Update README.md
Browse files
README.md
CHANGED
|
@@ -58,9 +58,6 @@ Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source
|
|
| 58 |
Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training.
|
| 59 |
In addition, its training process is remarkably stable.
|
| 60 |
Throughout the entire training process, we did not experience any irrecoverable loss spikes or perform any rollbacks.
|
| 61 |
-
<p align="center">
|
| 62 |
-
<img width="80%" src="figures/benchmark.png">
|
| 63 |
-
</p>
|
| 64 |
|
| 65 |
## 2. Model Summary
|
| 66 |
|
|
|
|
| 58 |
Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training.
|
| 59 |
In addition, its training process is remarkably stable.
|
| 60 |
Throughout the entire training process, we did not experience any irrecoverable loss spikes or perform any rollbacks.
|
|
|
|
|
|
|
|
|
|
| 61 |
|
| 62 |
## 2. Model Summary
|
| 63 |
|