File size: 3,672 Bytes
3c6df4d
 
 
 
 
 
 
 
 
4886e58
3c6df4d
 
 
41ac648
3c6df4d
 
 
4886e58
3c6df4d
 
 
 
 
 
 
 
41ac648
3c6df4d
 
 
 
 
 
 
 
01737fc
3c6df4d
 
 
 
3f2a21f
 
3c6df4d
 
 
c4303f2
3c6df4d
 
 
 
 
c4303f2
3c6df4d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
base_model: amazon/MistralLite
inference: false
license: apache-2.0
model_creator: Amazon Web Services
model_name: MistralLite 7B
model_type: mistral
quantized_by: Second State Inc.
---

<!-- 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 -->

# MistralLite-7B-GGUF

## Original Model

[amazon/MistralLite](https://huggingface.co/amazon/MistralLite)

## Run with LlamaEdge

- LlamaEdge version: [v0.2.8](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.2.8) and above

- Prompt template

  - Prompt type: `mistrallite`

  - Prompt string

    ```text
    <|prompter|>{user_message}</s><|assistant|>{assistant_message}</s>
    ```

  - Reverse prompt: `</s>`

- Context size: `4096`

- Run as LlamaEdge service

  ```bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:MistralLite-Q5_K_M.gguf llama-api-server.wasm -p mistrallite -r '</s>'
  ```

- Run as LlamaEdge command app

  ```bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:MistralLite-Q5_K_M.gguf llama-chat.wasm -p mistrallite -r '</s>'
  ```

## Quantized GGUF Models

| Name | Quant method | Bits | Size | Use case |
| ---- | ---- | ---- | ---- | ----- |
| [MistralLite-Q2_K.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q2_K.gguf)     | Q2_K   | 2 | 2.7 GB| smallest, significant quality loss - not recommended for most purposes |
| [MistralLite-Q3_K_L.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q3_K_L.gguf) | Q3_K_L | 3 | 3.82 GB| small, substantial quality loss |
| [MistralLite-Q3_K_M.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q3_K_M.gguf) | Q3_K_M | 3 | 3.52 GB| very small, high quality loss |
| [MistralLite-Q3_K_S.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q3_K_S.gguf) | Q3_K_S | 3 | 3.16 GB| very small, high quality loss |
| [MistralLite-Q4_0.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q4_0.gguf)     | Q4_0   | 4 | 4.11 GB| legacy; small, very high quality loss - prefer using Q3_K_M |
| [MistralLite-Q4_K_M.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q4_K_M.gguf) | Q4_K_M | 4 | 4.37 GB| medium, balanced quality - recommended |
| [MistralLite-Q4_K_S.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q4_K_S.gguf) | Q4_K_S | 4 | 4.14 GB| small, greater quality loss |
| [MistralLite-Q5_0.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q5_0.gguf)     | Q5_0   | 5 | 5.00 GB| legacy; medium, balanced quality - prefer using Q4_K_M |
| [MistralLite-Q5_K_M.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q5_K_M.gguf) | Q5_K_M | 5 | 5.13 GB| large, very low quality loss - recommended |
| [MistralLite-Q5_K_S.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q5_K_S.gguf) | Q5_K_S | 5 | 5.00 GB| large, low quality loss - recommended |
| [MistralLite-Q6_K.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q6_K.gguf)     | Q6_K   | 6 | 5.94 GB| very large, extremely low quality loss |
| [MistralLite-Q8_0.gguf](https://huggingface.co/second-state/MistralLite-7B-GGUF/blob/main/MistralLite-Q8_0.gguf)     | Q8_0   | 8 | 7.70 GB| very large, extremely low quality loss - not recommended |