Hugging Face's logo

bartowski
/
CodeQwen1.5-7B-Chat-GGUF

Text Generation
GGUF
English
chat
Inference Endpoints
conversational
Model card Files Community
bartowski commited on
Commit
552b091
•
1 Parent(s): 7fc1284

Llamacpp quants

Browse files
Files changed (25) hide show
  1. .gitattributes +23 -0
  2. CodeQwen1.5-7B-Chat-IQ1_M.gguf +3 -0
  3. CodeQwen1.5-7B-Chat-IQ1_S.gguf +3 -0
  4. CodeQwen1.5-7B-Chat-IQ2_M.gguf +3 -0
  5. CodeQwen1.5-7B-Chat-IQ2_S.gguf +3 -0
  6. CodeQwen1.5-7B-Chat-IQ2_XS.gguf +3 -0
  7. CodeQwen1.5-7B-Chat-IQ2_XXS.gguf +3 -0
  8. CodeQwen1.5-7B-Chat-IQ3_M.gguf +3 -0
  9. CodeQwen1.5-7B-Chat-IQ3_S.gguf +3 -0
  10. CodeQwen1.5-7B-Chat-IQ3_XS.gguf +3 -0
  11. CodeQwen1.5-7B-Chat-IQ3_XXS.gguf +3 -0
  12. CodeQwen1.5-7B-Chat-IQ4_NL.gguf +3 -0
  13. CodeQwen1.5-7B-Chat-IQ4_XS.gguf +3 -0
  14. CodeQwen1.5-7B-Chat-Q2_K.gguf +3 -0
  15. CodeQwen1.5-7B-Chat-Q3_K_L.gguf +3 -0
  16. CodeQwen1.5-7B-Chat-Q3_K_M.gguf +3 -0
  17. CodeQwen1.5-7B-Chat-Q3_K_S.gguf +3 -0
  18. CodeQwen1.5-7B-Chat-Q4_K_M.gguf +3 -0
  19. CodeQwen1.5-7B-Chat-Q4_K_S.gguf +3 -0
  20. CodeQwen1.5-7B-Chat-Q5_K_M.gguf +3 -0
  21. CodeQwen1.5-7B-Chat-Q5_K_S.gguf +3 -0
  22. CodeQwen1.5-7B-Chat-Q6_K.gguf +3 -0
  23. CodeQwen1.5-7B-Chat-Q8_0.gguf +3 -0
  24. CodeQwen1.5-7B-Chat.imatrix +3 -0
  25. README.md +84 -0
.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
+ CodeQwen1.5-7B-Chat-IQ1_M.gguf filter=lfs diff=lfs merge=lfs -text
37
+ CodeQwen1.5-7B-Chat-IQ1_S.gguf filter=lfs diff=lfs merge=lfs -text
38
+ CodeQwen1.5-7B-Chat-IQ2_M.gguf filter=lfs diff=lfs merge=lfs -text
39
+ CodeQwen1.5-7B-Chat-IQ2_S.gguf filter=lfs diff=lfs merge=lfs -text
40
+ CodeQwen1.5-7B-Chat-IQ2_XS.gguf filter=lfs diff=lfs merge=lfs -text
41
+ CodeQwen1.5-7B-Chat-IQ2_XXS.gguf filter=lfs diff=lfs merge=lfs -text
42
+ CodeQwen1.5-7B-Chat-IQ3_M.gguf filter=lfs diff=lfs merge=lfs -text
43
+ CodeQwen1.5-7B-Chat-IQ3_S.gguf filter=lfs diff=lfs merge=lfs -text
44
+ CodeQwen1.5-7B-Chat-IQ3_XS.gguf filter=lfs diff=lfs merge=lfs -text
45
+ CodeQwen1.5-7B-Chat-IQ3_XXS.gguf filter=lfs diff=lfs merge=lfs -text
46
+ CodeQwen1.5-7B-Chat-IQ4_NL.gguf filter=lfs diff=lfs merge=lfs -text
47
+ CodeQwen1.5-7B-Chat-IQ4_XS.gguf filter=lfs diff=lfs merge=lfs -text
48
+ CodeQwen1.5-7B-Chat-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text
49
+ CodeQwen1.5-7B-Chat-Q3_K_L.gguf filter=lfs diff=lfs merge=lfs -text
50
+ CodeQwen1.5-7B-Chat-Q3_K_M.gguf filter=lfs diff=lfs merge=lfs -text
51
+ CodeQwen1.5-7B-Chat-Q3_K_S.gguf filter=lfs diff=lfs merge=lfs -text
52
+ CodeQwen1.5-7B-Chat-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
53
+ CodeQwen1.5-7B-Chat-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text
54
+ CodeQwen1.5-7B-Chat-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text
55
+ CodeQwen1.5-7B-Chat-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text
56
+ CodeQwen1.5-7B-Chat-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text
57
+ CodeQwen1.5-7B-Chat-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text
58
+ CodeQwen1.5-7B-Chat.imatrix filter=lfs diff=lfs merge=lfs -text
CodeQwen1.5-7B-Chat-IQ1_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e40a0f97250db1ba25571e346b4f1b54f2c6e3c8c7c6546d475931897861dbd
3
+ size 2458142464
CodeQwen1.5-7B-Chat-IQ1_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bc61aed0174730f678c860fd3432428ead7a4e35b97dbc0daa3c13b0fbfe5f2
3
+ size 2361411328
CodeQwen1.5-7B-Chat-IQ2_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4d175136ad6f6e0f3ac5a510d905d05aa9e890f87cfae252fd6aca00dd99de8
3
+ size 3008030464
CodeQwen1.5-7B-Chat-IQ2_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4b6f10a8e929d4000517942462f556315a62256ec1577a4d6c53df65f1414b1
3
+ size 2879055616
CodeQwen1.5-7B-Chat-IQ2_XS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:58c2b09bfa93608ed02d673c1053d5aff4abee9829d7af7ec954976885b2458e
3
+ size 2765113088
CodeQwen1.5-7B-Chat-IQ2_XXS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7be201543f8a32ec56626da56bfdaaeb13ea4dbf57d4a7f0aa1ce218d6e32313
3
+ size 2619361024
CodeQwen1.5-7B-Chat-IQ3_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76032bdefb4b55fe113c69dd0e3dbd1b6e482929a49a5b4c0d181b6c58520faa
3
+ size 3608545024
CodeQwen1.5-7B-Chat-IQ3_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:afe446593f44ed6d78caa0df505688adb30c680e1b16ceea0acf1d14f6700031
3
+ size 3509716736
CodeQwen1.5-7B-Chat-IQ3_XS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d0027ee12e886d49143fb14a2515a0486e61b53502eee93a8013ab25d4567b5
3
+ size 3357542144
CodeQwen1.5-7B-Chat-IQ3_XXS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:11e2644b993639d87e617fa966587d318036fc18e315024459bb02f0e7eb2c89
3
+ size 3228231424
CodeQwen1.5-7B-Chat-IQ4_NL.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ef44a1cce875593039753fa01780379c71fb16d50f91c771369c31d2548e217
3
+ size 4187825920
CodeQwen1.5-7B-Chat-IQ4_XS.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d0f9dc29cbe3102ea5e18515ed86dd70f03c46e83718d96b37f5513fb02f16f
3
+ size 4030244608
CodeQwen1.5-7B-Chat-Q2_K.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d114a4f2483d8a0ebbfb8df8a52d5aa8d4c8ab723f4b0d7dffbf58692ebeb62a
3
+ size 3052103424
CodeQwen1.5-7B-Chat-Q3_K_L.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7097bd36160b8b2e32a44296409b20130b2a1652795e74b38c2dfed5279cdcee
3
+ size 3986818816
CodeQwen1.5-7B-Chat-Q3_K_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:676c529471a282bd43398695a8b8808c67ee99a091665a9fcf21096b6ca1bffa
3
+ size 3808626432
CodeQwen1.5-7B-Chat-Q3_K_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63200134db790075f75b8a63cb520611732aaeea07b7a15a39364bf362d78680
3
+ size 3500803840
CodeQwen1.5-7B-Chat-Q4_K_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:377d950f7f0b42662f84b19308e1939fd39b0bac82f486c73c89327cff07403b
3
+ size 4738590464
CodeQwen1.5-7B-Chat-Q4_K_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:948f0031bc275bf9e91d73952ac3e44c76f8fd3b6df048f90876ec33265d02d7
3
+ size 4414449408
CodeQwen1.5-7B-Chat-Q5_K_M.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ffc9674d3b6ee3dee58cf40838cb82fea13ff46d3c96ea5002ce9c6accc169ec
3
+ size 5428160256
CodeQwen1.5-7B-Chat-Q5_K_S.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:56b9f1977f0b82879fd80621985de51faf71d7b11756e89475607ebd9dcd8400
3
+ size 5148452608
CodeQwen1.5-7B-Chat-Q6_K.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19d6b0ca3eda8d6b04ec4018a730680493e72d52d8f4730ce3ed1313320a7750
3
+ size 6377588480
CodeQwen1.5-7B-Chat-Q8_0.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09a17459e3538c33a0499c7d0f9fd12bfc12ac5632e0b207fe0986c5a92b1388
3
+ size 7706773248
CodeQwen1.5-7B-Chat.imatrix ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ab641069b68fac36355f31779afe410300b6bf5272cd165efb24b722f3962b5
3
+ size 4873438
README.md ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ license_name: tongyi-qianwen
4
+ license_link: >-
5
+ https://huggingface.co/Qwen/CodeQwen1.5-7B-Chat/blob/main/LICENSE
6
+ language:
7
+ - en
8
+ pipeline_tag: text-generation
9
+ tags:
10
+ - chat
11
+ quantized_by: bartowski
12
+ ---
13
+
14
+ ## Llamacpp Quantizations of CodeQwen1.5-7B-Chat
15
+
16
+ Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/experimental">experimental</a> for quantization.
17
+
18
+ Original model: https://huggingface.co/Qwen/CodeQwen1.5-7B-Chat
19
+
20
+ All quants made using imatrix option with dataset provided by Kalomaze [here](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384)
21
+
22
+ ## Prompt format
23
+
24
+ ```
25
+ <|im_start|>system
26
+ {system_prompt}<|im_end|>
27
+ <|im_start|>user
28
+ {prompt}<|im_end|>
29
+ <|im_start|>assistant
30
+
31
+ ```
32
+
33
+ ## Download a file (not the whole branch) from below:
34
+
35
+ | Filename | Quant type | File Size | Description |
36
+ | -------- | ---------- | --------- | ----------- |
37
+ | [CodeQwen1.5-7B-Chat-Q8_0.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q8_0.gguf) | Q8_0 | 7.70GB | Extremely high quality, generally unneeded but max available quant. |
38
+ | [CodeQwen1.5-7B-Chat-Q6_K.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q6_K.gguf) | Q6_K | 6.37GB | Very high quality, near perfect, *recommended*. |
39
+ | [CodeQwen1.5-7B-Chat-Q5_K_M.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q5_K_M.gguf) | Q5_K_M | 5.42GB | High quality, *recommended*. |
40
+ | [CodeQwen1.5-7B-Chat-Q5_K_S.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q5_K_S.gguf) | Q5_K_S | 5.14GB | High quality, *recommended*. |
41
+ | [CodeQwen1.5-7B-Chat-Q4_K_M.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q4_K_M.gguf) | Q4_K_M | 4.73GB | Good quality, uses about 4.83 bits per weight, *recommended*. |
42
+ | [CodeQwen1.5-7B-Chat-Q4_K_S.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q4_K_S.gguf) | Q4_K_S | 4.41GB | Slightly lower quality with more space savings, *recommended*. |
43
+ | [CodeQwen1.5-7B-Chat-IQ4_NL.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ4_NL.gguf) | IQ4_NL | 4.18GB | Decent quality, slightly smaller than Q4_K_S with similar performance *recommended*. |
44
+ | [CodeQwen1.5-7B-Chat-IQ4_XS.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ4_XS.gguf) | IQ4_XS | 4.03GB | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
45
+ | [CodeQwen1.5-7B-Chat-Q3_K_L.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q3_K_L.gguf) | Q3_K_L | 3.98GB | Lower quality but usable, good for low RAM availability. |
46
+ | [CodeQwen1.5-7B-Chat-Q3_K_M.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q3_K_M.gguf) | Q3_K_M | 3.80GB | Even lower quality. |
47
+ | [CodeQwen1.5-7B-Chat-IQ3_M.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ3_M.gguf) | IQ3_M | 3.60GB | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
48
+ | [CodeQwen1.5-7B-Chat-IQ3_S.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ3_S.gguf) | IQ3_S | 3.50GB | Lower quality, new method with decent performance, recommended over Q3_K_S quant, same size with better performance. |
49
+ | [CodeQwen1.5-7B-Chat-Q3_K_S.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q3_K_S.gguf) | Q3_K_S | 3.50GB | Low quality, not recommended. |
50
+ | [CodeQwen1.5-7B-Chat-IQ3_XS.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ3_XS.gguf) | IQ3_XS | 3.35GB | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
51
+ | [CodeQwen1.5-7B-Chat-IQ3_XXS.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ3_XXS.gguf) | IQ3_XXS | 3.22GB | Lower quality, new method with decent performance, comparable to Q3 quants. |
52
+ | [CodeQwen1.5-7B-Chat-Q2_K.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-Q2_K.gguf) | Q2_K | 3.05GB | Very low quality but surprisingly usable. |
53
+ | [CodeQwen1.5-7B-Chat-IQ2_M.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ2_M.gguf) | IQ2_M | 3.00GB | Very low quality, uses SOTA techniques to also be surprisingly usable. |
54
+ | [CodeQwen1.5-7B-Chat-IQ2_S.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ2_S.gguf) | IQ2_S | 2.87GB | Very low quality, uses SOTA techniques to be usable. |
55
+ | [CodeQwen1.5-7B-Chat-IQ2_XS.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ2_XS.gguf) | IQ2_XS | 2.76GB | Very low quality, uses SOTA techniques to be usable. |
56
+ | [CodeQwen1.5-7B-Chat-IQ2_XXS.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ2_XXS.gguf) | IQ2_XXS | 2.61GB | Lower quality, uses SOTA techniques to be usable. |
57
+ | [CodeQwen1.5-7B-Chat-IQ1_M.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ1_M.gguf) | IQ1_M | 2.45GB | Extremely low quality, *not* recommended. |
58
+ | [CodeQwen1.5-7B-Chat-IQ1_S.gguf](https://huggingface.co/bartowski/CodeQwen1.5-7B-Chat-GGUF/blob/main/CodeQwen1.5-7B-Chat-IQ1_S.gguf) | IQ1_S | 2.36GB | Extremely low quality, *not* recommended. |
59
+
60
+ ## Which file should I choose?
61
+
62
+ A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
63
+
64
+ 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.
65
+
66
+ 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.
67
+
68
+ 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.
69
+
70
+ Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
71
+
72
+ 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.
73
+
74
+ If you want to get more into the weeds, you can check out this extremely useful feature chart:
75
+
76
+ [llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)
77
+
78
+ 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.
79
+
80
+ 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.
81
+
82
+ 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.
83
+
84
+ Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski