End of training

README.md

@@ -2,6 +2,8 @@
 library_name: transformers
 base_model: andro-flock/b2-classification
 tags:
+- image-segmentation
+- vision
 - generated_from_trainer
 model-index:
 - name: b2-segmentation
@@ -13,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # b2-segmentation
 
-This model is a fine-tuned version of [andro-flock/b2-classification](https://huggingface.co/andro-flock/b2-classification) on an unknown dataset.
+This model is a fine-tuned version of [andro-flock/b2-classification](https://huggingface.co/andro-flock/b2-classification) on the andro-flock/semantic-segment-4class dataset.
 
 ## Model description
 

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 500.0,
-    "total_flos": 7.8886713950208e+16,
-    "train_loss": 0.1895984401702881,
-    "train_runtime": 397.1334,
-    "train_samples_per_second": 20.144,
-    "train_steps_per_second": 2.518
+    "epoch": 250.0,
+    "total_flos": 2.903671701504e+17,
+    "train_loss": 0.08589911651611327,
+    "train_runtime": 1399.8181,
+    "train_samples_per_second": 2.858,
+    "train_steps_per_second": 0.357
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 500.0,
-    "total_flos": 7.8886713950208e+16,
-    "train_loss": 0.1895984401702881,
-    "train_runtime": 397.1334,
-    "train_samples_per_second": 20.144,
-    "train_steps_per_second": 2.518
+    "epoch": 250.0,
+    "total_flos": 2.903671701504e+17,
+    "train_loss": 0.08589911651611327,
+    "train_runtime": 1399.8181,
+    "train_samples_per_second": 2.858,
+    "train_steps_per_second": 0.357
 }

trainer_state.json

@@ -1,62 +1,41 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 500.0,
+  "epoch": 250.0,
   "eval_steps": 500,
-  "global_step": 1000,
+  "global_step": 500,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 100.0,
-      "grad_norm": 1.8578704595565796,
-      "learning_rate": 4.8020000000000004e-05,
-      "loss": 0.4234,
+      "grad_norm": 0.17752039432525635,
+      "learning_rate": 3.604e-05,
+      "loss": 0.1262,
       "step": 200
     },
     {
       "epoch": 200.0,
-      "grad_norm": 0.6928507685661316,
-      "learning_rate": 3.604e-05,
-      "loss": 0.1814,
-      "step": 400
-    },
-    {
-      "epoch": 300.0,
-      "grad_norm": 1.1086845397949219,
-      "learning_rate": 2.406e-05,
-      "loss": 0.1281,
-      "step": 600
-    },
-    {
-      "epoch": 400.0,
-      "grad_norm": 1.4631288051605225,
+      "grad_norm": 0.3985879421234131,
       "learning_rate": 1.2079999999999998e-05,
-      "loss": 0.1144,
-      "step": 800
-    },
-    {
-      "epoch": 500.0,
-      "grad_norm": 0.9606817364692688,
-      "learning_rate": 1e-07,
-      "loss": 0.1007,
-      "step": 1000
+      "loss": 0.062,
+      "step": 400
     },
     {
-      "epoch": 500.0,
-      "step": 1000,
-      "total_flos": 7.8886713950208e+16,
-      "train_loss": 0.1895984401702881,
-      "train_runtime": 397.1334,
-      "train_samples_per_second": 20.144,
-      "train_steps_per_second": 2.518
+      "epoch": 250.0,
+      "step": 500,
+      "total_flos": 2.903671701504e+17,
+      "train_loss": 0.08589911651611327,
+      "train_runtime": 1399.8181,
+      "train_samples_per_second": 2.858,
+      "train_steps_per_second": 0.357
     }
   ],
   "logging_steps": 200,
-  "max_steps": 1000,
+  "max_steps": 500,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 500,
+  "num_train_epochs": 250,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -70,7 +49,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 7.8886713950208e+16,
+  "total_flos": 2.903671701504e+17,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null