Calibration-free VQ-256 compression. Transformers, SSMs, hybrids. Beats GPTQ/AWQ quality. Zero calibration data.
EchoLabs
non-profit
AI & ML interests
AI cross compression architecture
Recent Activity
View all activity
models 18
EchoLabs33/Zamba2-7B-Instruct-v2-GGUF
11B • Updated • 45
EchoLabs33/Zamba2-1.2B-Instruct-v2-GGUF
2B • Updated • 63
EchoLabs33/Zamba2-2.7B-Instruct-v2-GGUF
4B • Updated • 63
EchoLabs33/Zamba2-1.2B-Instruct-GGUF
2B • Updated • 47
EchoLabs33/olmoe-1b-7b-instruct-hxq
Text Generation • Updated • 413
EchoLabs33/zamba2-7b-instruct-hxq
Text Generation • Updated • 616
EchoLabs33/zamba2-2.7b-instruct-hxq
Text Generation • Updated • 639
EchoLabs33/zamba2-1.2b-hxq
Text Generation • Updated • 1.31k
EchoLabs33/mamba2-1.3b-hxq
Text Generation • Updated • 1.15k
EchoLabs33/mamba-130m-hxq
Text Generation • Updated • 1.37k
datasets 0
None public yet