File size: 4,582 Bytes
31a7b6b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
45d4401
31a7b6b
 
 
 
 
 
 
 
9c09f86
 
 
 
 
31a7b6b
 
 
 
 
 
 
 
 
 
8afb473
 
31a7b6b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e2e85ec
31a7b6b
9c09f86
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
---
base_model: HuggingFaceTB/SmolVLM2-2.2B-Instruct
library_name: transformers
license: apache-2.0
model_creator: HuggingFaceTB
model_name: SmolVLM2-2.2B-Instruct
quantized_by: Second State Inc.
pipeline_tag: image-text-to-text
tags:
- video-text-to-text
language:
- en
---

<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->

# SmolVLM2-2.2B-Instruct-GGUF

## Original Model

[HuggingFaceTB/SmolVLM2-2.2B-Instruct](https://huggingface.co/HuggingFaceTB/SmolVLM2-2.2B-Instruct)

## Run with LlamaEdge

- LlamaEdge version: [v0.21.0](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.21.0) and above

- Prompt template

  - Prompt type: `smol-vision`

  - Prompt string

    ```text
    <|im_start|>
    User: {user_message_1}<image>
    Assistant: {assistant_message_1}
    User: {user_message_2}<image>
    Assistant:
    ```

- Context size: `2048`

- Run as LlamaEdge service

  ```bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:SmolVLM2-2.2B-Instruct-Q5_K_M.gguf \
    llama-api-server.wasm \
    --prompt-template smol-vision \
    --llava-mmproj SmolVLM2-2.2B-Instruct-mmproj-f16.gguf \
    --model-name SmolVLM2-2.2B-Instruct \
    --ctx-size 2048
  ```

## Quantized GGUF Models

| Name | Quant method | Bits | Size | Use case |
| ---- | ---- | ---- | ---- | ----- |
| [SmolVLM2-2.2B-Instruct-Q2_K.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q2_K.gguf)     | Q2_K   | 2 | 708 MB| smallest, significant quality loss - not recommended for most purposes |
| [SmolVLM2-2.2B-Instruct-Q3_K_L.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q3_K_L.gguf) | Q3_K_L | 3 | 976 MB| small, substantial quality loss |
| [SmolVLM2-2.2B-Instruct-Q3_K_M.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q3_K_M.gguf) | Q3_K_M | 3 | 904 MB| very small, high quality loss |
| [SmolVLM2-2.2B-Instruct-Q3_K_S.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q3_K_S.gguf) | Q3_K_S | 3 | 820 MB| very small, high quality loss |
| [SmolVLM2-2.2B-Instruct-Q4_0.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q4_0.gguf)     | Q4_0   | 4 | 1.05 GB| legacy; small, very high quality loss - prefer using Q3_K_M |
| [SmolVLM2-2.2B-Instruct-Q4_K_M.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q4_K_M.gguf) | Q4_K_M | 4 | 1.11 GB| medium, balanced quality - recommended |
| [SmolVLM2-2.2B-Instruct-Q4_K_S.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q4_K_S.gguf) | Q4_K_S | 4 | 1.06 GB| small, greater quality loss |
| [SmolVLM2-2.2B-Instruct-Q5_0.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q5_0.gguf)     | Q5_0   | 5 | 1.26 GB| legacy; medium, balanced quality - prefer using Q4_K_M |
| [SmolVLM2-2.2B-Instruct-Q5_K_M.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q5_K_M.gguf) | Q5_K_M | 5 | 1.30 GB| large, very low quality loss - recommended |
| [SmolVLM2-2.2B-Instruct-Q5_K_S.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q5_K_S.gguf) | Q5_K_S | 5 | 1.26 GB| large, low quality loss - recommended |
| [SmolVLM2-2.2B-Instruct-Q6_K.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q6_K.gguf)     | Q6_K   | 6 | 1.49 GB| very large, extremely low quality loss |
| [SmolVLM2-2.2B-Instruct-Q8_0.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-Q8_0.gguf)     | Q8_0   | 8 | 1.93 GB| very large, extremely low quality loss - not recommended |
| [SmolVLM2-2.2B-Instruct-f16.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-f16.gguf)       | f16    | 16 | 3.63 GB| |
| [SmolVLM2-2.2B-Instruct-mmproj-f16.gguf](https://huggingface.co/second-state/SmolVLM2-2.2B-Instruct-GGUF/blob/main/SmolVLM2-2.2B-Instruct-mmproj-f16.gguf)       | f16    | 16 | 872 MB| |

*Quantized with llama.cpp b5501*