Baichuan-13B-Chat
by baichuan-inc
13.3B params · text-generation · 632 likes · 7.9k downloads
Baichuan-13B-Chat is a 13.3B parameter model. At Q4 quantization it requires 7GB of VRAM. It runs comfortably on GeForce RTX 4090 (98 tok/s), GeForce RTX 5090 (148 tok/s), M4 Max 128GB (36 tok/s).
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| GeForce RTX 4090 | 24GB | 98 tok/s | comfortable |
| GeForce RTX 5090 | 32GB | 148 tok/s | comfortable |
| M4 Max 128GB | 128GB | 36 tok/s | comfortable |
| M4 Pro 48GB | 48GB | 18 tok/s | tight |
| M4 Pro 24GB | 24GB | 18 tok/s | tight |
| A100 PCIe 80 GB | 80GB | 181 tok/s | comfortable |
| H100 SXM5 80 GB | 80GB | 364 tok/s | comfortable |
| GeForce RTX 3090 | 24GB | 87 tok/s | comfortable |
| Radeon RX 7900 XTX | 24GB | 72 tok/s | comfortable |
| GeForce RTX 4080 | 16GB | 70 tok/s | comfortable |