3 ms·
Generated by GLM 5.3 NVFP4 (https://huggingface.co/local-inference-lab/GLM-5.3-NVFP4 https://huggingface.co/local-inference-lab/GLM-5.3-NVFP4) in 6 minutes, run
by CamperBob2 2mo ago
Generated by GLM 5.3 NVFP4 (https://huggingface.co/local-inference-lab/GLM-5.3-NVFP4 https://huggingface.co/local-inference-lab/GLM-5.3-NVFP4) in 6 minutes, running locally on 4x RTX6000: https://crimson-jeri-74.tiiny.site/ https://crimson-jeri-74.tiiny.site/
So, who trained on who?
- tmikaeld 2mo agoThat’s really impressive! I’d say even better than fable managed
- CamperBob2 2mo agoTo be fair, I misspoke, the NVFP4 quant I used for that test runs on 8 cards. Still...!