dphn/dolphin-2.9.1-yi-1.5-34b
dphn/dolphin-2.9.1-yi-1.5-34b
Parameters
37.5B
Context
8K
Downloads
4.5M
Likes
65
Architecture
llama
60 layers · 8 KV heads · head 128 · hidden 7168
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 98.7 GB RAM 148.1 · Disk 147.8 | 142.0 GB RAM 213.0 · Disk 173.8 | 368.7 GB RAM 553.0 · Disk 212.8 |
| FP88bit | 54.6 GB RAM 81.9 · Disk 93.4 | 95.9 GB RAM 143.8 · Disk 119.4 | 320.5 GB RAM 480.8 · Disk 158.4 |
| INT88bit | 54.6 GB RAM 81.9 · Disk 93.4 | 95.9 GB RAM 143.8 · Disk 119.4 | 320.5 GB RAM 480.8 · Disk 158.4 |
| AWQ4bit | 33.3 GB RAM 50.0 · Disk 67.2 | 73.7 GB RAM 110.5 · Disk 93.2 | 297.4 GB RAM 446.1 · Disk 132.2 |
| GPTQ4bit | 33.6 GB RAM 50.4 · Disk 67.6 | 74.0 GB RAM 110.9 · Disk 93.6 | 297.7 GB RAM 446.5 · Disk 132.6 |
| GGUF4bit | 34.2 GB RAM 51.3 · Disk 68.3 | 74.5 GB RAM 111.8 · Disk 94.3 | 298.3 GB RAM 447.4 · Disk 133.3 |
GPU Recommendations
1× H200 SXM
141 GB VRAM
≈ 265.3 tok/s
Effective production throughput ≈ 185.7 tok/s
score 0.6104
4× RTX 3090
96 GB VRAM
≈ 124.2 tok/s
Effective production throughput ≈ 86.9 tok/s
score 0.605
8× H200 SXM
1128 GB VRAM
≈ 849.1 tok/s
Effective production throughput ≈ 594.4 tok/s
score 0.2197
1× A100 80GB
80 GB VRAM
≈ 112.7 tok/s
Effective production throughput ≈ 78.9 tok/s
score 0.6407
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
A100 80GB
80 GB VRAM
$876
/mo · vast
H100 SXM
80 GB VRAM
$1,170
/mo · vast
H200 SXM
141 GB VRAM
$2,555
/mo · lambda
B200
180 GB VRAM
$3,013
/mo · vast
B200 SXM
180 GB VRAM
—