_per_second": 25.467,
+    "eval_steps_per_second": 0.606
 }

test_results.json

@@ -1,6 +1,6 @@
 {
-    "test_loss": 4.871317386627197,
-    "test_runtime": 2.4966,
-    "test_samples_per_second": 33.646,
-    "test_steps_per_second": 0.801
+    "test_loss": 4.198638439178467,
+    "test_runtime": 3.9304,
+    "test_samples_per_second": 21.372,
+    "test_steps_per_second": 0.509
 }

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 15.0,
-    "train_loss": 1.6000991821289063,
-    "train_runtime": 149.3142,
-    "train_samples_per_second": 34.056,
-    "train_steps_per_second": 0.603
+    "epoch": 10.0,
+    "train_loss": 2.1348843892415363,
+    "train_runtime": 100.2817,
+    "train_samples_per_second": 33.805,
+    "train_steps_per_second": 0.598
 }

trainer_state.json

@@ -1,29 +1,29 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 15.0,
+  "epoch": 10.0,
   "eval_steps": 500,
-  "global_step": 90,
+  "global_step": 60,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 15.0,
-      "step": 90,
-      "total_flos": 295779184657110.0,
-      "train_loss": 1.6000991821289063,
-      "train_runtime": 149.3142,
-      "train_samples_per_second": 34.056,
-      "train_steps_per_second": 0.603
+      "epoch": 10.0,
+      "step": 60,
+      "total_flos": 197186123104740.0,
+      "train_loss": 2.1348843892415363,
+      "train_runtime": 100.2817,
+      "train_samples_per_second": 33.805,
+      "train_steps_per_second": 0.598
     }
   ],
   "logging_steps": 500,
-  "max_steps": 90,
+  "max_steps": 60,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 15,
+  "num_train_epochs": 10,
   "save_steps": 500,
-  "total_flos": 295779184657110.0,
+  "total_flos": 197186123104740.0,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null