{
  "id": 332140,
  "title": "Train Accuracy is 0.074 and Val Accuracy is 0.000000 why ?",
  "url": "/competitions/bird-watch-summer-2022/discussion/332140",
  "author_name": "",
  "post_date": "2022-06-20T16:09:02.826866200Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I am Trying to Classify 280 Image species with  a 71-layer deep network CNN but model traning performace is too low……</p>\n<p>Here is my notebook,<br>\n<a href=\"https://www.kaggle.com/code/muki2003/280-birds-species-prediction-cnn\" target=\"_blank\">https://www.kaggle.com/code/muki2003/280-birds-species-prediction-cnn</a></p>",
  "messages": [
    {
      "id": "1826803",
      "postDate": "06/20/2022 16:09:02",
      "content": "<p>I am Trying to Classify 280 Image species with  a 71-layer deep network CNN but model traning performace is too low……</p>\n<p>Here is my notebook,<br>\n<a href=\"https://www.kaggle.com/code/muki2003/280-birds-species-prediction-cnn\" target=\"_blank\">https://www.kaggle.com/code/muki2003/280-birds-species-prediction-cnn</a></p>",
      "rawMarkdown": "I am Trying to Classify 280 Image species with  a 71-layer deep network CNN but model traning performace is too low......\n\nHere is my notebook,\nhttps://www.kaggle.com/code/muki2003/280-birds-species-prediction-cnn",
      "votes": null
    },
    {
      "id": "1827467",
      "postDate": "06/21/2022 04:29:46",
      "content": "<p>I saw your notebooks.<br>\nAnd I think because your model is so complicated.<br>\nAfter I removed some layers after basemodel, the accuracy was improved.</p>\n<p>I tried )<br>\nmodel = tf.keras.Sequential()<br>\nmodel.add(base_model)<br>\nmodel.add(layers.GlobalAveragePooling2D())<br>\nmodel.add(layers.Dense(270, activation='softmax'))<br>\nmodel.summary()</p>\n<p>improved…! :)<br>\nEpoch 1/100<br>\n3298/3298 [==============================] - 255s 76ms/step - loss: 4.9924 - accuracy: 0.0277 - val_loss: 10.8435 - val_accuracy: 0.0000e+00</p>\n<p>Epoch 00001: val_accuracy improved from -inf to 0.00000, saving model to Model.h5<br>\nEpoch 2/100<br>\n3298/3298 [==============================] - 250s 76ms/step - loss: 3.8538 - accuracy: 0.1343 - val_loss: 9.0936 - val_accuracy: 0.0027</p>\n<p>Epoch 00002: val_accuracy improved from 0.00000 to 0.00273, saving model to Model.h5<br>\nEpoch 3/100<br>\n 174/3298 [&gt;………………………..] - ETA: 3:16 - loss: 3.1605 - accuracy: 0.2356</p>",
      "rawMarkdown": "I saw your notebooks.\nAnd I think because your model is so complicated.\nAfter I removed some layers after basemodel, the accuracy was improved.\n\nI tried )\nmodel = tf.keras.Sequential()\nmodel.add(base_model)\nmodel.add(layers.GlobalAveragePooling2D())\nmodel.add(layers.Dense(270, activation='softmax'))\nmodel.summary()\n\nimproved...! :)\nEpoch 1/100\n3298/3298 [==============================] - 255s 76ms/step - loss: 4.9924 - accuracy: 0.0277 - val_loss: 10.8435 - val_accuracy: 0.0000e+00\n\nEpoch 00001: val_accuracy improved from -inf to 0.00000, saving model to Model.h5\nEpoch 2/100\n3298/3298 [==============================] - 250s 76ms/step - loss: 3.8538 - accuracy: 0.1343 - val_loss: 9.0936 - val_accuracy: 0.0027\n\nEpoch 00002: val_accuracy improved from 0.00000 to 0.00273, saving model to Model.h5\nEpoch 3/100\n 174/3298 [>.............................] - ETA: 3:16 - loss: 3.1605 - accuracy: 0.2356",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1827467,
      "author_name": "contrabase",
      "author_url": "",
      "post_date": "06/21/2022 04:29:46",
      "content": "<p>I saw your notebooks.<br>\nAnd I think because your model is so complicated.<br>\nAfter I removed some layers after basemodel, the accuracy was improved.</p>\n<p>I tried )<br>\nmodel = tf.keras.Sequential()<br>\nmodel.add(base_model)<br>\nmodel.add(layers.GlobalAveragePooling2D())<br>\nmodel.add(layers.Dense(270, activation='softmax'))<br>\nmodel.summary()</p>\n<p>improved…! :)<br>\nEpoch 1/100<br>\n3298/3298 [==============================] - 255s 76ms/step - loss: 4.9924 - accuracy: 0.0277 - val_loss: 10.8435 - val_accuracy: 0.0000e+00</p>\n<p>Epoch 00001: val_accuracy improved from -inf to 0.00000, saving model to Model.h5<br>\nEpoch 2/100<br>\n3298/3298 [==============================] - 250s 76ms/step - loss: 3.8538 - accuracy: 0.1343 - val_loss: 9.0936 - val_accuracy: 0.0027</p>\n<p>Epoch 00002: val_accuracy improved from 0.00000 to 0.00273, saving model to Model.h5<br>\nEpoch 3/100<br>\n 174/3298 [&gt;………………………..] - ETA: 3:16 - loss: 3.1605 - accuracy: 0.2356</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1826803": "I am Trying to Classify 280 Image species with  a 71-layer deep network CNN but model traning performace is too low......\n\nHere is my notebook,\nhttps://www.kaggle.com/code/muki2003/280-birds-species-prediction-cnn",
    "1827467": "I saw your notebooks.\nAnd I think because your model is so complicated.\nAfter I removed some layers after basemodel, the accuracy was improved.\n\nI tried )\nmodel = tf.keras.Sequential()\nmodel.add(base_model)\nmodel.add(layers.GlobalAveragePooling2D())\nmodel.add(layers.Dense(270, activation='softmax'))\nmodel.summary()\n\nimproved...! :)\nEpoch 1/100\n3298/3298 [==============================] - 255s 76ms/step - loss: 4.9924 - accuracy: 0.0277 - val_loss: 10.8435 - val_accuracy: 0.0000e+00\n\nEpoch 00001: val_accuracy improved from -inf to 0.00000, saving model to Model.h5\nEpoch 2/100\n3298/3298 [==============================] - 250s 76ms/step - loss: 3.8538 - accuracy: 0.1343 - val_loss: 9.0936 - val_accuracy: 0.0027\n\nEpoch 00002: val_accuracy improved from 0.00000 to 0.00273, saving model to Model.h5\nEpoch 3/100\n 174/3298 [>.............................] - ETA: 3:16 - loss: 3.1605 - accuracy: 0.2356"
  },
  "source": "meta"
}