Hugging Face's logo

legraphista
/
internlm2-math-plus-1_8b-IMat-GGUF

Text Generation
GGUF
English
Chinese
math
quantized
GGUF
imatrix
quantization
imat
static
conversational
Model card Files Community
legraphista commited on
Commit
34acb01
•
1 Parent(s): 48aac75

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +156 -0
README.md ADDED
@@ -0,0 +1,156 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: internlm/internlm2-math-plus-1_8b
3
+ inference: false
4
+ language:
5
+ - en
6
+ - zh
7
+ library_name: gguf
8
+ license: other
9
+ pipeline_tag: text-generation
10
+ quantized_by: legraphista
11
+ tags:
12
+ - quantized
13
+ - GGUF
14
+ - imatrix
15
+ - quantization
16
+ - imat
17
+ - imatrix
18
+ - static
19
+ ---
20
+
21
+ # internlm2-math-plus-1_8b-IMat-GGUF
22
+ _Llama.cpp imatrix quantization of internlm/internlm2-math-plus-1_8b_
23
+
24
+ Original Model: [internlm/internlm2-math-plus-1_8b](https://huggingface.co/internlm/internlm2-math-plus-1_8b)
25
+ Original dtype: `BF16` (`bfloat16`)
26
+ Quantized by: llama.cpp [b3008](https://github.com/ggerganov/llama.cpp/releases/tag/b3008)
27
+ IMatrix dataset: [here](https://gist.githubusercontent.com/legraphista/d6d93f1a254bcfc58e0af3777eaec41e/raw/d380e7002cea4a51c33fffd47db851942754e7cc/imatrix.calibration.medium.raw)
28
+
29
+ - [internlm2-math-plus-1_8b-IMat-GGUF](#internlm2-math-plus-1-8b-imat-gguf)
30
+ - [Files](#files)
31
+ - [IMatrix](#imatrix)
32
+ - [Common Quants](#common-quants)
33
+ - [All Quants](#all-quants)
34
+ - [Downloading using huggingface-cli](#downloading-using-huggingface-cli)
35
+ - [Inference](#inference)
36
+ - [Simple chat template](#simple-chat-template)
37
+ - [Chat template with system prompt](#chat-template-with-system-prompt)
38
+ - [Llama.cpp](#llama-cpp)
39
+ - [FAQ](#faq)
40
+ - [Why is the IMatrix not applied everywhere?](#why-is-the-imatrix-not-applied-everywhere)
41
+ - [How do I merge a split GGUF?](#how-do-i-merge-a-split-gguf)
42
+
43
+ ---
44
+
45
+ ## Files
46
+
47
+ ### IMatrix
48
+ Status: ⏳ Processing
49
+ Link: [here](https://huggingface.co/legraphista/internlm2-math-plus-1_8b-IMat-GGUF/blob/main/imatrix.dat)
50
+
51
+ ### Common Quants
52
+ | Filename | Quant type | File Size | Status | Uses IMatrix | Is Split |
53
+ | -------- | ---------- | --------- | ------ | ------------ | -------- |
54
+ | internlm2-math-plus-1_8b.Q8_0 | Q8_0 | - | ⏳ Processing | ⚪ Static | -
55
+ | internlm2-math-plus-1_8b.Q6_K | Q6_K | - | ⏳ Processing | ⚪ Static | -
56
+ | internlm2-math-plus-1_8b.Q4_K | Q4_K | - | ⏳ Processing | 🟢 IMatrix | -
57
+ | internlm2-math-plus-1_8b.Q3_K | Q3_K | - | ⏳ Processing | 🟢 IMatrix | -
58
+ | internlm2-math-plus-1_8b.Q2_K | Q2_K | - | ⏳ Processing | 🟢 IMatrix | -
59
+
60
+
61
+ ### All Quants
62
+ | Filename | Quant type | File Size | Status | Uses IMatrix | Is Split |
63
+ | -------- | ---------- | --------- | ------ | ------------ | -------- |
64
+ | internlm2-math-plus-1_8b.FP16 | F16 | - | ⏳ Processing | ⚪ Static | -
65
+ | internlm2-math-plus-1_8b.BF16 | BF16 | - | ⏳ Processing | ⚪ Static | -
66
+ | internlm2-math-plus-1_8b.Q5_K | Q5_K | - | ⏳ Processing | ⚪ Static | -
67
+ | internlm2-math-plus-1_8b.Q5_K_S | Q5_K_S | - | ⏳ Processing | ⚪ Static | -
68
+ | internlm2-math-plus-1_8b.Q4_K_S | Q4_K_S | - | ⏳ Processing | 🟢 IMatrix | -
69
+ | internlm2-math-plus-1_8b.Q3_K_L | Q3_K_L | - | ⏳ Processing | 🟢 IMatrix | -
70
+ | internlm2-math-plus-1_8b.Q3_K_S | Q3_K_S | - | ⏳ Processing | 🟢 IMatrix | -
71
+ | internlm2-math-plus-1_8b.Q2_K_S | Q2_K_S | - | ⏳ Processing | 🟢 IMatrix | -
72
+ | internlm2-math-plus-1_8b.IQ4_NL | IQ4_NL | - | ⏳ Processing | 🟢 IMatrix | -
73
+ | internlm2-math-plus-1_8b.IQ4_XS | IQ4_XS | - | ⏳ Processing | 🟢 IMatrix | -
74
+ | internlm2-math-plus-1_8b.IQ3_M | IQ3_M | - | ⏳ Processing | 🟢 IMatrix | -
75
+ | internlm2-math-plus-1_8b.IQ3_S | IQ3_S | - | ⏳ Processing | 🟢 IMatrix | -
76
+ | internlm2-math-plus-1_8b.IQ3_XS | IQ3_XS | - | ⏳ Processing | 🟢 IMatrix | -
77
+ | internlm2-math-plus-1_8b.IQ3_XXS | IQ3_XXS | - | ⏳ Processing | 🟢 IMatrix | -
78
+ | internlm2-math-plus-1_8b.IQ2_M | IQ2_M | - | ⏳ Processing | 🟢 IMatrix | -
79
+ | internlm2-math-plus-1_8b.IQ2_S | IQ2_S | - | ⏳ Processing | 🟢 IMatrix | -
80
+ | internlm2-math-plus-1_8b.IQ2_XS | IQ2_XS | - | ⏳ Processing | 🟢 IMatrix | -
81
+ | internlm2-math-plus-1_8b.IQ2_XXS | IQ2_XXS | - | ⏳ Processing | 🟢 IMatrix | -
82
+ | internlm2-math-plus-1_8b.IQ1_M | IQ1_M | - | ⏳ Processing | 🟢 IMatrix | -
83
+ | internlm2-math-plus-1_8b.IQ1_S | IQ1_S | - | ⏳ Processing | 🟢 IMatrix | -
84
+
85
+
86
+ ## Downloading using huggingface-cli
87
+ If you do not have hugginface-cli installed:
88
+ ```
89
+ pip install -U "huggingface_hub[cli]"
90
+ ```
91
+ Download the specific file you want:
92
+ ```
93
+ huggingface-cli download legraphista/internlm2-math-plus-1_8b-IMat-GGUF --include "internlm2-math-plus-1_8b.Q8_0.gguf" --local-dir ./
94
+ ```
95
+ If the model file is big, it has been split into multiple files. In order to download them all to a local folder, run:
96
+ ```
97
+ huggingface-cli download legraphista/internlm2-math-plus-1_8b-IMat-GGUF --include "internlm2-math-plus-1_8b.Q8_0/*" --local-dir internlm2-math-plus-1_8b.Q8_0
98
+ # see FAQ for merging GGUF's
99
+ ```
100
+
101
+ ---
102
+
103
+ ## Inference
104
+
105
+ ### Simple chat template
106
+ ```
107
+ <s><|im_start|>user
108
+ Can you provide ways to eat combinations of bananas and dragonfruits?<|im_end|>
109
+ <|im_start|>assistant
110
+ Sure! Here are some ways to eat bananas and dragonfruits together:
111
+ 1. Banana and dragonfruit smoothie: Blend bananas and dragonfruits together with some milk and honey.
112
+ 2. Banana and dragonfruit salad: Mix sliced bananas and dragonfruits together with some lemon juice and honey.<|im_end|>
113
+ <|im_start|>user
114
+ What about solving an 2x + 3 = 7 equation?<|im_end|>
115
+
116
+ ```
117
+
118
+ ### Chat template with system prompt
119
+ ```
120
+ <s><|im_start|>system
121
+ You are a helpful AI.<|im_end|>
122
+ <|im_start|>user
123
+ Can you provide ways to eat combinations of bananas and dragonfruits?<|im_end|>
124
+ <|im_start|>assistant
125
+ Sure! Here are some ways to eat bananas and dragonfruits together:
126
+ 1. Banana and dragonfruit smoothie: Blend bananas and dragonfruits together with some milk and honey.
127
+ 2. Banana and dragonfruit salad: Mix sliced bananas and dragonfruits together with some lemon juice and honey.<|im_end|>
128
+ <|im_start|>user
129
+ What about solving an 2x + 3 = 7 equation?<|im_end|>
130
+
131
+ ```
132
+
133
+ ### Llama.cpp
134
+ ```
135
+ llama.cpp/main -m internlm2-math-plus-1_8b.Q8_0.gguf --color -i -p "prompt here (according to the chat template)"
136
+ ```
137
+
138
+ ---
139
+
140
+ ## FAQ
141
+
142
+ ### Why is the IMatrix not applied everywhere?
143
+ According to [this investigation](https://www.reddit.com/r/LocalLLaMA/comments/1993iro/ggufs_quants_can_punch_above_their_weights_now/), it appears that lower quantizations are the only ones that benefit from the imatrix input (as per hellaswag results).
144
+
145
+ ### How do I merge a split GGUF?
146
+ 1. Make sure you have `gguf-split` available
147
+ - To get hold of `gguf-split`, navigate to https://github.com/ggerganov/llama.cpp/releases
148
+ - Download the appropriate zip for your system from the latest release
149
+ - Unzip the archive and you should be able to find `gguf-split`
150
+ 2. Locate your GGUF chunks folder (ex: `internlm2-math-plus-1_8b.Q8_0`)
151
+ 3. Run `gguf-split --merge internlm2-math-plus-1_8b.Q8_0/internlm2-math-plus-1_8b.Q8_0-00001-of-XXXXX.gguf internlm2-math-plus-1_8b.Q8_0.gguf`
152
+ - Make sure to point `gguf-split` to the first chunk of the split.
153
+
154
+ ---
155
+
156
+ Got a suggestion? Ping me [@legraphista](https://x.com/legraphista)!