KaleiNeely commited on
Commit
0549757
·
1 Parent(s): 6935388

add ncnn int8 models

Browse files

Signed-off-by: daquexian <daquexian566@gmail.com>

RWKV-4-World-CHNtuned-1.5B-v1-20230620-ctx4096-ncnn-int8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c952e8c22a09771da59a8d3f7907d0c2ab74462848c87f9901c0267d9c8f9f34
3
+ size 2163229348
RWKV-4-World-CHNtuned-1.5B-v1-20230620-ctx4096-ncnn-int8.config ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ version: 4
2
+ act_dtype: fp32
3
+ weight_dtype: int8
4
+ head_size: 0
5
+ n_layer: 24
6
+ n_embd: 2048
7
+ n_att: 2048
8
+ n_ffn: 0
RWKV-4-World-CHNtuned-1.5B-v1-20230620-ctx4096-ncnn-int8.param ADDED
The diff for this file is too large to render. See raw diff
 
RWKV-5-World-0.1B-v1-20230803-ctx4096-ncnn-int8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fc45bb2ff7ce8db0570d57bdb2ee8e4f72968a680a8a807349cb317bd2b925a3
3
+ size 364309972
RWKV-5-World-0.1B-v1-20230803-ctx4096-ncnn-int8.config ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ version: 5
2
+ act_dtype: fp32
3
+ weight_dtype: int8
4
+ head_size: 12
5
+ n_layer: 12
6
+ n_embd: 768
7
+ n_att: 768
8
+ n_ffn: 3072
RWKV-5-World-0.1B-v1-20230803-ctx4096-ncnn-int8.param ADDED
The diff for this file is too large to render. See raw diff