TeeZee
/

GALAXY-16B-v1.0-bpw8.0-h8-exl2

Text Generation

Not-For-All-Audiences

Inference Endpoints

text-generation-inference

Model card Files Files and versions Community

Edit model card

GALAXY-16B-v1.0

Technical notes

72 layers,DUS procedure, mistral(32)->SOLAR(48)->GALAXY(72)
16B parameters
model created as a extension of depth upscaling procedure used for SOLAR by upstage

Results

model can and will produce NSFW content
waiting for eval results

Downloads last month: 0

Datasets used to train TeeZee/GALAXY-16B-v1.0-bpw8.0-h8-exl2

Collection including TeeZee/GALAXY-16B-v1.0-bpw8.0-h8-exl2

24 GB VRAM

Quants that run fast on single 3090/4090 card with 24GB of VRAM and 4096 context length • 14 items • Updated Apr 23 • 4