Uploaded July 2026 | Updated September 2026, 2 weeks ago

---

Running a 284 billion parameter DeepSeek V4 Flash model completely locally on a single AMD Ryzen AI MAX+ 395 with 128GB unified memory, hitting 32 tokens per second using the Lucebox inference engine with DSpark speculative decoding.

🔥 Get 50% Discount on any A6000 or A5000 GPU rental, use following link and coupon:

bit.ly/fahd-mirza
Coupon code: FahdMirza

🔥 Buy Me a Coffee to support the channel: ko-fi.com/fahdmirza

#lucebox #dflash #halostrix #deepseekv4flash

PLEASE FOLLOW ME:
▶ LinkedIn: / fahdmirza
▶ YouTube: / @fahdmirza
▶ Blog: fahdmirza.com

RESOURCES:

▶ fahdmirza.com

All rights reserved © Fahd Mirza