CVE-2026-133428 (GCVE-0-2026-133428)

Vulnerability from cvelistv5 – Published: 2026-11-20 07:32 – Updated: 2026-11-20 17:19
VLAI
Title
Quadratic time and memory multi-host URL validation
Summary
Validating an untrusted URL with a long scheme and many hosts into a multi-host URL field (e.g. PostgresDsn) takes quadratic time and memory: an input of about a hundred kilobytes takes around a gigabyte of memory to validate. The parse_multihost_url() function (in pydantic-core/src/validators/url.rs) parses each host separately, prepending the whole URL prefix (including the scheme) to every one of them. As the scheme can be arbitrarily long, and is parsed and stored once per host, both time and memory grow quadratically with the input size. Allowed schemes (e.g. postgres for PostgresDsn) are only checked after parsing, so they don't prevent the issue. The issue was fixed in commit b4028f9d87e34750e53be0ba011d70f3c0fcc94e. Impact: denial of service through validation of untrusted input. Affected are applications validating untrusted strings into multi-host URL fields (PostgresDsn, MongoDsn, NatsDsn, or custom types built on MultiHostUrl).
SSVC
Exploitation: none Automatable: no Technical Impact: partial
CISA Coordinator · CISA-ADP (v2.0.3)
Decision recorded 2026-11-20 17:19 UTC
CWE
  • CWE-407 - Inefficient Algorithmic Complexity
Impacted products
Vendor Product Version
pydantic pydantic Affected: 2.0.0 , < 2.14.0 (python)
Create a notification for this product.
Credits
Show details on NVD website

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-133428",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T17:19:11.695942Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-11-20T17:19:22.719Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "collectionURL": "https://pypi.org",
          "defaultStatus": "unaffected",
          "packageName": "pydantic",
          "product": "pydantic",
          "repo": "https://github.com/pydantic/pydantic",
          "vendor": "pydantic",
          "versions": [
            {
              "lessThan": "2.14.0",
              "status": "affected",
              "version": "2.0.0",
              "versionType": "python"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "finder",
          "value": "iamibi"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "supportingMedia": [
            {
              "base64": false,
              "type": "text/html",
              "value": "\u003ch3\u003eSummary\u003c/h3\u003e\n\u003cp\u003eValidating an untrusted URL with a long scheme and many hosts into a multi-host URL field (e.g. \u003ccode\u003ePostgresDsn\u003c/code\u003e) takes quadratic time and memory: an input of about a hundred kilobytes takes around a gigabyte of memory to validate.\u003c/p\u003e\n\u003ch3\u003eDetails\u003c/h3\u003e\n\u003cp\u003eThe \u003ca href=\"https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L537\"\u003e\u003ccode\u003eparse_multihost_url()\u003c/code\u003e\u003c/a\u003e function (in \u003ccode\u003epydantic-core/src/validators/url.rs\u003c/code\u003e) parses each host separately, prepending the whole URL prefix (including the scheme) to every one of them (\u003ca href=\"https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L644-L649\"\u003eL644-L649\u003c/a\u003e):\u003c/p\u003e\n\u003cpre\u003e\u003ccode class=\"language-rust\"\u003elet extra_urls: Vec\u0026lt;Url\u0026gt; = hosts\n    .iter()\n    .map(|host| {\n        let reconstructed_url = format!(\"{prefix}{host}\");\n        parse_url(\u0026amp;reconstructed_url, input, strict)\n    })\n    .collect::\u0026lt;ValResult\u0026lt;_\u0026gt;\u0026gt;()?;\n\u003c/code\u003e\u003c/pre\u003e\n\u003cp\u003eAs the scheme can be arbitrarily long, and is parsed and stored once per host, both time and memory grow quadratically with the input size. Allowed schemes (e.g. \u003ccode\u003epostgres\u003c/code\u003e for \u003ccode\u003ePostgresDsn\u003c/code\u003e) are only checked after parsing, so they don\u0027t prevent the issue.\u003c/p\u003e\n\u003cp\u003eThe issue was fixed in commit \u003ca href=\"https://github.com/pydantic/pydantic/commit/b4028f9d87e34750e53be0ba011d70f3c0fcc94e\"\u003eb4028f9d87e34750e53be0ba011d70f3c0fcc94e\u003c/a\u003e.\u003c/p\u003e\n\u003ch3\u003eImpact\u003c/h3\u003e\n\u003cp\u003eDenial of service through validation of untrusted input. Affected are applications validating untrusted strings into multi-host URL fields (\u003ccode\u003ePostgresDsn\u003c/code\u003e, \u003ccode\u003eMongoDsn\u003c/code\u003e, \u003ccode\u003eNatsDsn\u003c/code\u003e, or custom types built on \u003ccode\u003eMultiHostUrl\u003c/code\u003e).\u003c/p\u003e"
            }
          ],
          "value": "Validating an untrusted URL with a long scheme and many hosts into a multi-host URL field (e.g. PostgresDsn) takes quadratic time and memory: an input of about a hundred kilobytes takes around a gigabyte of memory to validate.\n\nThe parse_multihost_url() function (in pydantic-core/src/validators/url.rs) parses each host separately, prepending the whole URL prefix (including the scheme) to every one of them. As the scheme can be arbitrarily long, and is parsed and stored once per host, both time and memory grow quadratically with the input size. Allowed schemes (e.g. postgres for PostgresDsn) are only checked after parsing, so they don\u0027t prevent the issue.\n\nThe issue was fixed in commit b4028f9d87e34750e53be0ba011d70f3c0fcc94e.\n\nImpact: denial of service through validation of untrusted input. Affected are applications validating untrusted strings into multi-host URL fields (PostgresDsn, MongoDsn, NatsDsn, or custom types built on MultiHostUrl)."
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "baseScore": 6.3,
            "baseSeverity": "MEDIUM",
            "vectorString": "CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N",
            "version": "4.0"
          }
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-407",
              "description": "CWE-407: Inefficient Algorithmic Complexity",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-11-20T07:32:16.979Z",
        "orgId": "7cec5f2f-3032-4006-8740-31f605ffab5f",
        "shortName": "Pydantic"
      },
      "references": [
        {
          "tags": [
            "vendor-advisory"
          ],
          "url": "https://github.com/pydantic/pydantic/security/advisories/GHSA-955g-x97g-8mcp"
        },
        {
          "url": "https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L537"
        },
        {
          "url": "https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L644-L649"
        },
        {
          "url": "https://github.com/pydantic/pydantic/commit/b4028f9d87e34750e53be0ba011d70f3c0fcc94e"
        }
      ],
      "source": {
        "advisory": "GHSA-955g-x97g-8mcp",
        "discovery": "UNKNOWN"
      },
      "title": "Quadratic time and memory multi-host URL validation"
    }
  },
  "cveMetadata": {
    "assignerOrgId": "7cec5f2f-3032-4006-8740-31f605ffab5f",
    "assignerShortName": "Pydantic",
    "cveId": "CVE-2026-133428",
    "datePublished": "2026-11-20T07:32:16.979Z",
    "dateReserved": "2026-11-19T19:47:33.202Z",
    "dateUpdated": "2026-11-20T17:19:22.719Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2",
  "vulnerability-lookup:meta": {
    "epss": {
      "cve": "CVE-2026-133428",
      "date": "2026-11-22",
      "epss": "0.00861",
      "percentile": "0.59446"
    },
    "microsoft_vex": {
      "current_release_date": "2026-11-22T01:04:01.000Z",
      "cve": "CVE-2026-133428",
      "id": "msrc_CVE-2026-133428",
      "initial_release_date": "2026-11-20T18:07:03.000Z",
      "product_status:under_investigation": "2",
      "source": "Microsoft CSAF VEX",
      "status": "final",
      "title": "Quadratic time and memory multi-host URL validation",
      "url": "https://msrc.microsoft.com/csaf/vex/2026/msrc_cve-2026-133428.json",
      "version": "2"
    },
    "nvd": {
      "cve": {
        "affected": [
          {
            "affectedData": [
              {
                "collectionURL": "https://pypi.org",
                "defaultStatus": "unaffected",
                "packageName": "pydantic",
                "product": "pydantic",
                "repo": "https://github.com/pydantic/pydantic",
                "vendor": "pydantic",
                "versions": [
                  {
                    "lessThan": "2.14.0",
                    "status": "affected",
                    "version": "2.0.0",
                    "versionType": "python"
                  }
                ]
              }
            ],
            "source": "7cec5f2f-3032-4006-8740-31f605ffab5f"
          }
        ],
        "cveTags": [],
        "descriptions": [
          {
            "lang": "en",
            "value": "Validating an untrusted URL with a long scheme and many hosts into a multi-host URL field (e.g. PostgresDsn) takes quadratic time and memory: an input of about a hundred kilobytes takes around a gigabyte of memory to validate.\n\nThe parse_multihost_url() function (in pydantic-core/src/validators/url.rs) parses each host separately, prepending the whole URL prefix (including the scheme) to every one of them. As the scheme can be arbitrarily long, and is parsed and stored once per host, both time and memory grow quadratically with the input size. Allowed schemes (e.g. postgres for PostgresDsn) are only checked after parsing, so they don\u0027t prevent the issue.\n\nThe issue was fixed in commit b4028f9d87e34750e53be0ba011d70f3c0fcc94e.\n\nImpact: denial of service through validation of untrusted input. Affected are applications validating untrusted strings into multi-host URL fields (PostgresDsn, MongoDsn, NatsDsn, or custom types built on MultiHostUrl)."
          }
        ],
        "id": "CVE-2026-133428",
        "lastModified": "2026-11-20T18:18:48.273",
        "metrics": {
          "cvssMetricV40": [
            {
              "cvssData": {
                "Automatable": "NOT_DEFINED",
                "Recovery": "NOT_DEFINED",
                "Safety": "NOT_DEFINED",
                "attackComplexity": "LOW",
                "attackRequirements": "PRESENT",
                "attackVector": "NETWORK",
                "availabilityRequirement": "NOT_DEFINED",
                "baseScore": 6.3,
                "baseSeverity": "MEDIUM",
                "confidentialityRequirement": "NOT_DEFINED",
                "exploitMaturity": "NOT_DEFINED",
                "integrityRequirement": "NOT_DEFINED",
                "modifiedAttackComplexity": "NOT_DEFINED",
                "modifiedAttackRequirements": "NOT_DEFINED",
                "modifiedAttackVector": "NOT_DEFINED",
                "modifiedPrivilegesRequired": "NOT_DEFINED",
                "modifiedSubAvailabilityImpact": "NOT_DEFINED",
                "modifiedSubConfidentialityImpact": "NOT_DEFINED",
                "modifiedSubIntegrityImpact": "NOT_DEFINED",
                "modifiedUserInteraction": "NOT_DEFINED",
                "modifiedVulnAvailabilityImpact": "NOT_DEFINED",
                "modifiedVulnConfidentialityImpact": "NOT_DEFINED",
                "modifiedVulnIntegrityImpact": "NOT_DEFINED",
                "privilegesRequired": "NONE",
                "providerUrgency": "NOT_DEFINED",
                "subAvailabilityImpact": "NONE",
                "subConfidentialityImpact": "NONE",
                "subIntegrityImpact": "NONE",
                "userInteraction": "NONE",
                "valueDensity": "NOT_DEFINED",
                "vectorString": "CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
                "version": "4.0",
                "vulnAvailabilityImpact": "LOW",
                "vulnConfidentialityImpact": "NONE",
                "vulnIntegrityImpact": "NONE",
                "vulnerabilityResponseEffort": "NOT_DEFINED"
              },
              "source": "7cec5f2f-3032-4006-8740-31f605ffab5f",
              "type": "Secondary"
            }
          ],
          "ssvcV203": [
            {
              "source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
              "ssvcData": {
                "id": "CVE-2026-133428",
                "options": [
                  {
                    "exploitation": "none"
                  },
                  {
                    "automatable": "no"
                  },
                  {
                    "technicalImpact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-11-20T17:19:11.695942Z",
                "version": "2.0.3"
              }
            }
          ]
        },
        "published": "2026-11-20T08:18:32.597",
        "references": [
          {
            "source": "7cec5f2f-3032-4006-8740-31f605ffab5f",
            "url": "https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L537"
          },
          {
            "source": "7cec5f2f-3032-4006-8740-31f605ffab5f",
            "url": "https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L644-L649"
          },
          {
            "source": "7cec5f2f-3032-4006-8740-31f605ffab5f",
            "url": "https://github.com/pydantic/pydantic/commit/b4028f9d87e34750e53be0ba011d70f3c0fcc94e"
          },
          {
            "source": "7cec5f2f-3032-4006-8740-31f605ffab5f",
            "url": "https://github.com/pydantic/pydantic/security/advisories/GHSA-955g-x97g-8mcp"
          }
        ],
        "sourceIdentifier": "7cec5f2f-3032-4006-8740-31f605ffab5f",
        "vulnStatus": "Received",
        "weaknesses": [
          {
            "description": [
              {
                "lang": "en",
                "value": "CWE-407"
              }
            ],
            "source": "7cec5f2f-3032-4006-8740-31f605ffab5f",
            "type": "Secondary"
          }
        ]
      }
    },
    "redhat_vex": {
      "aggregate_severity": "Moderate",
      "current_release_date": "2026-11-20T11:28:48+00:00",
      "cve": "CVE-2026-133428",
      "id": "CVE-2026-133428",
      "initial_release_date": "2026-11-20T07:32:16.979000+00:00",
      "product_status:known_affected": "134",
      "source": "Red Hat CSAF VEX",
      "status": "final",
      "title": "pydantic: pydantic-core: pydantic-core: Denial of Service via crafted multi-host URL parsing",
      "url": "https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-133428.json",
      "version": "3"
    },
    "vulnrichment": {
      "containers": {
        "adp": [
          {
            "metrics": [
              {
                "other": {
                  "content": {
                    "id": "CVE-2026-133428",
                    "options": [
                      {
                        "Exploitation": "none"
                      },
                      {
                        "Automatable": "no"
                      },
                      {
                        "Technical Impact": "partial"
                      }
                    ],
                    "role": "CISA Coordinator",
                    "timestamp": "2026-11-20T17:19:11.695942Z",
                    "version": "2.0.3"
                  },
                  "type": "ssvc"
                }
              }
            ],
            "providerMetadata": {
              "dateUpdated": "2026-11-20T17:19:17.002Z",
              "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
              "shortName": "CISA-ADP"
            },
            "title": "CISA ADP Vulnrichment"
          }
        ],
        "cna": {
          "affected": [
            {
              "collectionURL": "https://pypi.org",
              "defaultStatus": "unaffected",
              "packageName": "pydantic",
              "product": "pydantic",
              "repo": "https://github.com/pydantic/pydantic",
              "vendor": "pydantic",
              "versions": [
                {
                  "lessThan": "2.14.0",
                  "status": "affected",
                  "version": "2.0.0",
                  "versionType": "python"
                }
              ]
            }
          ],
          "credits": [
            {
              "lang": "en",
              "type": "finder",
              "value": "iamibi"
            }
          ],
          "descriptions": [
            {
              "lang": "en",
              "supportingMedia": [
                {
                  "base64": false,
                  "type": "text/html",
                  "value": "\u003ch3\u003eSummary\u003c/h3\u003e\n\u003cp\u003eValidating an untrusted URL with a long scheme and many hosts into a multi-host URL field (e.g. \u003ccode\u003ePostgresDsn\u003c/code\u003e) takes quadratic time and memory: an input of about a hundred kilobytes takes around a gigabyte of memory to validate.\u003c/p\u003e\n\u003ch3\u003eDetails\u003c/h3\u003e\n\u003cp\u003eThe \u003ca href=\"https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L537\"\u003e\u003ccode\u003eparse_multihost_url()\u003c/code\u003e\u003c/a\u003e function (in \u003ccode\u003epydantic-core/src/validators/url.rs\u003c/code\u003e) parses each host separately, prepending the whole URL prefix (including the scheme) to every one of them (\u003ca href=\"https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L644-L649\"\u003eL644-L649\u003c/a\u003e):\u003c/p\u003e\n\u003cpre\u003e\u003ccode class=\"language-rust\"\u003elet extra_urls: Vec\u0026lt;Url\u0026gt; = hosts\n    .iter()\n    .map(|host| {\n        let reconstructed_url = format!(\"{prefix}{host}\");\n        parse_url(\u0026amp;reconstructed_url, input, strict)\n    })\n    .collect::\u0026lt;ValResult\u0026lt;_\u0026gt;\u0026gt;()?;\n\u003c/code\u003e\u003c/pre\u003e\n\u003cp\u003eAs the scheme can be arbitrarily long, and is parsed and stored once per host, both time and memory grow quadratically with the input size. Allowed schemes (e.g. \u003ccode\u003epostgres\u003c/code\u003e for \u003ccode\u003ePostgresDsn\u003c/code\u003e) are only checked after parsing, so they don\u0027t prevent the issue.\u003c/p\u003e\n\u003cp\u003eThe issue was fixed in commit \u003ca href=\"https://github.com/pydantic/pydantic/commit/b4028f9d87e34750e53be0ba011d70f3c0fcc94e\"\u003eb4028f9d87e34750e53be0ba011d70f3c0fcc94e\u003c/a\u003e.\u003c/p\u003e\n\u003ch3\u003eImpact\u003c/h3\u003e\n\u003cp\u003eDenial of service through validation of untrusted input. Affected are applications validating untrusted strings into multi-host URL fields (\u003ccode\u003ePostgresDsn\u003c/code\u003e, \u003ccode\u003eMongoDsn\u003c/code\u003e, \u003ccode\u003eNatsDsn\u003c/code\u003e, or custom types built on \u003ccode\u003eMultiHostUrl\u003c/code\u003e).\u003c/p\u003e"
                }
              ],
              "value": "Validating an untrusted URL with a long scheme and many hosts into a multi-host URL field (e.g. PostgresDsn) takes quadratic time and memory: an input of about a hundred kilobytes takes around a gigabyte of memory to validate.\n\nThe parse_multihost_url() function (in pydantic-core/src/validators/url.rs) parses each host separately, prepending the whole URL prefix (including the scheme) to every one of them. As the scheme can be arbitrarily long, and is parsed and stored once per host, both time and memory grow quadratically with the input size. Allowed schemes (e.g. postgres for PostgresDsn) are only checked after parsing, so they don\u0027t prevent the issue.\n\nThe issue was fixed in commit b4028f9d87e34750e53be0ba011d70f3c0fcc94e.\n\nImpact: denial of service through validation of untrusted input. Affected are applications validating untrusted strings into multi-host URL fields (PostgresDsn, MongoDsn, NatsDsn, or custom types built on MultiHostUrl)."
            }
          ],
          "metrics": [
            {
              "cvssV4_0": {
                "baseScore": 6.3,
                "baseSeverity": "MEDIUM",
                "vectorString": "CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N",
                "version": "4.0"
              }
            }
          ],
          "problemTypes": [
            {
              "descriptions": [
                {
                  "cweId": "CWE-407",
                  "description": "CWE-407: Inefficient Algorithmic Complexity",
                  "lang": "en",
                  "type": "CWE"
                }
              ]
            }
          ],
          "providerMetadata": {
            "dateUpdated": "2026-11-20T07:32:16.979Z",
            "orgId": "7cec5f2f-3032-4006-8740-31f605ffab5f",
            "shortName": "Pydantic"
          },
          "references": [
            {
              "tags": [
                "vendor-advisory"
              ],
              "url": "https://github.com/pydantic/pydantic/security/advisories/GHSA-955g-x97g-8mcp"
            },
            {
              "url": "https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L537"
            },
            {
              "url": "https://github.com/pydantic/pydantic/blob/v2.13.5/pydantic-core/src/validators/url.rs#L644-L649"
            },
            {
              "url": "https://github.com/pydantic/pydantic/commit/b4028f9d87e34750e53be0ba011d70f3c0fcc94e"
            }
          ],
          "source": {
            "advisory": "GHSA-955g-x97g-8mcp",
            "discovery": "UNKNOWN"
          },
          "title": "Quadratic time and memory multi-host URL validation"
        }
      },
      "cveMetadata": {
        "assignerOrgId": "7cec5f2f-3032-4006-8740-31f605ffab5f",
        "assignerShortName": "Pydantic",
        "cveId": "CVE-2026-133428",
        "datePublished": "2026-11-20T07:32:16.979Z",
        "dateReserved": "2026-11-19T19:47:33.202Z",
        "dateUpdated": "2026-11-20T17:19:22.719Z",
        "state": "PUBLISHED"
      },
      "dataType": "CVE_RECORD",
      "dataVersion": "5.2"
    }
  }
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

Loading…

Loading…

Loading…

Related by attack behaviour

Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.


Loading…