A.I Joe

AI-Joe-git

33 5 23

https://www.blackriver-ai.com/

AI & ML interests

https://www.blackriver-ai.com/

Recent Activity

liked a model about 2 hours ago

FINAL-Bench/POCKET-EN-GGUF

liked a model 1 day ago

baseten/glm-5-2-projector

updated a Space 5 days ago

AI-Joe-git/blackriver-gguf-tools

View all activity

Organizations

None yet

liked a model about 2 hours ago

FINAL-Bench/POCKET-EN-GGUF

Text Generation • 23B • Updated about 15 hours ago • 359 • 24

liked a model 1 day ago

baseten/glm-5-2-projector

Updated 8 days ago • 3

updated a Space 5 days ago

BlackRiver GGUF Tools

🏴

Convert and quantize HuggingFace models to GGUF format

liked a Space 9 days ago

BlackRiver Gemma 4 Multimodal

🌊

Gemma 4 E2B/E4B multimodal inference on WebGPU.

updated a Space 9 days ago

BlackRiver Gemma 4 Multimodal

🌊

Gemma 4 E2B/E4B multimodal inference on WebGPU.

published a Space 9 days ago

BlackRiver Gemma 4 Multimodal

🌊

Gemma 4 E2B/E4B multimodal inference on WebGPU.

New activity in Intel/Qwen3.6-35B-A3B-int4-mixed-AutoRound 11 days ago

Intel hardware compatible?

👀 2

#3 opened 2 months ago by

AI-Joe-git

liked a model about 2 months ago

huihui-ai/Huihui-gemma-4-26B-A4B-it-abliterated

Any-to-Any • 27B • Updated May 12 • 1.35k • 47

New activity in SassyDiffusion/Qwen3.6-35B-A3B-heretic-MTP-GGUF 2 months ago

New version?

#1 opened 2 months ago by

AI-Joe-git

New activity in unsloth/Qwen3.6-27B 2 months ago

What is the difference between this and the original Qwen3.6-27B?

#1 opened 3 months ago by

crazyi

replied to DavidAU's post 3 months ago

Promising Qwen3.6-43b-a3b variants could outperform the 27B while retaining the 35B’s speed.

liked a model 3 months ago

mradermacher/Darwin-36B-Opus-ABLITERATED-HERETIC-i1-GGUF

35B • Updated Apr 30 • 816 • 18

replied to DavidAU's post 3 months ago

THANKS for sharing this ! any plan for the Qwen 3.6 35B A3B Variant ?

reacted to DavidAU's post with ❤️ 3 months ago

Post

22792

Uncensored, Heretic, Qwen 3.6 27B GGUFs - Exceeds all quant metrics and core model metrics too.

Tuned 27B Heretic Uncensored quants from IQ2M to Q8.
IQ2M is 83% of BF16, with Q6 just under 98% of BF16 precision.
Q8: 98.47% of BF16 precision.
NEO/Code DI-Imatrix Quants.

Exceeds all 5 metrics for "censored" quants too.

All metrics posted.

Tuned model -from which the quants were built- also exceeds Qwen 3.6 27B core metrics too.

DavidAU/Qwen3.6-27B-Heretic-Uncensored-FINETUNE-NEO-CODE-Di-IMatrix-MAX-GGUF