{
  "id": 469511,
  "title": "About TPU...",
  "url": "/competitions/hms-harmful-brain-activity-classification/discussion/469511",
  "author_name": "",
  "post_date": "2024-01-20T22:59:06.897710400Z",
  "votes": 2,
  "comment_count": 4,
  "views": 0,
  "content": "<p>who can tell me why I can not use the TPU,,,,</p>",
  "messages": [
    {
      "id": "2611715",
      "postDate": "01/20/2024 22:59:06",
      "content": "<p>who can tell me why I can not use the TPU,,,,</p>",
      "rawMarkdown": "who can tell me why I can not use the TPU,,,,",
      "votes": null
    },
    {
      "id": "2611722",
      "postDate": "01/20/2024 23:29:00",
      "content": "<p>More details are needed…</p>",
      "rawMarkdown": "More details are needed...",
      "votes": null
    },
    {
      "id": "2611740",
      "postDate": "01/20/2024 23:55:22",
      "content": "<p>Hi Liuyanfeng,</p>\n<p>Last time I tried I got that message:  TPU are POP</p>\n<p>\"TPUs are popular right now. You are #48 in the queue. You can wait, try connecting again later, or use another accelerator.\"</p>\n<p>Then humbly 😆 I changed to GPU cause I didn't have time to wait for TPU.</p>",
      "rawMarkdown": "Hi Liuyanfeng,\n\nLast time I tried I got that message:  TPU are POP\n\n\"TPUs are popular right now. You are #48 in the queue. You can wait, try connecting again later, or use another accelerator.\"\n\nThen humbly 😆 I changed to GPU cause I didn't have time to wait for TPU.",
      "votes": null
    },
    {
      "id": "2611758",
      "postDate": "01/21/2024 00:42:56",
      "content": "<p>same source as below that was working in the [Stanford Ribonanza RNA Folding].<br>\nbut this time not worked, has error as below …<br>\nby information of error , the tup is not even opened…<br>\nis something changed?<br>\nthank you very much..<br>\nsource:<br>\ntry:<br>\n    tpu = tf.distribute.cluster_resolver.TPUClusterResolver.connect(tpu=\"local\") # \"local\" for 1VM TPU<br>\n    print('Running on TPU ')#, tpu.cluster_spec().as_dict()['worker'])<br>\nexcept:<br>\n    tpu = None<br>\nif tpu:<br>\n    strategy = tf.distribute.TPUStrategy(tpu)<br>\n    print(\"on TPU\")<br>\n    print(\"REPLICAS: \", strategy.num_replicas_in_sync)<br>\nelse:<br>\n    print(\"on GPU\")<br>\n    strategy = tf.distribute.get_strategy()<br>\nwith strategy.scope():<br>\n    …..<br>\nerror:<br>\non GPU<br>\n0 / 3 ###################<br>\n2024-01-21 00:41:09.001994: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.<br>\n2024-01-21 00:41:09.002120: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.<br>\n2024-01-21 00:41:09.002228: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.<br>\n2024-01-21 00:41:09.002329: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.</p>",
      "rawMarkdown": "same source as below that was working in the [Stanford Ribonanza RNA Folding].\nbut this time not worked, has error as below ...\nby information of error , the tup is not even opened...\nis something changed?\nthank you very much..\nsource:\ntry:\n    tpu = tf.distribute.cluster_resolver.TPUClusterResolver.connect(tpu=\"local\") # \"local\" for 1VM TPU\n    print('Running on TPU ')#, tpu.cluster_spec().as_dict()['worker'])\nexcept:\n    tpu = None\nif tpu:\n    strategy = tf.distribute.TPUStrategy(tpu)\n    print(\"on TPU\")\n    print(\"REPLICAS: \", strategy.num_replicas_in_sync)\nelse:\n    print(\"on GPU\")\n    strategy = tf.distribute.get_strategy()\nwith strategy.scope():\n    .....\nerror:\non GPU\n0 / 3 ###################\n2024-01-21 00:41:09.001994: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.\n2024-01-21 00:41:09.002120: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.\n2024-01-21 00:41:09.002228: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.\n2024-01-21 00:41:09.002329: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.",
      "votes": null
    },
    {
      "id": "2611777",
      "postDate": "01/21/2024 01:33:07",
      "content": "<p>Ir prints 'on GPU,' so some error occurs in the 'try' (which will not get printed since it's inside 'try').<br>\nRun the block that is inside the try and report the error. (while you are in a tpu session, of course).<br>\nEven better- make a public version of your notebook that demonstrates the problem, and I will check it.</p>",
      "rawMarkdown": "Ir prints 'on GPU,' so some error occurs in the 'try' (which will not get printed since it's inside 'try').\nRun the block that is inside the try and report the error. (while you are in a tpu session, of course).\nEven better- make a public version of your notebook that demonstrates the problem, and I will check it.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2611722,
      "author_name": "shlomoron",
      "author_url": "",
      "post_date": "01/20/2024 23:29:00",
      "content": "<p>More details are needed…</p>",
      "votes": null,
      "replies": [
        {
          "id": 2611758,
          "author_name": "liuyanfeng",
          "author_url": "",
          "post_date": "01/21/2024 00:42:56",
          "content": "<p>same source as below that was working in the [Stanford Ribonanza RNA Folding].<br>\nbut this time not worked, has error as below …<br>\nby information of error , the tup is not even opened…<br>\nis something changed?<br>\nthank you very much..<br>\nsource:<br>\ntry:<br>\n    tpu = tf.distribute.cluster_resolver.TPUClusterResolver.connect(tpu=\"local\") # \"local\" for 1VM TPU<br>\n    print('Running on TPU ')#, tpu.cluster_spec().as_dict()['worker'])<br>\nexcept:<br>\n    tpu = None<br>\nif tpu:<br>\n    strategy = tf.distribute.TPUStrategy(tpu)<br>\n    print(\"on TPU\")<br>\n    print(\"REPLICAS: \", strategy.num_replicas_in_sync)<br>\nelse:<br>\n    print(\"on GPU\")<br>\n    strategy = tf.distribute.get_strategy()<br>\nwith strategy.scope():<br>\n    …..<br>\nerror:<br>\non GPU<br>\n0 / 3 ###################<br>\n2024-01-21 00:41:09.001994: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.<br>\n2024-01-21 00:41:09.002120: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.<br>\n2024-01-21 00:41:09.002228: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.<br>\n2024-01-21 00:41:09.002329: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2611777,
              "author_name": "shlomoron",
              "author_url": "",
              "post_date": "01/21/2024 01:33:07",
              "content": "<p>Ir prints 'on GPU,' so some error occurs in the 'try' (which will not get printed since it's inside 'try').<br>\nRun the block that is inside the try and report the error. (while you are in a tpu session, of course).<br>\nEven better- make a public version of your notebook that demonstrates the problem, and I will check it.</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 2611740,
      "author_name": "mpwolke",
      "author_url": "",
      "post_date": "01/20/2024 23:55:22",
      "content": "<p>Hi Liuyanfeng,</p>\n<p>Last time I tried I got that message:  TPU are POP</p>\n<p>\"TPUs are popular right now. You are #48 in the queue. You can wait, try connecting again later, or use another accelerator.\"</p>\n<p>Then humbly 😆 I changed to GPU cause I didn't have time to wait for TPU.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2611715": "who can tell me why I can not use the TPU,,,,",
    "2611722": "More details are needed...",
    "2611740": "Hi Liuyanfeng,\n\nLast time I tried I got that message:  TPU are POP\n\n\"TPUs are popular right now. You are #48 in the queue. You can wait, try connecting again later, or use another accelerator.\"\n\nThen humbly 😆 I changed to GPU cause I didn't have time to wait for TPU.",
    "2611758": "same source as below that was working in the [Stanford Ribonanza RNA Folding].\nbut this time not worked, has error as below ...\nby information of error , the tup is not even opened...\nis something changed?\nthank you very much..\nsource:\ntry:\n    tpu = tf.distribute.cluster_resolver.TPUClusterResolver.connect(tpu=\"local\") # \"local\" for 1VM TPU\n    print('Running on TPU ')#, tpu.cluster_spec().as_dict()['worker'])\nexcept:\n    tpu = None\nif tpu:\n    strategy = tf.distribute.TPUStrategy(tpu)\n    print(\"on TPU\")\n    print(\"REPLICAS: \", strategy.num_replicas_in_sync)\nelse:\n    print(\"on GPU\")\n    strategy = tf.distribute.get_strategy()\nwith strategy.scope():\n    .....\nerror:\non GPU\n0 / 3 ###################\n2024-01-21 00:41:09.001994: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.\n2024-01-21 00:41:09.002120: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.\n2024-01-21 00:41:09.002228: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.\n2024-01-21 00:41:09.002329: E external/local_xla/xla/stream_executor/stream_executor_internal.h:177] SetPriority unimplemented for this stream.",
    "2611777": "Ir prints 'on GPU,' so some error occurs in the 'try' (which will not get printed since it's inside 'try').\nRun the block that is inside the try and report the error. (while you are in a tpu session, of course).\nEven better- make a public version of your notebook that demonstrates the problem, and I will check it."
  },
  "source": "meta"
}