Hugging Face's logo

second-state
/
Megrez-3B-Instruct-GGUF

Text Generation
GGUF
English
Chinese
llama
Inference Endpoints
conversational
Model card Files Community
apepkuss79 commited on
Commit
e0bc976
·
verified ·
1 Parent(s): 00af38d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +84 -81
README.md CHANGED
@@ -1,82 +1,85 @@
1
- ---
2
- base_model: Infinigence/Megrez-3B-Instruct
3
- license: apache-2.0
4
- model_creator: Infinigence
5
- model_name: Megrez-3B-Instruct
6
- quantized_by: Second State Inc.
7
- language:
8
- - en
9
- - zh
10
- pipeline_tag: text-generation
11
- ---
12
-
13
- <!-- header start -->
14
- <!-- 200823 -->
15
- <div style="width: auto; margin-left: auto; margin-right: auto">
16
- <img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;">
17
- </div>
18
- <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
19
- <!-- header end -->
20
-
21
- # Megrez-3B-Instruct-GGUF
22
-
23
- ## Original Model
24
-
25
- [Infinigence/Megrez-3B-Instruct](https://huggingface.co/Infinigence/Megrez-3B-Instruct)
26
-
27
- ## Run with LlamaEdge
28
-
29
- - LlamaEdge version: coming soon
30
-
31
- <!-- - LlamaEdge version: [v0.14.3](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.14.3) and above -->
32
-
33
- - Prompt template
34
-
35
- - Prompt type: `megrez`
36
-
37
- - Prompt string
38
-
39
- ```text
40
- <|role_start|>system<|role_end|>{system_message}<|turn_end|><|role_start|>user<|role_end|>{user_message}<|turn_end|><|role_start|>assistant<|role_end|>
41
- ```
42
-
43
- - Context size: `32000`
44
-
45
- - Run as LlamaEdge service
46
-
47
- ```bash
48
- wasmedge --dir .:. --nn-preload default:GGML:AUTO:Megrez-3B-Instruct-Q5_K_M.gguf \
49
- llama-api-server.wasm \
50
- --model-name Megrez-3B-Instruct \
51
- --prompt-template megrez \
52
- --ctx-size 32000
53
- ```
54
-
55
- - Run as LlamaEdge command app
56
-
57
- ```bash
58
- wasmedge --dir .:. --nn-preload default:GGML:AUTO:Megrez-3B-Instruct-Q5_K_M.gguf \
59
- llama-chat.wasm \
60
- --prompt-template megrez \
61
- --ctx-size 32000
62
- ```
63
-
64
- ## Quantized GGUF Models
65
-
66
- | Name | Quant method | Bits | Size | Use case |
67
- | ---- | ---- | ---- | ---- | ----- |
68
- | [Megrez-3B-Instruct-Q2_K.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q2_K.gguf) | Q2_K | 2 | 1.21 GB| smallest, significant quality loss - not recommended for most purposes |
69
- | [Megrez-3B-Instruct-Q3_K_L.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q3_K_L.gguf) | Q3_K_L | 3 | 1.60 GB| small, substantial quality loss |
70
- | [Megrez-3B-Instruct-Q3_K_M.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q3_K_M.gguf) | Q3_K_M | 3 | 1.50 GB| very small, high quality loss |
71
- | [Megrez-3B-Instruct-Q3_K_S.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q3_K_S.gguf) | Q3_K_S | 3 | 1.38 GB| very small, high quality loss |
72
- | [Megrez-3B-Instruct-Q4_0.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q4_0.gguf) | Q4_0 | 4 | 1.73 GB| legacy; small, very high quality loss - prefer using Q3_K_M |
73
- | [Megrez-3B-Instruct-Q4_K_M.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q4_K_M.gguf) | Q4_K_M | 4 | 1.81 GB| medium, balanced quality - recommended |
74
- | [Megrez-3B-Instruct-Q4_K_S.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q4_K_S.gguf) | Q4_K_S | 4 | 1.74 GB| small, greater quality loss |
75
- | [Megrez-3B-Instruct-Q5_0.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q5_0.gguf) | Q5_0 | 5 | 2.05 GB| legacy; medium, balanced quality - prefer using Q4_K_M |
76
- | [Megrez-3B-Instruct-Q5_K_M.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q5_K_M.gguf) | Q5_K_M | 5 | 2.09 GB| large, very low quality loss - recommended |
77
- | [Megrez-3B-Instruct-Q5_K_S.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q5_K_S.gguf) | Q5_K_S | 5 | 2.05 GB| large, low quality loss - recommended |
78
- | [Megrez-3B-Instruct-Q6_K.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q6_K.gguf) | Q6_K | 6 | 2.40 GB| very large, extremely low quality loss |
79
- | [Megrez-3B-Instruct-Q8_0.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q8_0.gguf) | Q8_0 | 8 | 3.10 GB| very large, extremely low quality loss - not recommended |
80
- | [Megrez-3B-Instruct-f16.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-f16.gguf) | f16 | 16 | 5.84 GB| |
81
-
 
 
 
82
  *Quantized with llama.cpp (commit: b92a14a8)*
 
1
+ ---
2
+ base_model: Infinigence/Megrez-3B-Instruct
3
+ license: apache-2.0
4
+ model_creator: Infinigence
5
+ model_name: Megrez-3B-Instruct
6
+ quantized_by: Second State Inc.
7
+ language:
8
+ - en
9
+ - zh
10
+ pipeline_tag: text-generation
11
+ ---
12
+
13
+ <!-- header start -->
14
+ <!-- 200823 -->
15
+ <div style="width: auto; margin-left: auto; margin-right: auto">
16
+ <img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;">
17
+ </div>
18
+ <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
19
+ <!-- header end -->
20
+
21
+ # Megrez-3B-Instruct-GGUF
22
+
23
+ ## Original Model
24
+
25
+ [Infinigence/Megrez-3B-Instruct](https://huggingface.co/Infinigence/Megrez-3B-Instruct)
26
+
27
+ ## Run with LlamaEdge
28
+
29
+ - LlamaEdge version: coming soon
30
+
31
+ <!-- - LlamaEdge version: [v0.14.3](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.14.3) and above -->
32
+
33
+ - Prompt template
34
+
35
+ - Prompt type: `megrez`
36
+
37
+ - Prompt string
38
+
39
+ ```text
40
+ <|role_start|>system<|role_end|>{system_message}<|turn_end|><|role_start|>user<|role_end|>{user_message}<|turn_end|><|role_start|>assistant<|role_end|>
41
+ ```
42
+
43
+ - Context size: `32000`
44
+
45
+ - Run as LlamaEdge service
46
+
47
+ ```bash
48
+ wasmedge --dir .:. --nn-preload default:GGML:AUTO:Megrez-3B-Instruct-Q5_K_M.gguf \
49
+ llama-api-server.wasm \
50
+ --model-name Megrez-3B-Instruct \
51
+ --prompt-template megrez \
52
+ --ctx-size 32000
53
+ ```
54
+
55
+ > For use cases of conversations or article writing, `temperature=0.7` is strongly recommended.
56
+ > For use cases of mathematics or logical reasoning, `temperature=0.2` is strongly recommended.
57
+
58
+ - Run as LlamaEdge command app
59
+
60
+ ```bash
61
+ wasmedge --dir .:. --nn-preload default:GGML:AUTO:Megrez-3B-Instruct-Q5_K_M.gguf \
62
+ llama-chat.wasm \
63
+ --prompt-template megrez \
64
+ --ctx-size 32000
65
+ ```
66
+
67
+ ## Quantized GGUF Models
68
+
69
+ | Name | Quant method | Bits | Size | Use case |
70
+ | ---- | ---- | ---- | ---- | ----- |
71
+ | [Megrez-3B-Instruct-Q2_K.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q2_K.gguf) | Q2_K | 2 | 1.21 GB| smallest, significant quality loss - not recommended for most purposes |
72
+ | [Megrez-3B-Instruct-Q3_K_L.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q3_K_L.gguf) | Q3_K_L | 3 | 1.60 GB| small, substantial quality loss |
73
+ | [Megrez-3B-Instruct-Q3_K_M.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q3_K_M.gguf) | Q3_K_M | 3 | 1.50 GB| very small, high quality loss |
74
+ | [Megrez-3B-Instruct-Q3_K_S.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q3_K_S.gguf) | Q3_K_S | 3 | 1.38 GB| very small, high quality loss |
75
+ | [Megrez-3B-Instruct-Q4_0.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q4_0.gguf) | Q4_0 | 4 | 1.73 GB| legacy; small, very high quality loss - prefer using Q3_K_M |
76
+ | [Megrez-3B-Instruct-Q4_K_M.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q4_K_M.gguf) | Q4_K_M | 4 | 1.81 GB| medium, balanced quality - recommended |
77
+ | [Megrez-3B-Instruct-Q4_K_S.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q4_K_S.gguf) | Q4_K_S | 4 | 1.74 GB| small, greater quality loss |
78
+ | [Megrez-3B-Instruct-Q5_0.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q5_0.gguf) | Q5_0 | 5 | 2.05 GB| legacy; medium, balanced quality - prefer using Q4_K_M |
79
+ | [Megrez-3B-Instruct-Q5_K_M.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q5_K_M.gguf) | Q5_K_M | 5 | 2.09 GB| large, very low quality loss - recommended |
80
+ | [Megrez-3B-Instruct-Q5_K_S.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q5_K_S.gguf) | Q5_K_S | 5 | 2.05 GB| large, low quality loss - recommended |
81
+ | [Megrez-3B-Instruct-Q6_K.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q6_K.gguf) | Q6_K | 6 | 2.40 GB| very large, extremely low quality loss |
82
+ | [Megrez-3B-Instruct-Q8_0.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-Q8_0.gguf) | Q8_0 | 8 | 3.10 GB| very large, extremely low quality loss - not recommended |
83
+ | [Megrez-3B-Instruct-f16.gguf](https://huggingface.co/second-state/Megrez-3B-Instruct-GGUF/blob/main/Megrez-3B-Instruct-f16.gguf) | f16 | 16 | 5.84 GB| |
84
+
85
  *Quantized with llama.cpp (commit: b92a14a8)*