viraldi.
Video details
How it works
Back to discovery
12:32
Technology & Computing / VIDEO

This Model Can Generate 1000+ Tokens/sec - How to Run DiffusionGemma Locally

AsapGuide· 26.1K subscribers

3d ago · en-US audio

👉 In this video, I will show you Google's DiffusionGemma, a diffusion language model that turns random characters into a full answer instead of writing one token at a time, with speeds above 1000 tokens per second in the best setup. I cover running it locally on your PC with Unsloth Desktop, what hardware it needs, and the limits I hit, like no tool calling and memory spilling into system RAM. I also look at Mercury…

Watch on YouTube
3.2×Score at discovery
3.4KViews
1.1KMedian · 41 videos

Outlier score

This score was measured on Tue, 27 Oct 2026 10:33:51 GMT. Current view counts may differ.

3.4K views ÷ a 1.1K baseline = 3.2×. Compared with 41 earlier videos from the same channel, matching the format recorded at scoring time. Older-upload baseline: reference videos had at least as long to gain views. This comparison is not age-matched.

YouTube snapshot: Tue, 27 Oct 2026 12:09:05 GMT. Counters may have changed.

Read the methodology

Generative AI & Agents

Channel-balanced percentile: 40

Comparison details

32 videos from 30 other channels, with each channel weighted equally. Same format and language, with 10,000–99,999 subscribers. Observed at 91 hours after publication, within ±18 hours.

This describes the collected sample, not all of YouTube. Updated Tue, 27 Oct 2026 12:46:49 GMT.

Niches & tags

Transcripts and detailed analytics are available for recent outlier Shorts up to three minutes. Research tools

Breakouts from other channels