Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,11 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
datasets:
|
| 4 |
+
- livecodebench/code_generation_lite
|
| 5 |
+
- agentica-org/DeepCoder-Preview-Dataset
|
| 6 |
+
base_model:
|
| 7 |
+
- Qwen/Qwen3-14B
|
| 8 |
+
pipeline_tag: text-generation
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
We introduce HermesCoder-14B, a code reasoning model post-trained on Qwen3-14B via reinforcement learning with verifiable rewards (RLVR). On LiveCodeBench v5 (08/01/2024 - 05/01/2025), we achieve a Pass@1 accuracy of 67.87\%, up 7.08\% from the baseline Pass@1 accuracy of 60.79\% of Qwen3-14B. To the best of our knowledge, this is the highest-performing 14B model to date. We trained on 24K verifiable coding problems using 48 B200s over the course of four days.
|