Update README.md
Browse files
README.md
CHANGED
|
@@ -12,11 +12,11 @@ pipeline_tag: text-generation
|
|
| 12 |
[Project page](https://iprc-dip.github.io/CodeV-R1)
|
| 13 |
|
| 14 |
<div class="figure-container" style="display: flex; flex-direction: column; gap: 15px; max-width: 850px;">
|
| 15 |
-
<div style="display: flex; gap: 10px; justify-content: center;">
|
| 16 |
<img src="./assets/rtllm_tts.png" alt="RTLLM TTS Results" width="400">
|
| 17 |
<img src="./assets/rtllm_tts_flops.png" alt="RTLLM TTS FLOPs Results" width="400">
|
| 18 |
</div>
|
| 19 |
-
<figcaption class="caption
|
| 20 |
Test-time scaling curves. <strong>Left</strong>: Inference time as a function of token length. <strong>Right</strong>: Inference time vs. estimated FLOPs consumption.
|
| 21 |
When measured by FLOPs consumption, our <strong>CodeV-R1-Qwen-7B</strong> achieves better results with fewer computational resources than DeepSeek-R1, highlighting its superior efficiency.
|
| 22 |
</figcaption>
|
|
|
|
| 12 |
[Project page](https://iprc-dip.github.io/CodeV-R1)
|
| 13 |
|
| 14 |
<div class="figure-container" style="display: flex; flex-direction: column; gap: 15px; max-width: 850px;">
|
| 15 |
+
<div style="display: flex; gap: 10px; justify-content: center; margin-bottom: -3rem;">
|
| 16 |
<img src="./assets/rtllm_tts.png" alt="RTLLM TTS Results" width="400">
|
| 17 |
<img src="./assets/rtllm_tts_flops.png" alt="RTLLM TTS FLOPs Results" width="400">
|
| 18 |
</div>
|
| 19 |
+
<figcaption class="caption has-text-centered has-text-grey" style="font-size: 0.8rem;">
|
| 20 |
Test-time scaling curves. <strong>Left</strong>: Inference time as a function of token length. <strong>Right</strong>: Inference time vs. estimated FLOPs consumption.
|
| 21 |
When measured by FLOPs consumption, our <strong>CodeV-R1-Qwen-7B</strong> achieves better results with fewer computational resources than DeepSeek-R1, highlighting its superior efficiency.
|
| 22 |
</figcaption>
|