Hugging Face's logo

hfl
/
chinese-llama-2-13b-16k-gguf

GGUF
Chinese
English
Inference Endpoints
Model card Files Community
hfl-rc commited on
Commit
47d5a55
·
verified ·
1 Parent(s): de7e04f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -0
README.md CHANGED
@@ -18,9 +18,12 @@ Metric: PPL, lower is better
18
  |-----|------|------|
19
  | Q2_K | 11.8958 +/- 0.20739 | 13.0017 +/- 0.23003 |
20
  | Q3_K | 9.7130 +/- 0.17037 | 9.3443 +/- 0.16582 |
 
21
  | Q4_K | 9.0055 +/- 0.15918 | 8.9848 +/- 0.15908 |
 
22
  | Q5_K | 8.8999 +/- 0.15751 | 8.8983 +/- 0.15753 |
23
  | Q6_K | 8.8944 +/- 0.15776 | 8.8833 +/- 0.15760 |
 
24
 
25
  *The model with `-im` suffix is generated with important matrix, which has generally better performance (not always though).*
26
 
 
18
  |-----|------|------|
19
  | Q2_K | 11.8958 +/- 0.20739 | 13.0017 +/- 0.23003 |
20
  | Q3_K | 9.7130 +/- 0.17037 | 9.3443 +/- 0.16582 |
21
+ | Q4_0 | 9.2002 +/- 0.16219 | - |
22
  | Q4_K | 9.0055 +/- 0.15918 | 8.9848 +/- 0.15908 |
23
+ | Q5_0 | 8.8441 +/- 0.15690 | - |
24
  | Q5_K | 8.8999 +/- 0.15751 | 8.8983 +/- 0.15753 |
25
  | Q6_K | 8.8944 +/- 0.15776 | 8.8833 +/- 0.15760 |
26
+ | Q8_0 | 8.8745 +/- 0.15745 | - |
27
 
28
  *The model with `-im` suffix is generated with important matrix, which has generally better performance (not always though).*
29