DeepInfer-Mac good first issues (Python)

Exact indexer screening for long-context prefill (DoSA) / 无损 indexer 加速（DoSA） 1 day ago

help wanted performance

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

Profile the indexer's share of prefill at 32K–128K / 测量长上下文下 indexer 的耗时占比 2 days ago

help wanted benchmark performance

haoruilee/DeepInfer-Mac

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

Deterministic tie-breaking in indexer Top-K / indexer Top-K 的确定性并列顺序 3 days ago

good first issue quality

haoruilee/DeepInfer-Mac

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

Troubleshooting section in the READMEs / README 常见问题 3 days ago

documentation good first issue

haoruilee/DeepInfer-Mac

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

Chat CLI: /stats, /save, /load, /system / 终端聊天的实用命令 3 days ago

enhancement good first issue

haoruilee/DeepInfer-Mac

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

doctor: --json, swap warning, machine info / 改进 doctor 命令 3 days ago

enhancement good first issue

haoruilee/DeepInfer-Mac

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

Prompt logprobs / prompt 对数概率 3 days ago

enhancement good first issue

haoruilee/DeepInfer-Mac

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

Grammar-constrained tool calls in the server / 服务端工具调用的约束解码 3 days ago

enhancement good first issue server

haoruilee/DeepInfer-Mac

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

Benchmark your Mac / 提交你的 Mac 跑分 3 days ago

help wanted good first issue benchmark

haoruilee/DeepInfer-Mac

1

DeepSeek-V4.1-Flash (754B) on a single Apple Silicon Mac — DeepInfer-style MLX inference with SSD expert streaming, DSpark, SWA bounded replay, structured outputs

Python

#apple-silicon#deepseek#deepseek-v4#inference-engine#llm#llm-inference#macos#metal#mixture-of-experts#mlx#moe#quantization#speculative-decoding

haoruilee/DeepInfer-Mac

Open Issues Need Help