The project deepseek-ai/DeepSeek-V4-Flash-0731 appears to be missing from GitHub, indicating that it is either not available anymore or has never existed under that name. However, several related repositories featuring the "DeepSeek V4 Flash" are actively maintained. Here's a closer look:
| Project | Stars | Maintenance Status | Last Pushed | Open Issues | License | Description |
|---|---|---|---|---|---|---|
| MiaAI-Lab/DeepSeek-v4-Flash-DSpark-2x-DGX-Spark | 1,103 | ✅ Actively Maintained | Recently (0 days ago) | 29 | MIT | DeepSeek-v4-Flash 0731 recipe for 2x DGX Sparks. |
| antirez/llama.cpp-deepseek-v4-flash | 343 | ⚠️ Maintaining | Last pushed (126 days ago) | 4 | MIT | Experimental implementation of DeepSeek v4 flash in llama.cpp. |
| MiaAI-Lab/DeepSeek-v4-Flash-One-DGX-Spark | 295 | ✅ Actively Maintained | Recently (7 days ago) | 5 | MIT | DeepSeek v4 Flash EXL3 on one DGX Spark. |
| tonyd2wild/DeepSeek-v4-Flash-Vision-Exp-DSpark-1M-NVFP4-KV-2x-DGX-Spark | 457 | ✅ Actively Maintained | Recently (0 days ago) | 16 | MIT | DeepSeek V4 Flash DSpark 1M NVFP4 KV recipe for 2x DGX Spark. |
| 0xSero/deepseek-v4-flash-0731-spark-sparkinfer | 163 | ✅ Actively Maintained | Recently (25 days ago) | 4 | Apache-2.0 | Pinned one-DGX-Spark Docker recipe for DeepSeek V4 Flash with EXL3, SparkInfer, and 262K NVFP4 MLA KV cache. |
Since the original repository is not accessible, it's safe to say that deepseek-ai/DeepSeek-V4-Flash-0731 is effectively dead. On the other hand, several actively maintained forks and related projects exist, like the two major ones from MiaAI-Lab, which maintain a similar focus.
If you are looking for implementation, consider using the actively maintained alternatives like MiaAI-Lab/DeepSeek-v4-Flash-DSpark-2x-DGX-Spark or MiaAI-Lab/DeepSeek-v4-Flash-One-DGX-Spark.