bartowski commited on
Commit
3fcb103
1 Parent(s): 2d6a112

Llamacpp quants

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +65 -0
  2. README.md +69 -0
  3. WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00001-of-00005.gguf +3 -0
  4. WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00002-of-00005.gguf +3 -0
  5. WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00003-of-00005.gguf +3 -0
  6. WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00004-of-00005.gguf +3 -0
  7. WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00005-of-00005.gguf +3 -0
  8. WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00001-of-00005.gguf +3 -0
  9. WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00002-of-00005.gguf +3 -0
  10. WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00003-of-00005.gguf +3 -0
  11. WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00004-of-00005.gguf +3 -0
  12. WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00005-of-00005.gguf +3 -0
  13. WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00001-of-00005.gguf +3 -0
  14. WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00002-of-00005.gguf +3 -0
  15. WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00003-of-00005.gguf +3 -0
  16. WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00004-of-00005.gguf +3 -0
  17. WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00005-of-00005.gguf +3 -0
  18. WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00001-of-00005.gguf +3 -0
  19. WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00002-of-00005.gguf +3 -0
  20. WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00003-of-00005.gguf +3 -0
  21. WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00004-of-00005.gguf +3 -0
  22. WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00005-of-00005.gguf +3 -0
  23. WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00001-of-00005.gguf +3 -0
  24. WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00002-of-00005.gguf +3 -0
  25. WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00003-of-00005.gguf +3 -0
  26. WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00004-of-00005.gguf +3 -0
  27. WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00005-of-00005.gguf +3 -0
  28. WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00001-of-00005.gguf +3 -0
  29. WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00002-of-00005.gguf +3 -0
  30. WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00003-of-00005.gguf +3 -0
  31. WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00004-of-00005.gguf +3 -0
  32. WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00005-of-00005.gguf +3 -0
  33. WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00001-of-00005.gguf +3 -0
  34. WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00002-of-00005.gguf +3 -0
  35. WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00003-of-00005.gguf +3 -0
  36. WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00004-of-00005.gguf +3 -0
  37. WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00005-of-00005.gguf +3 -0
  38. WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00001-of-00005.gguf +3 -0
  39. WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00002-of-00005.gguf +3 -0
  40. WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00003-of-00005.gguf +3 -0
  41. WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00004-of-00005.gguf +3 -0
  42. WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00005-of-00005.gguf +3 -0
  43. WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00001-of-00005.gguf +3 -0
  44. WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00002-of-00005.gguf +3 -0
  45. WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00003-of-00005.gguf +3 -0
  46. WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00004-of-00005.gguf +3 -0
  47. WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00005-of-00005.gguf +3 -0
  48. WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00001-of-00005.gguf +3 -0
  49. WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00002-of-00005.gguf +3 -0
  50. WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00003-of-00005.gguf +3 -0
.gitattributes CHANGED
@@ -33,3 +33,68 @@ 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
+ WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
37
+ WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
38
+ WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
39
+ WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
40
+ WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
41
+ WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
42
+ WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
43
+ WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
44
+ WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
45
+ WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
46
+ WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
47
+ WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
48
+ WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
49
+ WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
50
+ WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
51
+ WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
52
+ WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
53
+ WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
54
+ WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
55
+ WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
56
+ WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
57
+ WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
58
+ WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
59
+ WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
60
+ WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
61
+ WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
62
+ WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
63
+ WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
64
+ WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
65
+ WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
66
+ WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
67
+ WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
68
+ WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
69
+ WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
70
+ WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
71
+ WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
72
+ WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
73
+ WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
74
+ WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
75
+ WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
76
+ WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
77
+ WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
78
+ WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
79
+ WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
80
+ WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
81
+ WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
82
+ WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
83
+ WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
84
+ WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
85
+ WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
86
+ WizardLM-2-8x22B-Q4_K_S.gguf/WizardLM-2-8x22B-Q4_K_S-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
87
+ WizardLM-2-8x22B-Q4_K_S.gguf/WizardLM-2-8x22B-Q4_K_S-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
88
+ WizardLM-2-8x22B-Q4_K_S.gguf/WizardLM-2-8x22B-Q4_K_S-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
89
+ WizardLM-2-8x22B-Q4_K_S.gguf/WizardLM-2-8x22B-Q4_K_S-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
90
+ WizardLM-2-8x22B-Q4_K_S.gguf/WizardLM-2-8x22B-Q4_K_S-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
91
+ WizardLM-2-8x22B-Q5_K_M.gguf/WizardLM-2-8x22B-Q5_K_M-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
92
+ WizardLM-2-8x22B-Q5_K_M.gguf/WizardLM-2-8x22B-Q5_K_M-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
93
+ WizardLM-2-8x22B-Q5_K_M.gguf/WizardLM-2-8x22B-Q5_K_M-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
94
+ WizardLM-2-8x22B-Q5_K_M.gguf/WizardLM-2-8x22B-Q5_K_M-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
95
+ WizardLM-2-8x22B-Q5_K_M.gguf/WizardLM-2-8x22B-Q5_K_M-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
96
+ WizardLM-2-8x22B-Q5_K_S.gguf/WizardLM-2-8x22B-Q5_K_S-00001-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
97
+ WizardLM-2-8x22B-Q5_K_S.gguf/WizardLM-2-8x22B-Q5_K_S-00002-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
98
+ WizardLM-2-8x22B-Q5_K_S.gguf/WizardLM-2-8x22B-Q5_K_S-00003-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
99
+ WizardLM-2-8x22B-Q5_K_S.gguf/WizardLM-2-8x22B-Q5_K_S-00004-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
100
+ WizardLM-2-8x22B-Q5_K_S.gguf/WizardLM-2-8x22B-Q5_K_S-00005-of-00005.gguf filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ quantized_by: bartowski
4
+ pipeline_tag: text-generation
5
+ ---
6
+
7
+ ## Llamacpp Quantizations of WizardLM-2-8x22B
8
+
9
+ Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b2675">b2675</a> for quantization.
10
+
11
+ Original model: https://huggingface.co/microsoft/WizardLM-2-8x22B
12
+
13
+ ## Prompt format
14
+
15
+ No chat template specified so default is used. This may be incorrect, check original model card for details.
16
+
17
+ ```
18
+ <s> [INST] <<SYS>>
19
+ {system_prompt}
20
+ <</SYS>>
21
+
22
+ {prompt} [/INST] </s>
23
+ ```
24
+
25
+ ## Download a file (not the whole branch) from below:
26
+
27
+ | Filename | Quant type | File Size | Description |
28
+ | -------- | ---------- | --------- | ----------- |
29
+ | [WizardLM-2-8x22B-Q8_0.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF//main/WizardLM-2-8x22B-Q8_0.gguf) | Q8_0 | | Extremely high quality, generally unneeded but max available quant. |
30
+ | [WizardLM-2-8x22B-Q6_K.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF//main/WizardLM-2-8x22B-Q6_K.gguf) | Q6_K | | Very high quality, near perfect, *recommended*. |
31
+ | [WizardLM-2-8x22B-Q5_K_M.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-Q5_K_M.gguf) | Q5_K_M | 99.96GB | High quality, *recommended*. |
32
+ | [WizardLM-2-8x22B-Q5_K_S.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-Q5_K_S.gguf) | Q5_K_S | 96.97GB | High quality, *recommended*. |
33
+ | [WizardLM-2-8x22B-Q4_K_M.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-Q4_K_M.gguf) | Q4_K_M | 85.58GB | Good quality, uses about 4.83 bits per weight, *recommended*. |
34
+ | [WizardLM-2-8x22B-Q4_K_S.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-Q4_K_S.gguf) | Q4_K_S | 80.47GB | Slightly lower quality with more space savings, *recommended*. |
35
+ | [WizardLM-2-8x22B-IQ4_NL.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-IQ4_NL.gguf) | IQ4_NL | 80.47GB | Decent quality, slightly smaller than Q4_K_S with similar performance *recommended*. |
36
+ | [WizardLM-2-8x22B-IQ4_XS.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-IQ4_XS.gguf) | IQ4_XS | 76.35GB | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
37
+ | [WizardLM-2-8x22B-Q3_K_L.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-Q3_K_L.gguf) | Q3_K_L | 72.57GB | Lower quality but usable, good for low RAM availability. |
38
+ | [WizardLM-2-8x22B-Q3_K_M.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-Q3_K_M.gguf) | Q3_K_M | 67.78GB | Even lower quality. |
39
+ | [WizardLM-2-8x22B-IQ3_M.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-IQ3_M.gguf) | IQ3_M | 64.49GB | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
40
+ | [WizardLM-2-8x22B-IQ3_S.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-IQ3_S.gguf) | IQ3_S | 61.49GB | Lower quality, new method with decent performance, recommended over Q3_K_S quant, same size with better performance. |
41
+ | [WizardLM-2-8x22B-Q3_K_S.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-Q3_K_S.gguf) | Q3_K_S | 61.49GB | Low quality, not recommended. |
42
+ | [WizardLM-2-8x22B-IQ3_XS.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-IQ3_XS.gguf) | IQ3_XS | 58.22GB | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
43
+ | [WizardLM-2-8x22B-Q2_K.gguf](https://huggingface.co/bartowski/WizardLM-2-8x22B-GGUF/tree/main/WizardLM-2-8x22B-Q2_K.gguf) | Q2_K | 52.10GB | Very low quality but surprisingly usable. |
44
+
45
+ ## Which file should I choose?
46
+
47
+ A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
48
+
49
+ 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.
50
+
51
+ 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.
52
+
53
+ 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.
54
+
55
+ Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
56
+
57
+ 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.
58
+
59
+ If you want to get more into the weeds, you can check out this extremely useful feature chart:
60
+
61
+ [llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)
62
+
63
+ 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.
64
+
65
+ 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.
66
+
67
+ 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.
68
+
69
+ Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:66971fa47c15ef97dca7d3c68e227a71ac7d5debb768d740bbed4463d825b6ff
3
+ size 14920852384
WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2910280dced795296535ebcc99a8a0fc61042f25cb0e2e49c3371f4ae3d7d8ee
3
+ size 14953668800
WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e19f79c2df988303e339179bca129a81360a496671783310bb968841fc8ebbfd
3
+ size 13625286784
WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f80f980d5857b79e59104274ddf6f078003168c0847a20f867eaf419be2f98ea
3
+ size 14834819232
WizardLM-2-8x22B-IQ3_M.gguf/WizardLM-2-8x22B-IQ3_M-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b933c69ad628c08cb6b035acc98e3b02a74fdd505082586aee5818fae155480
3
+ size 6158294336
WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9dbda48303e757c2b7199ed138150f2b3bd85ee2505c5e9873b91bbbdaed7344
3
+ size 13958259616
WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5c294d9fcbc8ec811c5799ba5b7b72b2b770706a943f244db8091916e073949
3
+ size 14525849792
WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1839e86c7a30530b4167fd28ec704fdacd73120b614e498ed7808ddccf3d2599
3
+ size 13197467776
WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09c96178da60912612e0ff6523cae24cb762e4dbb5d4768d0d2a394d41911fd8
3
+ size 14193090720
WizardLM-2-8x22B-IQ3_S.gguf/WizardLM-2-8x22B-IQ3_S-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:64b09d7214f1cb11e92e0ae62f657852a07ec6c34072fd9d6597912e10e33864
3
+ size 5623520576
WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6d4400347be63961318fbf10f3277b1c2fc54a7e686c65f056c783df972927c0
3
+ size 13520020384
WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e566a04d762cddc4579b844e425c614b46e2266756e7255e99d03bd93c3144e
3
+ size 13483630784
WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fafb55b1e73da0c6ababd35b4086578f0c29b1a04b5492ad3e3475660f85ddab
3
+ size 12270264448
WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e421a93a1fcbd9ff769cc5eed1e74b8c3ab281ecdc7977f915c6929ffa9859b7
3
+ size 13339615392
WizardLM-2-8x22B-IQ3_XS.gguf/WizardLM-2-8x22B-IQ3_XS-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37afb9affd4192da35e6d4e4879a79571457b1d2b2d358f5ab269978c1bc506c
3
+ size 5614673216
WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf716ff0decaed635826b9ddcabb7dcb74f50a0fcf9c11bf11d2ab26a30b162c
3
+ size 18818470816
WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d337799bfdeed565551612d0d1adb6ca853d2b4fbe29fee3ce0650f32e8ffc7
3
+ size 18869215424
WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a2858411d2d0b083692e8f7006fc6a2980f466e152c547c957cd62a5c65f55b
3
+ size 17108000896
WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:225034b3ddc8ad3e48104ef09eacf6d4d1bfdcccb943a12a9b69f1aae2c85940
3
+ size 18429501600
WizardLM-2-8x22B-IQ4_NL.gguf/WizardLM-2-8x22B-IQ4_NL-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37c79d377fe14d94f711b209ff392cba7a2099b8f39b1381a32d4d94de3cf1eb
3
+ size 7252909376
WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5222f36c18ca13e4fb8b90d3c4a9c06b3df1b21201cafb8c4987929e91fdf254
3
+ size 18016850848
WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ea1881a0b42f37767af47c1d13b9eb0efe510853a731b678feabff97c015719
3
+ size 17847247040
WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52ae1b3b1b75dabaf96484ec1e0141f0ce65225f1b98a73d2261a50d6cae3f15
3
+ size 16187875456
WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ce912e4e6c0948169b3b192268cbdb04af0c1bd8f3790bfcb9cd2fc39d4f0291
3
+ size 17432699040
WizardLM-2-8x22B-IQ4_XS.gguf/WizardLM-2-8x22B-IQ4_XS-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:762ac686d247b8b7499ea8076c432df8f409807401819c47f7042be617ad8bce
3
+ size 6869523776
WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:823e851cf071e95a8397b64854a5207f88ae4fde16452a1481b25a09b3f34448
3
+ size 11843728288
WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:53a3fe6da0e7076c168ff7c94b454ec98d0c46c575d27ece5236be21c61af3f6
3
+ size 12267708608
WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d6434c42e263006300fddbd9f3b923a6cc690ddd350847579c65dc178ed0bea
3
+ size 11188527232
WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af766c6d298310821a4ab132bc8e6a21817023ca428ffa2bb6a54b9d2da45600
3
+ size 12016738464
WizardLM-2-8x22B-Q2_K.gguf/WizardLM-2-8x22B-Q2_K-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a7fd3949c13b60d1369f86663cef9ca4e900037e9c621ab39db2790ec64b519
3
+ size 4786462016
WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04355dc334eb9d2b3babd1ba7019c6e77128edd27cc8ebc572503747d136008c
3
+ size 16530776992
WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b3c810fd9c410f7d21effb1d75fb1255a3a06770a696755d8bdadd618f07bf58
3
+ size 17108099264
WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5129af244ebaaecbf40f89cb6eede1b2035c4f9a8f1df066ed97413e87fb4eec
3
+ size 15552635008
WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ecef6b5d484ae03d0684ec579fb086e3864cf2b50805c85dd3a6dbffb69fd8fe
3
+ size 16775340192
WizardLM-2-8x22B-Q3_K_L.gguf/WizardLM-2-8x22B-Q3_K_L-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af2364a5dc9bde78cfccf90063df9b01658410629124ceadb40eef2edb9f628b
3
+ size 6612950336
WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:265d1fdd0d6387132566ced74b854dcef7ffde3dc70126066541603fbfdbe0af
3
+ size 15650661280
WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:001d7eb7125217a694781fecf9ca2cb49c655fb5b0bd124a4025b198dd140bae
3
+ size 15916261568
WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:577ea20c1fcc89ee42e5fabdd372dc7974857bf0eaedfab111ca94c6ba0fc607
3
+ size 14480924800
WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a96a5ad47fa40eebd46723a2db791e5834430bbb1a25014839028fa64db87bf
3
+ size 15583502496
WizardLM-2-8x22B-Q3_K_M.gguf/WizardLM-2-8x22B-Q3_K_M-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f79c807121753e7c4d8312f02330c5315a37e47ac1e1feb6a64f395416fb6af1
3
+ size 6158294336
WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fffe9c98da87351123cac81e63d66b1174fbbcdc353ec6e83d253b44d91cb71a
3
+ size 13958259616
WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4578f7ef6150c93b7347bdb77c6f2c3f839bec145ec9022bccea8900d8e3caf2
3
+ size 14525849792
WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbf516d44516ac5f1478f48dfb37c45ebd91940111b32af93edbcf629441e8a5
3
+ size 13197467776
WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00004-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f4433ced22748b00fabffe18ef41b93aa164214029603286ecc55ad7ef6f8e1
3
+ size 14193090720
WizardLM-2-8x22B-Q3_K_S.gguf/WizardLM-2-8x22B-Q3_K_S-00005-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13b984ca43dcbf4d0a873d7658400060ab00629cc784716142819dbb045da01c
3
+ size 5623520576
WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00001-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d31fc573ada123bd80246c9bf8b2d5b916b0c74022d412fde3a6f08a6b6fb21
3
+ size 19982390176
WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00002-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cda9d26caeeb40eb916e04b4ef5efc0117b90ffe63d5f8fe997262fb04b4b333
3
+ size 19699687616
WizardLM-2-8x22B-Q4_K_M.gguf/WizardLM-2-8x22B-Q4_K_M-00003-of-00005.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2765d9a73e1d558781fcd77b908d3a26ce92772b02a2cf38ac0717f95f1ab58f
3
+ size 17938473088