falcon-180B-chat
by tiiuae
180B params · text-generation · 545 likes · 65 downloads
falcon-180B-chat is a 180B parameter model. At Q4 quantization it requires 90GB of VRAM. It runs comfortably on H200 SXM 141 GB (39 tok/s).
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| M4 Max 128GB | 128GB | 2 tok/s | tight |
| H200 SXM 141 GB | 141GB | 39 tok/s | comfortable |
| M2 Ultra 192GB | 192GB | 3 tok/s | tight |