Yuepixel/tensorfold-qwen38-exl3-lookup

Qwen3.8-Flash-Next EXL3 3.05bpw on ONE DGX Spark (GB10): ~2.4x prefill (700->1777 t/s), prompt-lookup drafter, native vision/video, 262k ctx. Unofficial TensorFold v0.6.5 fork, 18 patches, byte-exact receipts.

语言：Python · GitHub 星标：9 · Fork：0

话题：cuda · dgx-spark · exl3 · gb10 · llm-inference · prompt-lookup · qwen3 · speculative-decoding

在 GitHub 上查看原仓库