File size: 4,672 Bytes
d7e1f72
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
84e6b93
 
 
 
 
 
d7e1f72
 
 
 
 
 
 
 
 
 
 
a4ff1a1
 
d7e1f72
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
---
base_model: nomic-ai/nomic-embed-text-v1
inference: false
language:
  - en
license: apache-2.0
model_creator: Nomic
model_name: nomic-embed-text-v1
model_type: bert
pipeline_tag: sentence-similarity
quantized_by: Nomic
tags:
  - feature-extraction
  - sentence-similarity
---

***
**Warning**: There is a llama.cpp PR [about to be merged](https://github.com/ggerganov/llama.cpp/pull/5500) that will break compatibility with these files. Keep an eye out for updates to this repo.
***

<br/>

# nomic-embed-text-v1 - GGUF

Original model: [nomic-embed-text-v1](https://huggingface.co/nomic-ai/nomic-embed-text-v1)


## Description

This repo contains llama.cpp-compatible files for [nomic-embed-text-v1](https://huggingface.co/nomic-ai/nomic-embed-text-v1) in GGUF format.

llama.cpp will default to 2048 tokens of context with these files. To use the full 8192 tokens that Nomic Embed is benchmarked on, you will have to choose a context extension method. The original model uses Dynamic NTK-Aware RoPE scaling, but that is not currently available in llama.cpp. A combination of YaRN and linear scaling is an acceptable substitute.

These files were converted and quantized with llama.cpp commit [6c00a0669](https://github.com/ggerganov/llama.cpp/commit/6c00a066928b0475b865a2e3e709e2166e02d548).

## Example `llama.cpp` Command

Compute a single embedding:
```shell
./embedding -ngl 99 -m nomic-embed-text-v1.f16.gguf -c 8192 -b 8192 --rope-scaling yarn --rope-freq-scale .75 -p 'search_query: What is TSNE?'
```

You can also submit a batch of texts to embed, as long as the total number of tokens does not exceed the context length. Only the first three embeddings are shown by the `embedding` example.

texts.txt:
```
search_query: What is TSNE?
search_query: Who is Laurens Van der Maaten?
```

Compute multiple embeddings:
```shell
./embedding -ngl 99 -m nomic-embed-text-v1.f16.gguf -c 8192 -b 8192 --rope-scaling yarn --rope-freq-scale .75 -f texts.txt
```


## Compatibility

These files are compatible with llama.cpp as commit [ea9c8e114](https://github.com/ggerganov/llama.cpp/commit/ea9c8e11436ad50719987fa23a289c74b7b40d40) from 2/13/2024.


## Provided Files

The below table shows the mean squared error of the embeddings produced by these quantizations of Nomic Embed relative to the Sentence Transformers implementation.

Name | Quant | Size | MSE 
-----|-------|------|-----
[nomic-embed-text-v1.Q2\_K.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q2_K.gguf) | Q2\_K | 48 MiB | 2.36e-03
[nomic-embed-text-v1.Q3\_K\_S.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q3_K_S.gguf) | Q3\_K\_S | 57 MiB | 1.31e-03
[nomic-embed-text-v1.Q3\_K\_M.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q3_K_M.gguf) | Q3\_K\_M | 65 MiB | 8.73e-04
[nomic-embed-text-v1.Q3\_K\_L.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q3_K_L.gguf) | Q3\_K\_L | 69 MiB | 8.68e-04
[nomic-embed-text-v1.Q4\_0.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q4_0.gguf) | Q4\_0 | 75 MiB | 6.87e-04
[nomic-embed-text-v1.Q4\_K\_S.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q4_K_S.gguf) | Q4\_K\_S | 75 MiB | 6.81e-04
[nomic-embed-text-v1.Q4\_K\_M.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q4_K_M.gguf) | Q4\_K\_M | 81 MiB | 3.12e-04
[nomic-embed-text-v1.Q5\_0.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q5_0.gguf) | Q5\_0 | 91 MiB | 2.79e-04
[nomic-embed-text-v1.Q5\_K\_S.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q5_K_S.gguf) | Q5\_K\_S | 91 MiB | 2.61e-04
[nomic-embed-text-v1.Q5\_K\_M.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q5_K_M.gguf) | Q5\_K\_M | 95 MiB | 7.34e-05
[nomic-embed-text-v1.Q6\_K.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q6_K.gguf) | Q6\_K | 108 MiB | 6.29e-05
[nomic-embed-text-v1.Q8\_0.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.Q8_0.gguf) | Q8\_0 | 140 MiB | 6.34e-06
[nomic-embed-text-v1.f16.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.f16.gguf) | F16 | 262 MiB | 5.62e-10
[nomic-embed-text-v1.f32.gguf](https://huggingface.co/nomic-ai/nomic-embed-text-v1-GGUF/blob/main/nomic-embed-text-v1.f32.gguf) | F32 | 262 MiB | 9.34e-11