Aeon Forge

AEON-7

·

AI & ML interests

Systems engineer. Security researcher. Quiet architect of things that probably shouldn't work — but do. Small models doing big things. Quantization, fine-tuning, skill training, agentic workflows, and generative CAD. Been at neural nets since 2013, still surprised weekly. The GPU hums, the coffee's on, and the door's open.

Recent Activity

new activity 3 days ago

AEON-7/Ornith-1.0-35B-AEON-Ultimate-Uncensored-BF16:It may generate garbled characters.

updated a model 11 days ago

AEON-7/Gemma-4-26B-A4B-it-Uncensored-NVFP4

updated a model 11 days ago

AEON-7/Gemma-4-12B-it-AEON-Abliterated-K4-NVFP4-FP8

View all activity

Organizations

None yet

New activity in AEON-7/Ornith-1.0-35B-AEON-Ultimate-Uncensored-BF16 3 days ago

It may generate garbled characters.

#2 opened 7 days ago by

New activity in AEON-7/Ornith-1.0-35B-AEON-Ultimate-Uncensored-NVFP4 13 days ago

Not actually abliterated - avoid this model

#2 opened 15 days ago by

New activity in AEON-7/Ornith-1.0-35B-AEON-Ultimate-Uncensored-BF16 15 days ago

Solid work — testing for mobile deployment

#1 opened 17 days ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4 26 days ago

The first uncensored that's actually good

#8 opened about 1 month ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-Multimodal-NVFP4-MTP-XS about 1 month ago

Benchmarks not useful for real world usage

#6 opened about 1 month ago by

New activity in AEON-7/Step-3.7-Flash-AEON-Ultimate-Abliterated-NVFP4 about 1 month ago

output is garbled and nonsensical

#1 opened about 1 month ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-Multimodal-NVFP4-MTP-XS about 1 month ago

Error with aeon-vllm-ultimate:latest on dgx spark

#5 opened about 1 month ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4 about 1 month ago

is it possible to apply --kv-cache-dtype fp8

#6 opened about 1 month ago by

Kudos — combined your model with DFlash/AEON for GB10

#7 opened about 1 month ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-Multimodal-NVFP4-MTP-XS about 2 months ago

100 tok/s rtx 5090

#4 opened about 2 months ago by

New activity in AEON-7/Gemma-4-26B-A4B-it-Uncensored-NVFP4 about 2 months ago

This model is not uncensored at all

#2 opened about 2 months ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16 about 2 months ago

Benchmarks and Chain of Thought

#8 opened about 2 months ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-Multimodal-NVFP4-MTP-XS about 2 months ago

Issue with Agentic Coding / Tool Calling - Model Hangs

#2 opened 2 months ago by

New activity in AEON-7/Qwen3.6-35B-A3B-heretic-NVFP4 about 2 months ago

Image crash

#1 opened 3 months ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-Multimodal-NVFP4-MTP-XS 2 months ago

Feedback: Incredible project for Blackwell (ASUS GX10)

#3 opened 2 months ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-Multimodal-NVFP4-MTP 2 months ago

CUDA Graph capture failed

#1 opened 2 months ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4 2 months ago

The DGX Spark didn't reach the benchmark speed.

#5 opened 3 months ago by

New activity in AEON-7/Gemma-4-26B-A4B-it-Uncensored-NVFP4 2 months ago

FP8 abliterated variant on the roadmap?

#1 opened 2 months ago by

New activity in AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-NVFP4 2 months ago

Using RTX 5090

#3 opened 3 months ago by

[Request] modelopt NVFP4 + MTP variant with proper SSM layer preservation

#4 opened 3 months ago by