Update README.md
Browse files
README.md
CHANGED
|
@@ -12,9 +12,11 @@ tags:
|
|
| 12 |
|
| 13 |
LLM360 demystifies the training recipe used for Llama 2 - 70B with K2.
|
| 14 |
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
|
|
|
|
|
|
| 18 |
* Languages: English
|
| 19 |
* Models Released: base, chat model
|
| 20 |
* Trained in 2 stages
|
|
|
|
| 12 |
|
| 13 |
LLM360 demystifies the training recipe used for Llama 2 - 70B with K2.
|
| 14 |
|
| 15 |
+
<center><img src="k2_eval_table.png" alt="k2 eval table"/></center>
|
| 16 |
+
|
| 17 |
+
## About K2:
|
| 18 |
+
* 65 billion parameter LLM
|
| 19 |
+
* Tokens: 1.4T
|
| 20 |
* Languages: English
|
| 21 |
* Models Released: base, chat model
|
| 22 |
* Trained in 2 stages
|