imi2 commited on
Commit
0801f09
·
verified ·
1 Parent(s): 0702cd8

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +13 -0
README.md CHANGED
@@ -46,3 +46,16 @@ TODO:
46
  | Reka-Flash 21B | 19,648 |
47
  | Mistral 2503 | 32,768 |
48
  | Codestral 22B | 16,384 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
46
  | Reka-Flash 21B | 19,648 |
47
  | Mistral 2503 | 32,768 |
48
  | Codestral 22B | 16,384 |
49
+
50
+ ------------
51
+
52
+ ## T-MAC (larger groupsize 128?)
53
+
54
+ | Model | Size | Params | Backend | Threads | Test | t/s (tokens/sec) |
55
+ |-------------------------|---------|--------|---------|---------|--------|----------------------|
56
+ | qwen2 3B Q4_K - Medium | 1.95 GiB| 3.40 B | CPU | 4 | pp512 | 67.33 ± 0.10 |
57
+ | qwen2 3B Q4_K - Medium | 1.95 GiB| 3.40 B | CPU | 4 | tg128 | 22.72 ± 0.04 |
58
+ | qwen2 ?B INT_N Q4_K | 1.70 GiB| 3.40 B | CPU | 4 | pp512 | 59.66 ± 0.10 |
59
+ | qwen2 ?B INT_N Q4_K | 1.70 GiB| 3.40 B | CPU | 4 | tg128 | 26.43 ± 0.14 |
60
+
61
+ [Test Issue Link](https://github.com/microsoft/T-MAC/issues/79)