| --- |
| license: mit |
| base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B |
| pipeline_tag: text-generation |
| library_name: rkllm |
| tags: |
| - rkllm |
| - rockchip |
| - rk3576 |
| - rk3588 |
| - deepseek |
| - deepseek-r1 |
| - qwen |
| --- |
| |
| # DeepSeek-R1-Distill-Qwen-1.5B · RKLLM |
|
|
| Text generation on Rockchip RK3576 and RK3588, converted from [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B). |
|
|
| - **Conversion toolkit:** RKLLM Toolkit 1.3.0. |
| - **Inputs / output:** text → text. |
| - **License:** `mit`; the [upstream model card](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B) describes its terms and intended use. |
|
|
| ## Choose your files |
|
|
| Use the files for your board’s exact SoC. `w4a16` uses 4-bit weights and 16-bit activations; `w4a16_g128` adds a weight group size of 128. `w8a8` uses 8-bit weights and activations. File size below is download size, not total runtime RAM. |
|
|
| | Board | Quantization | Size | Language model | |
| | --- | --- | --- | --- | |
| | RK3576 | `w4a16_g128` | 1.43 GiB | [DeepSeek-R1-Distill-Qwen-1.5B_RK3576_w4a16_g128.rkllm](RK3576/DeepSeek-R1-Distill-Qwen-1.5B_RK3576_w4a16_g128.rkllm) | |
| | RK3576 | `w8a8` | 1.93 GiB | [DeepSeek-R1-Distill-Qwen-1.5B_RK3576_w8a8.rkllm](RK3576/DeepSeek-R1-Distill-Qwen-1.5B_RK3576_w8a8.rkllm) | |
| | RK3588 | `w8a8` | 1.91 GiB | [DeepSeek-R1-Distill-Qwen-1.5B_RK3588_w8a8.rkllm](RK3588/DeepSeek-R1-Distill-Qwen-1.5B_RK3588_w8a8.rkllm) | |
|
|
| ## Run on your board |
|
|
| With Docker installed on a matching Rockchip board, this example starts the **RK3576 w4a16_g128** model and its API on port 8001: |
| |
| ```bash |
| sudo docker run --rm -d --name rkllm \ |
| --privileged -v /dev:/dev -p 8001:8001 \ |
| ghcr.io/hanzo-huang/rkllm-docker/llm/deepseek-r1-distill-qwen-1.5b:rk3576-w4a16-g128 |
| ``` |
| |
| The image includes its model file. See [rkllm-docker](https://github.com/Hanzo-Huang/rkllm-docker#readme) for API requests, other board variants, and configuration. |
| |
| To download the files separately instead: |
| |
| ```bash |
| hf download HanzoHuang/DeepSeek-R1-Distill-Qwen-1.5B-RKLLM \ |
| RK3576/DeepSeek-R1-Distill-Qwen-1.5B_RK3576_w4a16_g128.rkllm \ |
| --local-dir ./models |
| ``` |
| |
| These are native RKLLM/RKNN artifacts. Use a compatible RKLLM runtime and the model’s correct chat template when integrating them in another application. |
| |
| ## Release notes |
| |
| No per-variant board benchmark is published here. Quantized output may differ from the original model; check responses on your own workloads. Use the toolkit version listed above when checking runtime compatibility. |
| |
| <details> |
| <summary>SHA-256 checksums</summary> |
| |
| ```text |
| e737c9aa8cfbfd0216e547ed67bc2325d5add2f8b2aa21037389e296ac3b8d80 RK3576/DeepSeek-R1-Distill-Qwen-1.5B_RK3576_w4a16_g128.rkllm |
| 6ceab2c6b93a8b55837ab75e4050bbce888030153c0bc83a465bd73ead44120f RK3576/DeepSeek-R1-Distill-Qwen-1.5B_RK3576_w8a8.rkllm |
| d00038d619c1905ac66ab5d499019d80889d37cc1f81be491cc503e284950c2d RK3588/DeepSeek-R1-Distill-Qwen-1.5B_RK3588_w8a8.rkllm |
| ``` |
| |
| </details> |
| |
| ## Credits |
| |
| Original model: [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B). Conversion tools: [Rockchip RKLLM](https://github.com/airockchip/rknn-llm). |
| |
