End of training

README.md

@@ -32,7 +32,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [facebook/convnext-large-384-22k-1k](https://huggingface.co/facebook/convnext-large-384-22k-1k) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0612
+- Loss: 0.0935
 - Accuracy: 0.9828
 
 ## Model description

all_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 2.938775510204082,
+    "eval_accuracy": 0.9827586206896551,
+    "eval_loss": 0.09347157180309296,
+    "eval_runtime": 6.1975,
+    "eval_samples_per_second": 28.076,
+    "eval_steps_per_second": 3.55,
+    "total_flos": 2.3959024845520896e+18,
+    "train_loss": 0.2514381897118356,
+    "train_runtime": 373.0041,
+    "train_samples_per_second": 12.579,
+    "train_steps_per_second": 0.097
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.938775510204082,
+    "eval_accuracy": 0.9827586206896551,
+    "eval_loss": 0.09347157180309296,
+    "eval_runtime": 6.1975,
+    "eval_samples_per_second": 28.076,
+    "eval_steps_per_second": 3.55
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.938775510204082,
+    "total_flos": 2.3959024845520896e+18,
+    "train_loss": 0.2514381897118356,
+    "train_runtime": 373.0041,
+    "train_samples_per_second": 12.579,
+    "train_steps_per_second": 0.097
+}

trainer_state.json

@@ -0,0 +1,78 @@
+{
+  "best_metric": 0.9827586206896551,
+  "best_model_checkpoint": "convnext-large-384-22k-1k-finetuned-climbing-test1\\checkpoint-12",
+  "epoch": 2.938775510204082,
+  "eval_steps": 500,
+  "global_step": 36,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.8163265306122449,
+      "grad_norm": 13.45537281036377,
+      "learning_rate": 4.0625000000000005e-05,
+      "loss": 0.6932,
+      "step": 10
+    },
+    {
+      "epoch": 0.9795918367346939,
+      "eval_accuracy": 0.9827586206896551,
+      "eval_loss": 0.09347157180309296,
+      "eval_runtime": 9.5962,
+      "eval_samples_per_second": 18.132,
+      "eval_steps_per_second": 2.293,
+      "step": 12
+    },
+    {
+      "epoch": 1.6326530612244898,
+      "grad_norm": 3.6985955238342285,
+      "learning_rate": 2.5e-05,
+      "loss": 0.1248,
+      "step": 20
+    },
+    {
+      "epoch": 1.9591836734693877,
+      "eval_accuracy": 0.9827586206896551,
+      "eval_loss": 0.06910645216703415,
+      "eval_runtime": 5.9283,
+      "eval_samples_per_second": 29.351,
+      "eval_steps_per_second": 3.711,
+      "step": 24
+    },
+    {
+      "epoch": 2.4489795918367347,
+      "grad_norm": 2.4517667293548584,
+      "learning_rate": 9.375000000000001e-06,
+      "loss": 0.0587,
+      "step": 30
+    },
+    {
+      "epoch": 2.938775510204082,
+      "eval_accuracy": 0.9827586206896551,
+      "eval_loss": 0.061170052736997604,
+      "eval_runtime": 5.7649,
+      "eval_samples_per_second": 30.183,
+      "eval_steps_per_second": 3.816,
+      "step": 36
+    },
+    {
+      "epoch": 2.938775510204082,
+      "step": 36,
+      "total_flos": 2.3959024845520896e+18,
+      "train_loss": 0.2514381897118356,
+      "train_runtime": 373.0041,
+      "train_samples_per_second": 12.579,
+      "train_steps_per_second": 0.097
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 36,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "total_flos": 2.3959024845520896e+18,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}