Hugging Face's logo

legraphista
/
Phi-3-mini-4k-instruct-IMat-GGUF

Text Generation
GGUF
English
quantized
GGUF
imatrix
quantization
imat
static
conversational
Model card Files Community
legraphista commited on
Commit
52aa2f4
•
1 Parent(s): f3eac58

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +94 -0
README.md ADDED
@@ -0,0 +1,94 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: microsoft/Phi-3-mini-4k-instruct
3
+ inference: false
4
+ language:
5
+ - en
6
+ library_name: gguf
7
+ license: mit
8
+ license_link: https://huggingface.co/microsoft/Phi-3-mini-4k-instruct/resolve/main/LICENSE
9
+ pipeline_tag: text-generation
10
+ quantized_by: legraphista
11
+ tags:
12
+ - quantized
13
+ - GGUF
14
+ - imatrix
15
+ - quantization
16
+ ---
17
+
18
+ # Phi-3-mini-4k-instruct-GGUF
19
+ _Llama.cpp imatrix quantization of Phi-3-mini-4k-instruct-GGUF_
20
+
21
+ Original Model: [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct)
22
+ Original dtype: `BF16` (`bfloat16`)
23
+ Quantized by: llama.cpp [b2989](https://github.com/ggerganov/llama.cpp/releases/tag/b2989)
24
+ IMatrix dataset: [here](https://gist.githubusercontent.com/legraphista/d6d93f1a254bcfc58e0af3777eaec41e/raw/d380e7002cea4a51c33fffd47db851942754e7cc/imatrix.calibration.medium.raw)
25
+
26
+ ## Files
27
+
28
+ ### IMatrix
29
+ Status: ⏳ Processing
30
+ Link: [here](https://huggingface.co/legraphista/Phi-3-mini-4k-instruct-GGUF/blob/main/imatrix.dat)
31
+
32
+ ### Common Quants
33
+ | Filename | Quant type | File Size | Status | Uses IMatrix | Is Split |
34
+ | -------- | ---------- | --------- | ------ | ------------ | -------- |
35
+ | Phi-3-mini-4k-instruct.Q8_0 | Q8_0 | - | ⏳ Processing | No | -
36
+ | Phi-3-mini-4k-instruct.Q6_K | Q6_K | - | ⏳ Processing | No | -
37
+ | Phi-3-mini-4k-instruct.Q4_K | Q4_K | - | ⏳ Processing | Yes | -
38
+ | Phi-3-mini-4k-instruct.Q3_K | Q3_K | - | ⏳ Processing | Yes | -
39
+ | Phi-3-mini-4k-instruct.Q2_K | Q2_K | - | ⏳ Processing | Yes | -
40
+
41
+
42
+ ### All Quants
43
+ | Filename | Quant type | File Size | Status | Uses IMatrix | Is Split |
44
+ | -------- | ---------- | --------- | ------ | ------------ | -------- |
45
+ | Phi-3-mini-4k-instruct.FP16 | F16 | - | ⏳ Processing | No | -
46
+ | Phi-3-mini-4k-instruct.BF16 | BF16 | - | ⏳ Processing | No | -
47
+ | Phi-3-mini-4k-instruct.Q5_K | Q5_K | - | ⏳ Processing | No | -
48
+ | Phi-3-mini-4k-instruct.Q5_K_S | Q5_K_S | - | ⏳ Processing | No | -
49
+ | Phi-3-mini-4k-instruct.Q4_K_S | Q4_K_S | - | ⏳ Processing | Yes | -
50
+ | Phi-3-mini-4k-instruct.Q3_K_L | Q3_K_L | - | ⏳ Processing | Yes | -
51
+ | Phi-3-mini-4k-instruct.Q3_K_S | Q3_K_S | - | ⏳ Processing | Yes | -
52
+ | Phi-3-mini-4k-instruct.Q2_K_S | Q2_K_S | - | ⏳ Processing | Yes | -
53
+ | Phi-3-mini-4k-instruct.IQ4_NL | IQ4_NL | - | ⏳ Processing | Yes | -
54
+ | Phi-3-mini-4k-instruct.IQ4_XS | IQ4_XS | - | ⏳ Processing | Yes | -
55
+ | Phi-3-mini-4k-instruct.IQ3_M | IQ3_M | - | ⏳ Processing | Yes | -
56
+ | Phi-3-mini-4k-instruct.IQ3_S | IQ3_S | - | ⏳ Processing | Yes | -
57
+ | Phi-3-mini-4k-instruct.IQ3_XS | IQ3_XS | - | ⏳ Processing | Yes | -
58
+ | Phi-3-mini-4k-instruct.IQ3_XXS | IQ3_XXS | - | ⏳ Processing | Yes | -
59
+ | Phi-3-mini-4k-instruct.IQ2_M | IQ2_M | - | ⏳ Processing | Yes | -
60
+ | Phi-3-mini-4k-instruct.IQ2_S | IQ2_S | - | ⏳ Processing | Yes | -
61
+ | Phi-3-mini-4k-instruct.IQ2_XS | IQ2_XS | - | ⏳ Processing | Yes | -
62
+ | Phi-3-mini-4k-instruct.IQ2_XXS | IQ2_XXS | - | ⏳ Processing | Yes | -
63
+ | Phi-3-mini-4k-instruct.IQ1_M | IQ1_M | - | ⏳ Processing | Yes | -
64
+ | Phi-3-mini-4k-instruct.IQ1_S | IQ1_S | - | ⏳ Processing | Yes | -
65
+
66
+
67
+ ## Downloading using huggingface-cli
68
+ First, make sure you have hugginface-cli installed:
69
+ ```
70
+ pip install -U "huggingface_hub[cli]"
71
+ ```
72
+ Then, you can target the specific file you want:
73
+ ```
74
+ huggingface-cli download legraphista/Phi-3-mini-4k-instruct-GGUF --include "Phi-3-mini-4k-instruct.Q8_0.gguf" --local-dir ./
75
+ ```
76
+ If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:
77
+ ```
78
+ huggingface-cli download legraphista/Phi-3-mini-4k-instruct-GGUF --include "Phi-3-mini-4k-instruct.Q8_0/*" --local-dir Phi-3-mini-4k-instruct.Q8_0
79
+ # see FAQ for merging GGUF's
80
+ ```
81
+
82
+ ## FAQ
83
+
84
+ ### Why is the IMatrix not applied everywhere?
85
+ 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).
86
+
87
+ ### How do I merge a split GGUF?
88
+ 1. Make sure you have `gguf-split` available
89
+ - To get hold of `gguf-split`, navigate to https://github.com/ggerganov/llama.cpp/releases
90
+ - Download the appropriate zip for your system from the latest release
91
+ - Unzip the archive and you should be able to find `gguf-split`
92
+ 2. Locate your GGUF chunks folder (ex: `Phi-3-mini-4k-instruct.Q8_0`)
93
+ 3. Run `gguf-split --merge Phi-3-mini-4k-instruct.Q8_0/Phi-3-mini-4k-instruct.Q8_0-00001-of-XXXXX.gguf Phi-3-mini-4k-instruct.Q8_0.gguf`
94
+ - Make sure to point `gguf-split` to the first chunk of the split.