{
  "id": "tensorflow/resource-exhausted-iterator-get-next",
  "signature": "ResourceExhaustedError: Failed to get next element from iterator: Out of memory while reading data",
  "signature_zh": "资源耗尽错误：从迭代器获取下一个元素失败：读取数据时内存不足",
  "regex": "ResourceExhaustedError: Failed to get next element from iterator: Out of memory while reading data",
  "domain": "tensorflow",
  "category": "resource_error",
  "subcategory": null,
  "root_cause": "Data pipeline prefetching consumes too much memory, often due to large batch size, large dataset elements, or infinite prefetch.",
  "root_cause_type": "generic",
  "root_cause_zh": "数据管道预取消耗过多内存，通常由于批大小过大、数据集元素过大或无限预取导致。",
  "versions": [
    {
      "version": "tensorflow 2.13.0",
      "introduced": null,
      "deprecated": null,
      "removed": null,
      "behavior_change": null,
      "status": "active"
    },
    {
      "version": "tensorflow 2.14.0",
      "introduced": null,
      "deprecated": null,
      "removed": null,
      "behavior_change": null,
      "status": "active"
    }
  ],
  "os_specific": {},
  "dead_ends": [
    {
      "action": "",
      "why_fails": "Memory is consumed by prefetch buffer, not worker count.",
      "fail_rate": 0.7,
      "condition": "",
      "sources": []
    },
    {
      "action": "",
      "why_fails": "The error is about TensorFlow's internal buffer limits, not system memory capacity.",
      "fail_rate": 0.9,
      "condition": "",
      "sources": []
    }
  ],
  "workarounds": [
    {
      "action": "Limit the prefetch buffer size using tf.data.Dataset.prefetch with a specific buffer size:\ndataset = dataset.batch(32).prefetch(buffer_size=tf.data.AUTOTUNE)\n# Or manually set buffer_size to a small number like 1 or 2\n# Alternatively, use .prefetch(1) to limit to one batch",
      "success_rate": 0.85,
      "how": "Limit the prefetch buffer size using tf.data.Dataset.prefetch with a specific buffer size:\ndataset = dataset.batch(32).prefetch(buffer_size=tf.data.AUTOTUNE)\n# Or manually set buffer_size to a small number like 1 or 2\n# Alternatively, use .prefetch(1) to limit to one batch",
      "condition": "",
      "sources": []
    },
    {
      "action": "Reduce batch size to lower memory usage per element:\ndataset = dataset.batch(16).prefetch(tf.data.AUTOTUNE)\n# Or use smaller batch size like 8",
      "success_rate": 0.8,
      "how": "Reduce batch size to lower memory usage per element:\ndataset = dataset.batch(16).prefetch(tf.data.AUTOTUNE)\n# Or use smaller batch size like 8",
      "condition": "",
      "sources": []
    },
    {
      "action": "Use tf.data.Dataset.cache to avoid re-reading large files, combined with controlled prefetch:\ndataset = dataset.cache().batch(32).prefetch(1)",
      "success_rate": 0.75,
      "how": "Use tf.data.Dataset.cache to avoid re-reading large files, combined with controlled prefetch:\ndataset = dataset.cache().batch(32).prefetch(1)",
      "condition": "",
      "sources": []
    }
  ],
  "workarounds_zh": [
    "Limit the prefetch buffer size using tf.data.Dataset.prefetch with a specific buffer size:\ndataset = dataset.batch(32).prefetch(buffer_size=tf.data.AUTOTUNE)\n# Or manually set buffer_size to a small number like 1 or 2\n# Alternatively, use .prefetch(1) to limit to one batch",
    "Reduce batch size to lower memory usage per element:\ndataset = dataset.batch(16).prefetch(tf.data.AUTOTUNE)\n# Or use smaller batch size like 8",
    "Use tf.data.Dataset.cache to avoid re-reading large files, combined with controlled prefetch:\ndataset = dataset.cache().batch(32).prefetch(1)"
  ],
  "transition_graph": {
    "leads_to": [],
    "preceded_by": [],
    "frequently_confused_with": []
  },
  "official_doc_url": "https://www.tensorflow.org/api_docs/python/tf/data/Dataset#prefetch",
  "official_doc_section": null,
  "error_code": "REI",
  "verification_tier": "ai_generated",
  "confidence": 0.86,
  "fix_success_rate": 0.82,
  "resolvable": "true",
  "first_seen": "2024-06-18",
  "last_confirmed": "2024-06-01",
  "last_updated": "2024-06-01",
  "evidence_count": 1,
  "tags": [],
  "locale": "en",
  "aliases": []
}