Hugging Face's logo

bartowski
/
granite-3.0-2b-instruct-GGUF

Text Generation
GGUF
language
granite-3.0
Eval Results
imatrix
conversational
Model card Files Community
granite-3.0-2b-instruct-GGUF
File size: 12,599 Bytes 
33f6cd9
ca09067
 
33f6cd9
ca09067
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33f6cd9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
 
---
base_model: ibm-granite/granite-3.0-2b-instruct
license: apache-2.0
pipeline_tag: text-generation
tags:
- language
- granite-3.0
quantized_by: bartowski
inference: false
model-index:
- name: granite-3.0-2b-instruct
  results:
  - task:
      type: text-generation
    dataset:
      name: IFEval
      type: instruction-following
    metrics:
    - type: pass@1
      value: 46.07
      name: pass@1
    - type: pass@1
      value: 7.66
      name: pass@1
  - task:
      type: text-generation
    dataset:
      name: AGI-Eval
      type: human-exams
    metrics:
    - type: pass@1
      value: 29.75
      name: pass@1
    - type: pass@1
      value: 56.03
      name: pass@1
    - type: pass@1
      value: 27.92
      name: pass@1
  - task:
      type: text-generation
    dataset:
      name: OBQA
      type: commonsense
    metrics:
    - type: pass@1
      value: 43.2
      name: pass@1
    - type: pass@1
      value: 66.36
      name: pass@1
    - type: pass@1
      value: 76.79
      name: pass@1
    - type: pass@1
      value: 71.9
      name: pass@1
    - type: pass@1
      value: 53.37
      name: pass@1
  - task:
      type: text-generation
    dataset:
      name: BoolQ
      type: reading-comprehension
    metrics:
    - type: pass@1
      value: 84.89
      name: pass@1
    - type: pass@1
      value: 19.73
      name: pass@1
  - task:
      type: text-generation
    dataset:
      name: ARC-C
      type: reasoning
    metrics:
    - type: pass@1
      value: 54.35
      name: pass@1
    - type: pass@1
      value: 28.61
      name: pass@1
    - type: pass@1
      value: 43.74
      name: pass@1
  - task:
      type: text-generation
    dataset:
      name: HumanEvalSynthesis
      type: code
    metrics:
    - type: pass@1
      value: 50.61
      name: pass@1
    - type: pass@1
      value: 45.58
      name: pass@1
    - type: pass@1
      value: 51.83
      name: pass@1
    - type: pass@1
      value: 41.0
      name: pass@1
  - task:
      type: text-generation
    dataset:
      name: GSM8K
      type: math
    metrics:
    - type: pass@1
      value: 59.66
      name: pass@1
    - type: pass@1
      value: 23.66
      name: pass@1
  - task:
      type: text-generation
    dataset:
      name: PAWS-X (7 langs)
      type: multilingual
    metrics:
    - type: pass@1
      value: 61.42
      name: pass@1
    - type: pass@1
      value: 37.13
      name: pass@1
---

## Llamacpp imatrix Quantizations of granite-3.0-2b-instruct

Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b3930">b3930</a> for quantization.

Original model: https://huggingface.co/ibm-granite/granite-3.0-2b-instruct

All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)

Run them in [LM Studio](https://lmstudio.ai/)

## Prompt format

```
<|start_of_role|>system<|end_of_role|>{system_prompt}<|end_of_text|>
<|start_of_role|>user<|end_of_role|>{prompt}<|end_of_text|>
<|start_of_role|>assistant<|end_of_role|>
```

## Download a file (not the whole branch) from below:

| Filename | Quant type | File Size | Split | Description |
| -------- | ---------- | --------- | ----- | ----------- |
| [granite-3.0-2b-instruct-f16.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-f16.gguf) | f16 | 5.27GB | false | Full F16 weights. |
| [granite-3.0-2b-instruct-Q8_0.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q8_0.gguf) | Q8_0 | 2.80GB | false | Extremely high quality, generally unneeded but max available quant. |
| [granite-3.0-2b-instruct-Q6_K_L.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q6_K_L.gguf) | Q6_K_L | 2.21GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. |
| [granite-3.0-2b-instruct-Q6_K.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q6_K.gguf) | Q6_K | 2.16GB | false | Very high quality, near perfect, *recommended*. |
| [granite-3.0-2b-instruct-Q5_K_L.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q5_K_L.gguf) | Q5_K_L | 1.94GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. |
| [granite-3.0-2b-instruct-Q5_K_M.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q5_K_M.gguf) | Q5_K_M | 1.87GB | false | High quality, *recommended*. |
| [granite-3.0-2b-instruct-Q5_K_S.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q5_K_S.gguf) | Q5_K_S | 1.83GB | false | High quality, *recommended*. |
| [granite-3.0-2b-instruct-Q4_K_L.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q4_K_L.gguf) | Q4_K_L | 1.68GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. |
| [granite-3.0-2b-instruct-Q4_K_M.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q4_K_M.gguf) | Q4_K_M | 1.60GB | false | Good quality, default size for must use cases, *recommended*. |
| [granite-3.0-2b-instruct-Q4_K_S.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q4_K_S.gguf) | Q4_K_S | 1.52GB | false | Slightly lower quality with more space savings, *recommended*. |
| [granite-3.0-2b-instruct-Q4_0.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q4_0.gguf) | Q4_0 | 1.52GB | false | Legacy format, generally not worth using over similarly sized formats |
| [granite-3.0-2b-instruct-Q4_0_8_8.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q4_0_8_8.gguf) | Q4_0_8_8 | 1.51GB | false | Optimized for ARM inference. Requires 'sve' support (see link below). *Don't use on Mac or Windows*. |
| [granite-3.0-2b-instruct-Q4_0_4_8.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q4_0_4_8.gguf) | Q4_0_4_8 | 1.51GB | false | Optimized for ARM inference. Requires 'i8mm' support (see link below). *Don't use on Mac or Windows*. |
| [granite-3.0-2b-instruct-Q4_0_4_4.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q4_0_4_4.gguf) | Q4_0_4_4 | 1.51GB | false | Optimized for ARM inference. Should work well on all ARM chips, pick this if you're unsure. *Don't use on Mac or Windows*. |
| [granite-3.0-2b-instruct-Q3_K_XL.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q3_K_XL.gguf) | Q3_K_XL | 1.49GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
| [granite-3.0-2b-instruct-IQ4_XS.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-IQ4_XS.gguf) | IQ4_XS | 1.44GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
| [granite-3.0-2b-instruct-Q3_K_L.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q3_K_L.gguf) | Q3_K_L | 1.40GB | false | Lower quality but usable, good for low RAM availability. |
| [granite-3.0-2b-instruct-IQ3_M.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-IQ3_M.gguf) | IQ3_M | 1.21GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
| [granite-3.0-2b-instruct-Q3_K_S.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q3_K_S.gguf) | Q3_K_S | 1.17GB | false | Low quality, not recommended. |
| [granite-3.0-2b-instruct-IQ3_XS.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-IQ3_XS.gguf) | IQ3_XS | 1.12GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
| [granite-3.0-2b-instruct-Q2_K_L.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q2_K_L.gguf) | Q2_K_L | 1.11GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
| [granite-3.0-2b-instruct-IQ3_XXS.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-IQ3_XXS.gguf) | IQ3_XXS | 1.05GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. |
| [granite-3.0-2b-instruct-Q2_K.gguf](https://huggingface.co/bartowski/granite-3.0-2b-instruct-GGUF/blob/main/granite-3.0-2b-instruct-Q2_K.gguf) | Q2_K | 1.01GB | false | Very low quality but surprisingly usable. |

## Embed/output weights

Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.

Some say that this improves the quality, others don't notice any difference. If you use these models PLEASE COMMENT with your findings. I would like feedback that these are actually used and useful so I don't keep uploading quants no one is using.

Thanks!

## Downloading using huggingface-cli

First, make sure you have hugginface-cli installed:

```
pip install -U "huggingface_hub[cli]"
```

Then, you can target the specific file you want:

```
huggingface-cli download bartowski/granite-3.0-2b-instruct-GGUF --include "granite-3.0-2b-instruct-Q4_K_M.gguf" --local-dir ./
```

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:

```
huggingface-cli download bartowski/granite-3.0-2b-instruct-GGUF --include "granite-3.0-2b-instruct-Q8_0/*" --local-dir ./
```

You can either specify a new local-dir (granite-3.0-2b-instruct-Q8_0) or download them all in place (./)

## Q4_0_X_X

These are *NOT* for Metal (Apple) offloading, only ARM chips.

If you're using an ARM chip, the Q4_0_X_X quants will have a substantial speedup. Check out Q4_0_4_4 speed comparisons [on the original pull request](https://github.com/ggerganov/llama.cpp/pull/5780#pullrequestreview-21657544660)

To check which one would work best for your ARM chip, you can check [AArch64 SoC features](https://gpages.juszkiewicz.com.pl/arm-socs-table/arm-socs.html) (thanks EloyOn!).

## Which file should I choose?

A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)

The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.

If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.

If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.

Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.

If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.

If you want to get more into the weeds, you can check out this extremely useful feature chart:

[llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)

But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.

These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.

The I-quants are *not* compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm.

## Credits

Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset

Thank you ZeroWw for the inspiration to experiment with embed/output

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski