Hanhpt23/Clip-Soil-patch32

README.md

@@ -13,6 +13,8 @@ should probably proofread and complete it, then remove this comment. -->
 # Clip-Soil-patch32
 
 This model is a fine-tuned version of [openai/clip-vit-base-patch32](https://huggingface.co/openai/clip-vit-base-patch32) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 5.9887
 
 ## Model description
 

all_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 1.0,
-    "eval_loss": 1.8086835145950317,
-    "eval_runtime": 2.0062,
-    "eval_samples_per_second": 33.894,
-    "eval_steps_per_second": 4.486,
-    "total_flos": 35598202755192.0,
-    "train_loss": 2.0196751185825894,
-    "train_runtime": 33.4326,
-    "train_samples_per_second": 18.305,
-    "train_steps_per_second": 2.303
+    "epoch": 50.0,
+    "eval_loss": 5.988685131072998,
+    "eval_runtime": 0.9434,
+    "eval_samples_per_second": 72.078,
+    "eval_steps_per_second": 9.54,
+    "total_flos": 1779910137759600.0,
+    "train_loss": 0.6525665847976486,
+    "train_runtime": 847.6015,
+    "train_samples_per_second": 36.102,
+    "train_steps_per_second": 4.542
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 1.0,
-    "eval_loss": 1.8086835145950317,
-    "eval_runtime": 2.0062,
-    "eval_samples_per_second": 33.894,
-    "eval_steps_per_second": 4.486
+    "epoch": 50.0,
+    "eval_loss": 5.988685131072998,
+    "eval_runtime": 0.9434,
+    "eval_samples_per_second": 72.078,
+    "eval_steps_per_second": 9.54
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "total_flos": 35598202755192.0,
-    "train_loss": 2.0196751185825894,
-    "train_runtime": 33.4326,
-    "train_samples_per_second": 18.305,
-    "train_steps_per_second": 2.303
+    "epoch": 50.0,
+    "total_flos": 1779910137759600.0,
+    "train_loss": 0.6525665847976486,
+    "train_runtime": 847.6015,
+    "train_samples_per_second": 36.102,
+    "train_steps_per_second": 4.542
 }

trainer_state.json

@@ -1,27 +1,76 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 50.0,
   "eval_steps": 500,
-  "global_step": 77,
+  "global_step": 3850,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.0,
-      "step": 77,
-      "total_flos": 35598202755192.0,
-      "train_loss": 2.0196751185825894,
-      "train_runtime": 33.4326,
-      "train_samples_per_second": 18.305,
-      "train_steps_per_second": 2.303
+      "epoch": 6.4935064935064934,
+      "grad_norm": 17.624969482421875,
+      "learning_rate": 4.3506493506493503e-05,
+      "loss": 1.7452,
+      "step": 500
+    },
+    {
+      "epoch": 12.987012987012987,
+      "grad_norm": 15.936542510986328,
+      "learning_rate": 3.701298701298702e-05,
+      "loss": 1.1727,
+      "step": 1000
+    },
+    {
+      "epoch": 19.48051948051948,
+      "grad_norm": 22.365955352783203,
+      "learning_rate": 3.051948051948052e-05,
+      "loss": 0.8119,
+      "step": 1500
+    },
+    {
+      "epoch": 25.974025974025974,
+      "grad_norm": 19.36739730834961,
+      "learning_rate": 2.4025974025974027e-05,
+      "loss": 0.5254,
+      "step": 2000
+    },
+    {
+      "epoch": 32.467532467532465,
+      "grad_norm": 4.489462852478027,
+      "learning_rate": 1.7532467532467535e-05,
+      "loss": 0.3811,
+      "step": 2500
+    },
+    {
+      "epoch": 38.96103896103896,
+      "grad_norm": 0.12353431433439255,
+      "learning_rate": 1.103896103896104e-05,
+      "loss": 0.2244,
+      "step": 3000
+    },
+    {
+      "epoch": 45.45454545454545,
+      "grad_norm": 4.508783340454102,
+      "learning_rate": 4.5454545454545455e-06,
+      "loss": 0.1191,
+      "step": 3500
+    },
+    {
+      "epoch": 50.0,
+      "step": 3850,
+      "total_flos": 1779910137759600.0,
+      "train_loss": 0.6525665847976486,
+      "train_runtime": 847.6015,
+      "train_samples_per_second": 36.102,
+      "train_steps_per_second": 4.542
     }
   ],
   "logging_steps": 500,
-  "max_steps": 77,
+  "max_steps": 3850,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 1,
+  "num_train_epochs": 50,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -29,13 +78,13 @@
         "should_epoch_stop": false,
         "should_evaluate": false,
         "should_log": false,
-        "should_save": false,
+        "should_save": true,
         "should_training_stop": false
       },
       "attributes": {}
     }
   },
-  "total_flos": 35598202755192.0,
+  "total_flos": 1779910137759600.0,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null