allover326 / deepseek-v4-cmp170hx
git clone https://github.com/allover326/deepseek-v4-cmp170hx.git
DeepSeek-V4-Flash-0731 on 4x CMP 170HX (sm_80): 98 tok/s decode, ~5300 tok/s prefill. DSpark speculative decoding under pipeline parallelism, which vLLM does not support upstream.
+0
stars, 24h · #26,143
+~0
stars, 7 days · #37,818
+~0
stars, 30 days · #41,170
118
★ total
26
forks · 6 open issues
Star history (as recorded here)
LanguagePython
LicenseApache-2.0
Created2026-08-03 (2mo ago)
Last push2mo ago
OwnerUser
Links
README
Loading README…