bartowski commited on
Commit
599b2dd
1 Parent(s): da3cd2b

Llamacpp quants

Browse files
.gitattributes CHANGED
@@ -33,3 +33,26 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ gemma-2-9b-it-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text
37
+ gemma-2-9b-it-IQ2_S.gguf filter=lfs diff=lfs merge=lfs -text
38
+ gemma-2-9b-it-IQ2_XS.gguf filter=lfs diff=lfs merge=lfs -text
39
+ gemma-2-9b-it-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text
40
+ gemma-2-9b-it-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text
41
+ gemma-2-9b-it-IQ3_XXS.gguf filter=lfs diff=lfs merge=lfs -text
42
+ gemma-2-9b-it-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text
43
+ gemma-2-9b-it-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text
44
+ gemma-2-9b-it-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text
45
+ gemma-2-9b-it-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text
46
+ gemma-2-9b-it-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text
47
+ gemma-2-9b-it-Q4_K_L.gguf filter=lfs diff=lfs merge=lfs -text
48
+ gemma-2-9b-it-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
49
+ gemma-2-9b-it-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text
50
+ gemma-2-9b-it-Q5_K_L.gguf filter=lfs diff=lfs merge=lfs -text
51
+ gemma-2-9b-it-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text
52
+ gemma-2-9b-it-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text
53
+ gemma-2-9b-it-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text
54
+ gemma-2-9b-it-Q6_K_L.gguf filter=lfs diff=lfs merge=lfs -text
55
+ gemma-2-9b-it-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text
56
+ gemma-2-9b-it-Q8_0_L.gguf filter=lfs diff=lfs merge=lfs -text
57
+ gemma-2-9b-it-f32.gguf filter=lfs diff=lfs merge=lfs -text
58
+ gemma-2-9b-it.imatrix filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,108 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: gemma
3
+ library_name: transformers
4
+ pipeline_tag: text-generation
5
+ extra_gated_heading: Access Gemma on Hugging Face
6
+ extra_gated_prompt: >-
7
+ To access Gemma on Hugging Face, you’re required to review and agree to
8
+ Google’s usage license. To do this, please ensure you’re logged in to Hugging
9
+ Face and click below. Requests are processed immediately.
10
+ extra_gated_button_content: Acknowledge license
11
+ tags:
12
+ - conversational
13
+ quantized_by: bartowski
14
+ ---
15
+
16
+ ## Llamacpp imatrix Quantizations of gemma-2-9b-it
17
+
18
+ Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> PR <a href="https://github.com/ggerganov/llama.cpp/pull/8156">8156</a> for quantization.
19
+
20
+ Original model: https://huggingface.co/google/gemma-2-9b-it
21
+
22
+ All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)
23
+
24
+ ## Prompt format
25
+
26
+ ```
27
+ <bos><start_of_turn>user
28
+ {prompt}<end_of_turn>
29
+ <start_of_turn>model
30
+
31
+ ```
32
+
33
+ Note that this model does not support a System prompt.
34
+
35
+ ## Download a file (not the whole branch) from below:
36
+
37
+ | Filename | Quant type | File Size | Description |
38
+ | -------- | ---------- | --------- | ----------- |
39
+ | [gemma-2-9b-it-Q8_0_L.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q8_1.gguf) | Q8_0_L | 10.68GB | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. Extremely high quality, generally unneeded but max available quant. |
40
+ | [gemma-2-9b-it-Q8_0.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q8_0.gguf) | Q8_0 | 9.82GB | Extremely high quality, generally unneeded but max available quant. |
41
+ | [gemma-2-9b-it-Q6_K_L.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q6_K_L.gguf) | Q6_K_L | 8.67GB | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. Very high quality, near perfect, *recommended*. |
42
+ | [gemma-2-9b-it-Q6_K.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q6_K.gguf) | Q6_K | 7.58GB | Very high quality, near perfect, *recommended*. |
43
+ | [gemma-2-9b-it-Q5_K_L.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q5_K_L.gguf) | Q5_K_L | 7.73GB | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. High quality, *recommended*. |
44
+ | [gemma-2-9b-it-Q5_K_M.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q5_K_M.gguf) | Q5_K_M | 6.64GB | High quality, *recommended*. |
45
+ | [gemma-2-9b-it-Q5_K_S.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q5_K_S.gguf) | Q5_K_S | 6.48GB | High quality, *recommended*. |
46
+ | [gemma-2-9b-it-Q4_K_L.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q4_K_L.gguf) | Q4_K_L | 6.84GB | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. Good quality, uses about 4.83 bits per weight, *recommended*. |
47
+ | [gemma-2-9b-it-Q4_K_M.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q4_K_M.gguf) | Q4_K_M | 5.76GB | Good quality, uses about 4.83 bits per weight, *recommended*. |
48
+ | [gemma-2-9b-it-Q4_K_S.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q4_K_S.gguf) | Q4_K_S | 5.47GB | Slightly lower quality with more space savings, *recommended*. |
49
+ | [gemma-2-9b-it-IQ4_XS.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-IQ4_XS.gguf) | IQ4_XS | 5.18GB | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
50
+ | [gemma-2-9b-it-Q3_K_XL.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF//main/gemma-2-9b-it-Q3_K_XL.gguf) | Q3_K_XL | | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. Lower quality but usable, good for low RAM availability. |
51
+ | [gemma-2-9b-it-Q3_K_L.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q3_K_L.gguf) | Q3_K_L | 5.13GB | Lower quality but usable, good for low RAM availability. |
52
+ | [gemma-2-9b-it-Q3_K_M.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q3_K_M.gguf) | Q3_K_M | 4.76GB | Even lower quality. |
53
+ | [gemma-2-9b-it-IQ3_M.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-IQ3_M.gguf) | IQ3_M | 4.49GB | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
54
+ | [gemma-2-9b-it-Q3_K_S.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q3_K_S.gguf) | Q3_K_S | 4.33GB | Low quality, not recommended. |
55
+ | [gemma-2-9b-it-IQ3_XS.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-IQ3_XS.gguf) | IQ3_XS | 4.14GB | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
56
+ | [gemma-2-9b-it-IQ3_XXS.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-IQ3_XXS.gguf) | IQ3_XXS | 3.79GB | Lower quality, new method with decent performance, comparable to Q3 quants. |
57
+ | [gemma-2-9b-it-Q2_K.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-Q2_K.gguf) | Q2_K | 3.80GB | Very low quality but surprisingly usable. |
58
+ | [gemma-2-9b-it-IQ2_M.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-IQ2_M.gguf) | IQ2_M | 3.43GB | Very low quality, uses SOTA techniques to also be surprisingly usable. |
59
+ | [gemma-2-9b-it-IQ2_S.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-IQ2_S.gguf) | IQ2_S | 3.21GB | Very low quality, uses SOTA techniques to be usable. |
60
+ | [gemma-2-9b-it-IQ2_XS.gguf](https://huggingface.co/bartowski/gemma-2-9b-it-GGUF/blob/main/gemma-2-9b-it-IQ2_XS.gguf) | IQ2_XS | 3.06GB | Very low quality, uses SOTA techniques to be usable. |
61
+
62
+ ## Downloading using huggingface-cli
63
+
64
+ First, make sure you have hugginface-cli installed:
65
+
66
+ ```
67
+ pip install -U "huggingface_hub[cli]"
68
+ ```
69
+
70
+ Then, you can target the specific file you want:
71
+
72
+ ```
73
+ huggingface-cli download bartowski/gemma-2-9b-it-GGUF --include "gemma-2-9b-it-Q4_K_M.gguf" --local-dir ./
74
+ ```
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
+ ```
79
+ huggingface-cli download bartowski/gemma-2-9b-it-GGUF --include "gemma-2-9b-it-Q8_0.gguf/*" --local-dir gemma-2-9b-it-Q8_0
80
+ ```
81
+
82
+ You can either specify a new local-dir (gemma-2-9b-it-Q8_0) or download them all in place (./)
83
+
84
+ ## Which file should I choose?
85
+
86
+ A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
87
+
88
+ 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.
89
+
90
+ 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.
91
+
92
+ 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.
93
+
94
+ Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
95
+
96
+ 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.
97
+
98
+ If you want to get more into the weeds, you can check out this extremely useful feature chart:
99
+
100
+ [llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)
101
+
102
+ 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.
103
+
104
+ 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.
105
+
106
+ 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.
107
+
108
+ Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
gemma-2-9b-it-IQ2_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94b6fe40ee612eb50159ad3e6fc97cc04c88ec7af8a813246d8d8c27c4522a00
3
+ size 3434988384
gemma-2-9b-it-IQ2_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee80dfba6b4b7aee3528d742526d984496e621640d3c75eb521a36afed2c6010
3
+ size 3211805536
gemma-2-9b-it-IQ2_XS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f3ae196ff0f2a34927db3ed93451575394fb47e0a843f5d7915f2a9da02e7fd
3
+ size 3067700064
gemma-2-9b-it-IQ3_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:744e3ead642fcbde768639cf2532f78da5b8ca80ce1718fed712c0ab38a5ebda
3
+ size 4494995808
gemma-2-9b-it-IQ3_XS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:294638961d516cbc520c0dcd6a143fe0226f9969ee483e3e0d89b0f6f3587e4d
3
+ size 4145369440
gemma-2-9b-it-IQ3_XXS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79dfb350e10a4c7598b957c00381b7a8459f64e99457d4d05203dc2528683824
3
+ size 3797058400
gemma-2-9b-it-IQ4_XS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8817cc2720f51f1130d994fcacd58b8860c88882d10bbffb29ed4b5f62f155b
3
+ size 5183410528
gemma-2-9b-it-Q2_K.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c3d884107e4d01bac8316cbe755137beac8988cc9c8a9c8806b61f0c4d8bcc5
3
+ size 3805778272
gemma-2-9b-it-Q3_K_L.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d97e40aef60b072b15d9bdd41b171f141f46b45a18eaee257d9bf91daea52e1f
3
+ size 5132833120
gemma-2-9b-it-Q3_K_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:066d5cd0c5ad7a8d2b7407fb53f56858e6736c8248a06111b5f65afcbc709c8f
3
+ size 4762161504
gemma-2-9b-it-Q3_K_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a0dc0e74df5cd76fb134249f4858bd4610049111482daa6dc1d8479ff283f21
3
+ size 4338045280
gemma-2-9b-it-Q4_K_L.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05dfbfc607fce0d039cac53098c387eb66c9c17587eb8bbf8b400c4769ae0252
3
+ size 6844347232
gemma-2-9b-it-Q4_K_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0874bf61be2e4b3d0a4a75e58fbd442dc410745d513c1e1e5de0b54ae33e65db
3
+ size 5761438048
gemma-2-9b-it-Q4_K_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ae8828f11ad498fa6fb0c31ccbd5f6ec0bc3b579613f164b8bebc663f9d1763
3
+ size 5479305568
gemma-2-9b-it-Q5_K_L.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:381bc6455ee6bd33dfa3987d001678e390019e214a6f3e7fb17e22853a0102da
3
+ size 7730656096
gemma-2-9b-it-Q5_K_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd2ef3823778e3c138aebb5892fcc4587313549ec3c21b84103090ce4b7617b4
3
+ size 6647746912
gemma-2-9b-it-Q5_K_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ba2748d33407c999dc1c2fe6845e39eb686792cb5e42d5d1bc20599cfb2bdb7
3
+ size 6483972448
gemma-2-9b-it-Q6_K.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f38ef3fe66173a34e05b7f356fa0880b3941e64c94618623f4c243e1a4fde12
3
+ size 7589450080
gemma-2-9b-it-Q6_K_L.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d69cac05ff85f04207610e19def2e741d0fda43c19e0bf64cc7dfef9f005a2ef
3
+ size 8672359264
gemma-2-9b-it-Q8_0.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68ef14eb96fb6ab0dfb04dc9d73e4b5ff42fda0011110a160881d20b3d84b594
3
+ size 9827640160
gemma-2-9b-it-Q8_0_L.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c2c8fe89751b1be794ce20ae92a764b3fe7fdfad735e5402650b5e40ad5788c
3
+ size 10688230240
gemma-2-9b-it-f32.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f7d9c68f33c338a0d98faf6bad67cb13d39a5d4f3c87965bff4d62620d27d70
3
+ size 36974719488
gemma-2-9b-it.imatrix ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ebc87750f8830146ae668032ca7e319bfbf89e4276c1a514aeee1be9e6addfd
3
+ size 6116901