{
  "id": 306610,
  "title": "UnavailableError: Socket closed ERROR helP.........",
  "url": "/competitions/happy-whale-and-dolphin/discussion/306610",
  "author_name": "Akhilesh D. Kapse",
  "post_date": "2022-02-10T06:11:23.493000",
  "votes": 3,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I'm trying custom training loop (TPU version) and getting this types of errors:<br>\n <code>UnavailableError: Socket closed</code></p>\n<p><code>UnavailableError: Socket closed\nAdditional GRPC error information:\n{\"created\":\"@1644424497.024981884\",\"description\":\"Error received from peer ipv4:10.0.0.2:8470\",\"file\":\"external/com_github_grpc_grpc/src/core/lib/surface/call.cc\",\"file_line\":1056,\"grpc_message\":\"Socket closed\",\"grpc_status\":14}</code></p>\n<p><code>UnavailableError: Socket closed\nAdditional GRPC error information from remote target /job:worker/replica:0/task:0:\n:{\"created\":\"@1644497781.356904322\",\"description\":\"Error received from peer ipv4:10.81.154.42:8470\",\"file\":\"external/com_github_grpc_grpc/src/core/lib/surface/call.cc\",\"file_line\":1056,\"grpc_message\":\"Socket closed\",\"grpc_status\":14}</code></p>\n<p>Tried various suggestion on net so far like-</p>\n<ul>\n<li>changing feed data to float32</li>\n<li>lowering batch_size, epoch. etc<br>\n-running with tf 2.2.0</li>\n<li><code>from cloud_tpu_client import Client\nClient().configure_tpu_version(tf.__version__, restart_type='ifNeeded')</code></li>\n</ul>\n<p>but it dosn't resolved the PROBLEM.</p>\n<p>PLease Help🙏<br>\n~Thanks</p>",
  "messages": [
    {
      "id": 1683899,
      "postDate": "2022-02-10T06:11:23.493Z",
      "content": "<p>I'm trying custom training loop (TPU version) and getting this types of errors:<br>\n <code>UnavailableError: Socket closed</code></p>\n<p><code>UnavailableError: Socket closed\nAdditional GRPC error information:\n{\"created\":\"@1644424497.024981884\",\"description\":\"Error received from peer ipv4:10.0.0.2:8470\",\"file\":\"external/com_github_grpc_grpc/src/core/lib/surface/call.cc\",\"file_line\":1056,\"grpc_message\":\"Socket closed\",\"grpc_status\":14}</code></p>\n<p><code>UnavailableError: Socket closed\nAdditional GRPC error information from remote target /job:worker/replica:0/task:0:\n:{\"created\":\"@1644497781.356904322\",\"description\":\"Error received from peer ipv4:10.81.154.42:8470\",\"file\":\"external/com_github_grpc_grpc/src/core/lib/surface/call.cc\",\"file_line\":1056,\"grpc_message\":\"Socket closed\",\"grpc_status\":14}</code></p>\n<p>Tried various suggestion on net so far like-</p>\n<ul>\n<li>changing feed data to float32</li>\n<li>lowering batch_size, epoch. etc<br>\n-running with tf 2.2.0</li>\n<li><code>from cloud_tpu_client import Client\nClient().configure_tpu_version(tf.__version__, restart_type='ifNeeded')</code></li>\n</ul>\n<p>but it dosn't resolved the PROBLEM.</p>\n<p>PLease Help🙏<br>\n~Thanks</p>",
      "rawMarkdown": "I'm trying custom training loop (TPU version) and getting this types of errors:\n `UnavailableError: Socket closed`\n\n`UnavailableError: Socket closed\nAdditional GRPC error information:\n{\"created\":\"@1644424497.024981884\",\"description\":\"Error received from peer ipv4:10.0.0.2:8470\",\"file\":\"external/com_github_grpc_grpc/src/core/lib/surface/call.cc\",\"file_line\":1056,\"grpc_message\":\"Socket closed\",\"grpc_status\":14}`\n\n`UnavailableError: Socket closed\nAdditional GRPC error information from remote target /job:worker/replica:0/task:0:\n:{\"created\":\"@1644497781.356904322\",\"description\":\"Error received from peer ipv4:10.81.154.42:8470\",\"file\":\"external/com_github_grpc_grpc/src/core/lib/surface/call.cc\",\"file_line\":1056,\"grpc_message\":\"Socket closed\",\"grpc_status\":14}`\n\nTried various suggestion on net so far like-\n- changing feed data to float32\n- lowering batch_size, epoch. etc\n-running with tf 2.2.0\n- `from cloud_tpu_client import Client\nClient().configure_tpu_version(tf.__version__, restart_type='ifNeeded')`\n\nbut it dosn't resolved the PROBLEM.\n\nPLease Help🙏\n~Thanks\n\n",
      "votes": 3
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1683899": "I'm trying custom training loop (TPU version) and getting this types of errors:\n `UnavailableError: Socket closed`\n\n`UnavailableError: Socket closed\nAdditional GRPC error information:\n{\"created\":\"@1644424497.024981884\",\"description\":\"Error received from peer ipv4:10.0.0.2:8470\",\"file\":\"external/com_github_grpc_grpc/src/core/lib/surface/call.cc\",\"file_line\":1056,\"grpc_message\":\"Socket closed\",\"grpc_status\":14}`\n\n`UnavailableError: Socket closed\nAdditional GRPC error information from remote target /job:worker/replica:0/task:0:\n:{\"created\":\"@1644497781.356904322\",\"description\":\"Error received from peer ipv4:10.81.154.42:8470\",\"file\":\"external/com_github_grpc_grpc/src/core/lib/surface/call.cc\",\"file_line\":1056,\"grpc_message\":\"Socket closed\",\"grpc_status\":14}`\n\nTried various suggestion on net so far like-\n- changing feed data to float32\n- lowering batch_size, epoch. etc\n-running with tf 2.2.0\n- `from cloud_tpu_client import Client\nClient().configure_tpu_version(tf.__version__, restart_type='ifNeeded')`\n\nbut it dosn't resolved the PROBLEM.\n\nPLease Help🙏\n~Thanks\n\n"
  }
}