darioooooo0o/Spark-X2.6-4B-CED

Source: 🤗🏴‍☠️ Rescued by: Sherpatext-generationapache-2.04B activated14 GBGGUF⚠ No SHA-256 weightsupdated 16d ago

🧲 Magnet🌱 2

Hugging Face didn't answer

Loading…

CommentsShare your thoughts and shape the future of sovereign AI.

@darioooooo0o is unclaimedSign in with the matching Hugging Face account or an organization member account to verify this handle.Claim @darioooooo0o

Files and sources

✓SourceAvailable on 🤗

2ChecksumsNone found

✓TorrentMagnet listed

✓Swarm2 seeders reported

Source revision: 2b48da1f2210…
Witnessed: Not yet
Swarm checked: 14 Nov 2026, 01:54 UTC
Weight files: 0
Model size: 14 GB
Downloads · last month: 282
Publisher: Unclaimed

Family

↑ Quantized from XHToken/Spark-X2.6-4B

Speed on real hardware

LocalMaxxing

Output speed for one request at a time, measured by people running this model on their own machines.

Intel Arc Pro B70✓llama.cpp · Q4_K_M · 2 runs90.1 tok/s

llama-server -m spark-ced-Q4_K_M.gguf -fa on -c 16384 -np 1 --jinja -ngl 99   (CED two-phase prefill, replay 1024; prompts under 1024 tokens do not show the prefill gain)

View run on LocalMaxxing →

Median of 2 community runs on 1 setups, from LocalMaxxing.

Run it yourself: setup guide →