Commit
·
200af1e
1
Parent(s):
1e8d7c0
Update README.md
Browse files
README.md
CHANGED
|
@@ -13,6 +13,45 @@ tags:
|
|
| 13 |
|
| 14 |
> [!NOTE]
|
| 15 |
> For DeepSeek-R1-0528-**Qwen3-8B** GGUFs, [see here](https://huggingface.co/unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
<div align="center">
|
| 18 |
<img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek-V3" />
|
|
|
|
| 13 |
|
| 14 |
> [!NOTE]
|
| 15 |
> For DeepSeek-R1-0528-**Qwen3-8B** GGUFs, [see here](https://huggingface.co/unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF).
|
| 16 |
+
<div>
|
| 17 |
+
<p style="margin-bottom: 0; margin-top: 0;">
|
| 18 |
+
<strong>Learn how to run DeepSeek-R1-0528 correctly - <a href="https://docs.unsloth.ai/basics/deepseek-r1-0528">Read our Guide</a>.</strong>
|
| 19 |
+
</p>
|
| 20 |
+
<p style="margin-bottom: 0;">
|
| 21 |
+
<em>See <a href="https://huggingface.co/collections/unsloth/deepseek-r1-all-versions-678e1c48f5d2fce87892ace5">our collection</a> for all versions of R1 including GGUF, 4-bit & 16-bit formats.</em>
|
| 22 |
+
</p>
|
| 23 |
+
<p style="margin-top: 0;margin-bottom: 0;">
|
| 24 |
+
<em><a href="https://docs.unsloth.ai/basics/unsloth-dynamic-v2.0-gguf">Unsloth Dynamic 2.0</a> achieves superior accuracy & outperforms other leading quants.</em>
|
| 25 |
+
</p>
|
| 26 |
+
<div style="display: flex; gap: 5px; align-items: center; ">
|
| 27 |
+
<a href="https://github.com/unslothai/unsloth/">
|
| 28 |
+
<img src="https://github.com/unslothai/unsloth/raw/main/images/unsloth%20new%20logo.png" width="133">
|
| 29 |
+
</a>
|
| 30 |
+
<a href="https://discord.gg/unsloth">
|
| 31 |
+
<img src="https://github.com/unslothai/unsloth/raw/main/images/Discord%20button.png" width="173">
|
| 32 |
+
</a>
|
| 33 |
+
<a href="https://docs.unsloth.ai/basics/deepseek-r1-0528-how-to-run-locally">
|
| 34 |
+
<img src="https://raw.githubusercontent.com/unslothai/unsloth/refs/heads/main/images/documentation%20green%20button.png" width="143">
|
| 35 |
+
</a>
|
| 36 |
+
</div>
|
| 37 |
+
<h1 style="margin-top: 0rem;">🐋 DeepSeek-R1-0528 Usage Guidelines</h1>
|
| 38 |
+
</div>
|
| 39 |
+
|
| 40 |
+
- Set the temperature between **0.5–0.7 (0.6 recommended)** to reduce repetition and incoherence.
|
| 41 |
+
- Set Top_P value of **0.95 (recommended)**
|
| 42 |
+
- R1-0528 uses the same chat template as the original R1 model:
|
| 43 |
+
```
|
| 44 |
+
<｜begin▁of▁sentence｜><｜User｜>What is 1+1?<｜Assistant｜>It's 2.<｜end▁of▁sentence｜><｜User｜>Explain more!<｜Assistant｜>
|
| 45 |
+
```
|
| 46 |
+
- For llama.cpp / GGUF inference, you should skip the BOS since it’ll auto add it:
|
| 47 |
+
```
|
| 48 |
+
<｜User｜>What is 1+1?<｜Assistant｜>
|
| 49 |
+
```
|
| 50 |
+
- For complete detailed instructions, see our guide: [unsloth.ai/blog/deepseek-r1-0528](https://docs.unsloth.ai/basics/deepseek-r1-0528-how-to-run-locally)
|
| 51 |
+
|
| 52 |
+
<br>
|
| 53 |
+
|
| 54 |
+
# DeepSeek-R1-0528 Model Card
|
| 55 |
|
| 56 |
<div align="center">
|
| 57 |
<img src="https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg?raw=true" width="60%" alt="DeepSeek-V3" />
|