Hugging Face's logo

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

GGUF
Chinese
English
Inference Endpoints
Model card Files Community
hfl-rc commited on
Commit
984a184
•
1 Parent(s): c05ab48

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -0
README.md CHANGED
@@ -17,9 +17,12 @@ Metric: PPL, lower is better
17
  |-----|------|------|
18
  | Q2_K | 12.7790 +/- 0.17943 | 13.8057 +/- 0.19614 |
19
  | Q3_K | 10.0834 +/- 0.14063 | 9.6355 +/- 0.13483 |
 
20
  | Q4_K | 9.2864 +/- 0.13001 | 9.2097 +/- 0.12874 |
 
21
  | Q5_K | 9.0912 +/- 0.12705 | 9.0701 +/- 0.12668 |
22
  | Q6_K | 9.0799 +/- 0.12681 | 9.0558 +/- 0.12653 |
 
23
 
24
  *The model with `-im` suffix is generated with important matrix, which has generally better performance (not always though).*
25
 
 
17
  |-----|------|------|
18
  | Q2_K | 12.7790 +/- 0.17943 | 13.8057 +/- 0.19614 |
19
  | Q3_K | 10.0834 +/- 0.14063 | 9.6355 +/- 0.13483 |
20
+ | Q4_0 | 9.7072 +/- 0.13563 | - |
21
  | Q4_K | 9.2864 +/- 0.13001 | 9.2097 +/- 0.12874 |
22
+ | Q5_0 | 9.2062 +/- 0.12846 | - |
23
  | Q5_K | 9.0912 +/- 0.12705 | 9.0701 +/- 0.12668 |
24
  | Q6_K | 9.0799 +/- 0.12681 | 9.0558 +/- 0.12653 |
25
+ | Q8_0 | 9.0200 +/- 0.12616 | - |
26
 
27
  *The model with `-im` suffix is generated with important matrix, which has generally better performance (not always though).*
28