Mamy Ratsimbazafy

mratsim

·

AI & ML interests

Slop Hunter, Creative Sampler, Compressor of Weights, Bearer of the Quantizer Torch

Recent Activity

new activity 5 days ago

RadixArk/Qwen3.8-Flash-Next-NVFP4:Optimized build 1x RTX Pro 6000: 177 to up to 340 tok/s, concurrency=4 @ 800 tok/s aggregate, 939K KV cache + RAM offloading

liked a model about 1 month ago

zerofata/G4-MeroMero-v2-31B

liked a model about 1 month ago

nvidia/NVIDIA-NemotronLabs-VoiceChat-11B

View all activity

Organizations

None yet

Collections 8

View 8 collections

models 44

mratsim/Qwen3.5-397B-A17B-EXL3

Image-Text-to-Text • Updated May 2 • 26 • 11

mratsim/GLM-4.7-EXL3

Text Generation • Updated Apr 17 • 48 • 21

mratsim/MiniMax-M2.5-BF16-INT4-AWQ

Text Generation • 39B • Updated Feb 17 • 245 • 43

mratsim/MiniMax-M2.5-FP8-INT4-AWQ

Text Generation • 39B • Updated Feb 17 • 1.21k • 21

mratsim/GLM-4.7-Flash-FP8

Text Generation • 31B • Updated Jan 26 • 791 • 1

mratsim/MiniMax-M2.1-BF16-INT4-AWQ

Text Generation • 39B • Updated Jan 14 • 19 • 7

mratsim/MiniMax-M2.1-FP8-INT4-AWQ

Text Generation • 39B • Updated Jan 14 • 31 • 39

mratsim/Hearthfire-24B-NVFP4A16

Text Generation • 14B • Updated Dec 19, 2025 • 8 • 1

mratsim/Codex-24B-Small-3.2-NVFP4A16

Text Generation • 14B • Updated Dec 19, 2025 • 13

mratsim/Hearthfire-24B-NVFP4

Text Generation • 14B • Updated Dec 19, 2025 • 38

datasets 0

None public yet