Nexesenex commited on
Commit
c12d2c0
·
verified ·
1 Parent(s): dd5d800

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +7 -7
README.md CHANGED
@@ -19,29 +19,29 @@ Edit : Due to a poor CPU (i7-6700k) for AI purpose, and only 36GB of VRAM, I rem
19
 
20
  More than 3% in Rope 8 on Q2_K
21
 
22
- WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,6.2489,512,
23
  WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-iMat-c32_ch25-Q2_K.gguf,-,wikitext,6.0482,512
24
 
25
  More than 2% in Rope 4 on Q2_K
26
 
27
- WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.8859
28
  WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-iMat-c32_ch25-Q2_K.gguf,-,wikitext,4.7739,512
29
 
30
  More than 1.5% in Rope 2 on Q2_K
31
 
32
- WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.5030,512,
33
  WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-iMat-c32_ch25-Q2_K.gguf,-,wikitext,4.42,512
34
 
35
  More than 1% with Rope 8 on Q3_K_S
36
 
37
- WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q3_K_S.gguf,-,wikitext,5.6127,512
38
  WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-iMat-c32_ch25-Q3_K_S.gguf,-,wikitext,5.5461,512
39
 
40
  Interestingly, Rope 2.5, 3, and 3.2 are quite good, here are the values with the normal Q2_K :
41
 
42
- Rope 2.5 (max context 10240) : WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.5246,512
43
- Rope 3 (max context 12288) : WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.6203,512
44
- Rope 3.2 (max context 13107) : WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.6679,512
45
 
46
  So the linear rope, at least on this model, is flexible, and you can lower it to have the best peplexity for your max context.
47
 
 
19
 
20
  More than 3% in Rope 8 on Q2_K
21
 
22
+ WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,6.2489,512
23
  WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-iMat-c32_ch25-Q2_K.gguf,-,wikitext,6.0482,512
24
 
25
  More than 2% in Rope 4 on Q2_K
26
 
27
+ WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.8859,512
28
  WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-iMat-c32_ch25-Q2_K.gguf,-,wikitext,4.7739,512
29
 
30
  More than 1.5% in Rope 2 on Q2_K
31
 
32
+ WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.5030,512
33
  WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-iMat-c32_ch25-Q2_K.gguf,-,wikitext,4.42,512
34
 
35
  More than 1% with Rope 8 on Q3_K_S
36
 
37
+ WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q3_K_S.gguf,-,wikitext,5.6127,512
38
  WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-iMat-c32_ch25-Q3_K_S.gguf,-,wikitext,5.5461,512
39
 
40
  Interestingly, Rope 2.5, 3, and 3.2 are quite good, here are the values with the normal Q2_K :
41
 
42
+ Rope 2.5 (max context 10240) : WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.5246,512
43
+ Rope 3 (max context 12288) : WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.6203,512
44
+ Rope 3.2 (max context 13107) : WinterGoddess-1.4x-limarpv3-70B-L2-32k-Requant-AR-b1924-Q2_K.gguf,-,wikitext,4.6679,512
45
 
46
  So the linear rope, at least on this model, is flexible, and you can lower it to have the best peplexity for your max context.
47