ContextQMD
Back to Redis

VEMB

content/commands/vemb.md

latest3.6 KB
Original Source

Return the approximate vector associated with a given element in the vector set.

shell
VEMB word_embeddings SQL
1) "0.18208661675453186"
2) "0.08535309880971909"
3) "0.1365649551153183"
4) "-0.16501599550247192"
5) "0.14225517213344574"
... 295 more elements ...

Vector sets normalize and may quantize vectors on insertion. VEMB reverses this process to approximate the original vector by de-normalizing and de-quantizing it.

To retrieve the raw internal representation, use the RAW option:

shell
VEMB word_embeddings apple RAW
1) int8
2) "\xf1\xdc\xfd\x1e\xcc%E...\xde\x1f\xfbN" # artificially shortened for this example
3) "3.1426539421081543"
4) "0.17898885905742645"

Required arguments

<details open> <summary><code>key</code></summary>

is the name of the key that holds the vector set.

</details> <details open> <summary><code>element</code></summary>

is the name of the element whose vector you want to retrieve.

</details>

Optional arguments

<details open> <summary><code>RAW</code></summary>

returns the raw vector data, its quantization type, and metadata such as norm and range.

</details>
  • [Vector sets]({{< relref "/develop/data-types/vector-sets" >}})

Redis Software and Redis Cloud compatibility

| Redis Software | Redis Cloud | <span style="min-width: 9em; display: table-cell">Notes</span> | |:----------------------|:-----------------|:------| | <span title="Supported">✅ Standard</span> <span title="Not supported"><nobr>❌ Active-Active</nobr></span> | <span title="Supported">✅ Standard</span> <span title="Not supported"><nobr>❌ Active-Active</nobr></span> | |

Return information

{{< multitabs id="vemb-return-info" tab1="RESP2" tab2="RESP3" >}}

One of the following:

  • Array reply: of real numbers as bulk strings, representing the vector.
  • Array reply: consisting of the following elements:
    1. The quantization type as a simple string: fp32, bin, or q8.
    2. A bulk string blob with the following raw data:
      • 4-byte floats for fp32 (little-endian encoding)
      • A bitmap for binary quantization
      • A byte array for q8
    3. The L2 norm, as a simple string, of the vector before normalization.
    4. (Only for q8): The quantization range as a simple string. Multiply this by integer components to recover normalized values.

-tab-sep-

One of the following:

  • Array reply: of doubles, representing the vector.
  • Array reply: consisting of the following elements:
    1. The quantization type as a simple string: fp32, bin, or q8.
    2. A bulk string blob with the following raw data:
      • 4-byte floats for fp32 (little-endian encoding)
      • A bitmap for binary quantization
      • A byte array for q8
    3. The double L2 norm of the vector before normalization.
    4. (Only for q8): The quantization range as a double. Multiply this by integer components to recover normalized values.

{{< /multitabs >}}