Hugging Face's logo

legraphista
/
unsloth-Phi-3.5-mini-instruct-IMat-GGUF

Text Generation
GGUF
Transformers
multilingual
unsloth
phi3
phi
quantized
GGUF
quantization
imat
imatrix
static
16bit
8bit
6bit
5bit
4bit
3bit
2bit
1bit
conversational
Model card Files Community
legraphista commited on
Commit
cae17aa
•
1 Parent(s): 71011fd

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +167 -0
README.md ADDED
@@ -0,0 +1,167 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: unsloth/Phi-3.5-mini-instruct
3
+ inference: false
4
+ language:
5
+ - multilingual
6
+ library_name: gguf
7
+ license: mit
8
+ license_link: https://huggingface.co/microsoft/Phi-3.5-mini-instruct/resolve/main/LICENSE
9
+ pipeline_tag: text-generation
10
+ quantized_by: legraphista
11
+ tags:
12
+ - unsloth
13
+ - transformers
14
+ - phi3
15
+ - phi
16
+ - quantized
17
+ - GGUF
18
+ - quantization
19
+ - imat
20
+ - imatrix
21
+ - static
22
+ - 16bit
23
+ - 8bit
24
+ - 6bit
25
+ - 5bit
26
+ - 4bit
27
+ - 3bit
28
+ - 2bit
29
+ - 1bit
30
+ ---
31
+
32
+ # unsloth-Phi-3.5-mini-instruct-IMat-GGUF
33
+ _Llama.cpp imatrix quantization of unsloth/Phi-3.5-mini-instruct_
34
+
35
+ Original Model: [unsloth/Phi-3.5-mini-instruct](https://huggingface.co/unsloth/Phi-3.5-mini-instruct)
36
+ Original dtype: `BF16` (`bfloat16`)
37
+ Quantized by: llama.cpp [b3620](https://github.com/ggerganov/llama.cpp/releases/tag/b3620)
38
+ IMatrix dataset: [here](https://gist.githubusercontent.com/bartowski1182/eb213dccb3571f863da82e99418f81e8/raw/b2869d80f5c16fd7082594248e80144677736635/calibration_datav3.txt)
39
+
40
+ - [Files](#files)
41
+ - [IMatrix](#imatrix)
42
+ - [Common Quants](#common-quants)
43
+ - [All Quants](#all-quants)
44
+ - [Downloading using huggingface-cli](#downloading-using-huggingface-cli)
45
+ - [Inference](#inference)
46
+ - [Simple chat template](#simple-chat-template)
47
+ - [Chat template with system prompt](#chat-template-with-system-prompt)
48
+ - [Llama.cpp](#llama-cpp)
49
+ - [FAQ](#faq)
50
+ - [Why is the IMatrix not applied everywhere?](#why-is-the-imatrix-not-applied-everywhere)
51
+ - [How do I merge a split GGUF?](#how-do-i-merge-a-split-gguf)
52
+
53
+ ---
54
+
55
+ ## Files
56
+
57
+ ### IMatrix
58
+ Status: ⏳ Processing
59
+ Link: [here](https://huggingface.co/legraphista/unsloth-Phi-3.5-mini-instruct-IMat-GGUF/blob/main/imatrix.dat)
60
+
61
+ ### Common Quants
62
+ | Filename | Quant type | File Size | Status | Uses IMatrix | Is Split |
63
+ | -------- | ---------- | --------- | ------ | ------------ | -------- |
64
+ | unsloth-Phi-3.5-mini-instruct.Q8_0 | Q8_0 | - | ⏳ Processing | ⚪ Static | -
65
+ | unsloth-Phi-3.5-mini-instruct.Q6_K | Q6_K | - | ⏳ Processing | ⚪ Static | -
66
+ | unsloth-Phi-3.5-mini-instruct.Q4_K | Q4_K | - | ⏳ Processing | 🟢 IMatrix | -
67
+ | unsloth-Phi-3.5-mini-instruct.Q3_K | Q3_K | - | ⏳ Processing | 🟢 IMatrix | -
68
+ | unsloth-Phi-3.5-mini-instruct.Q2_K | Q2_K | - | ⏳ Processing | 🟢 IMatrix | -
69
+
70
+
71
+ ### All Quants
72
+ | Filename | Quant type | File Size | Status | Uses IMatrix | Is Split |
73
+ | -------- | ---------- | --------- | ------ | ------------ | -------- |
74
+ | unsloth-Phi-3.5-mini-instruct.BF16 | BF16 | - | ⏳ Processing | ⚪ Static | -
75
+ | unsloth-Phi-3.5-mini-instruct.FP16 | F16 | - | ⏳ Processing | ⚪ Static | -
76
+ | unsloth-Phi-3.5-mini-instruct.Q8_0 | Q8_0 | - | ⏳ Processing | ⚪ Static | -
77
+ | unsloth-Phi-3.5-mini-instruct.Q6_K | Q6_K | - | ⏳ Processing | ⚪ Static | -
78
+ | unsloth-Phi-3.5-mini-instruct.Q5_K | Q5_K | - | ⏳ Processing | ⚪ Static | -
79
+ | unsloth-Phi-3.5-mini-instruct.Q5_K_S | Q5_K_S | - | ⏳ Processing | ⚪ Static | -
80
+ | unsloth-Phi-3.5-mini-instruct.Q4_K | Q4_K | - | ⏳ Processing | 🟢 IMatrix | -
81
+ | unsloth-Phi-3.5-mini-instruct.Q4_K_S | Q4_K_S | - | ⏳ Processing | 🟢 IMatrix | -
82
+ | unsloth-Phi-3.5-mini-instruct.IQ4_NL | IQ4_NL | - | ⏳ Processing | 🟢 IMatrix | -
83
+ | unsloth-Phi-3.5-mini-instruct.IQ4_XS | IQ4_XS | - | ⏳ Processing | 🟢 IMatrix | -
84
+ | unsloth-Phi-3.5-mini-instruct.Q3_K | Q3_K | - | ⏳ Processing | 🟢 IMatrix | -
85
+ | unsloth-Phi-3.5-mini-instruct.Q3_K_L | Q3_K_L | - | ⏳ Processing | 🟢 IMatrix | -
86
+ | unsloth-Phi-3.5-mini-instruct.Q3_K_S | Q3_K_S | - | ⏳ Processing | 🟢 IMatrix | -
87
+ | unsloth-Phi-3.5-mini-instruct.IQ3_M | IQ3_M | - | ⏳ Processing | 🟢 IMatrix | -
88
+ | unsloth-Phi-3.5-mini-instruct.IQ3_S | IQ3_S | - | ⏳ Processing | 🟢 IMatrix | -
89
+ | unsloth-Phi-3.5-mini-instruct.IQ3_XS | IQ3_XS | - | ⏳ Processing | 🟢 IMatrix | -
90
+ | unsloth-Phi-3.5-mini-instruct.IQ3_XXS | IQ3_XXS | - | ⏳ Processing | 🟢 IMatrix | -
91
+ | unsloth-Phi-3.5-mini-instruct.Q2_K | Q2_K | - | ⏳ Processing | 🟢 IMatrix | -
92
+ | unsloth-Phi-3.5-mini-instruct.Q2_K_S | Q2_K_S | - | ⏳ Processing | 🟢 IMatrix | -
93
+ | unsloth-Phi-3.5-mini-instruct.IQ2_M | IQ2_M | - | ⏳ Processing | 🟢 IMatrix | -
94
+ | unsloth-Phi-3.5-mini-instruct.IQ2_S | IQ2_S | - | ⏳ Processing | 🟢 IMatrix | -
95
+ | unsloth-Phi-3.5-mini-instruct.IQ2_XS | IQ2_XS | - | ⏳ Processing | 🟢 IMatrix | -
96
+ | unsloth-Phi-3.5-mini-instruct.IQ2_XXS | IQ2_XXS | - | ⏳ Processing | 🟢 IMatrix | -
97
+ | unsloth-Phi-3.5-mini-instruct.IQ1_M | IQ1_M | - | ⏳ Processing | 🟢 IMatrix | -
98
+ | unsloth-Phi-3.5-mini-instruct.IQ1_S | IQ1_S | - | ⏳ Processing | 🟢 IMatrix | -
99
+
100
+
101
+ ## Downloading using huggingface-cli
102
+ If you do not have hugginface-cli installed:
103
+ ```
104
+ pip install -U "huggingface_hub[cli]"
105
+ ```
106
+ Download the specific file you want:
107
+ ```
108
+ huggingface-cli download legraphista/unsloth-Phi-3.5-mini-instruct-IMat-GGUF --include "unsloth-Phi-3.5-mini-instruct.Q8_0.gguf" --local-dir ./
109
+ ```
110
+ If the model file is big, it has been split into multiple files. In order to download them all to a local folder, run:
111
+ ```
112
+ huggingface-cli download legraphista/unsloth-Phi-3.5-mini-instruct-IMat-GGUF --include "unsloth-Phi-3.5-mini-instruct.Q8_0/*" --local-dir ./
113
+ # see FAQ for merging GGUF's
114
+ ```
115
+
116
+ ---
117
+
118
+ ## Inference
119
+
120
+ ### Simple chat template
121
+ ```
122
+ <|user|>
123
+ {user_prompt}<|end|>
124
+ <|assistant|>
125
+ {assistant_response}<|end|>
126
+ <|user|>
127
+ {next_user_prompt}<|end|>
128
+ <|endoftext|>
129
+ ```
130
+
131
+ ### Chat template with system prompt
132
+ ```
133
+ <|system|>
134
+ {system_prompt}<|end|>
135
+ <|user|>
136
+ {user_prompt}<|end|>
137
+ <|assistant|>
138
+ {assistant_response}<|end|>
139
+ <|user|>
140
+ {next_user_prompt}<|end|>
141
+ <|endoftext|>
142
+ ```
143
+
144
+ ### Llama.cpp
145
+ ```
146
+ llama.cpp/main -m unsloth-Phi-3.5-mini-instruct.Q8_0.gguf --color -i -p "prompt here (according to the chat template)"
147
+ ```
148
+
149
+ ---
150
+
151
+ ## FAQ
152
+
153
+ ### Why is the IMatrix not applied everywhere?
154
+ 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).
155
+
156
+ ### How do I merge a split GGUF?
157
+ 1. Make sure you have `gguf-split` available
158
+ - To get hold of `gguf-split`, navigate to https://github.com/ggerganov/llama.cpp/releases
159
+ - Download the appropriate zip for your system from the latest release
160
+ - Unzip the archive and you should be able to find `gguf-split`
161
+ 2. Locate your GGUF chunks folder (ex: `unsloth-Phi-3.5-mini-instruct.Q8_0`)
162
+ 3. Run `gguf-split --merge unsloth-Phi-3.5-mini-instruct.Q8_0/unsloth-Phi-3.5-mini-instruct.Q8_0-00001-of-XXXXX.gguf unsloth-Phi-3.5-mini-instruct.Q8_0.gguf`
163
+ - Make sure to point `gguf-split` to the first chunk of the split.
164
+
165
+ ---
166
+
167
+ Got a suggestion? Ping me [@legraphista](https://x.com/legraphista)!