ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
Use caseGPU requirements
Recommended card
Link
2
Running Falcon-40BGPU with 85-100GB+ VRAM (Video RAM)
See Falcon-40B GPU table
https://huggingface.co/tiiuae/falcon-40b
3
Running MPT-30B80GB for 16-bit precision
See MPT-30B GPU table
https://www.mosaicml.com/blog/mpt-30b
4
Training LLaMA (65B)
"A friend of mine is on the LLaMA team at FAIR. They had 8,000 Nvidia A100s at the time."
Very large H100 cluster
https://news.ycombinator.com/item?id=36339417
5
Training Falcon (40B)"384 A100 40GB GPUs"
Large H100 cluster
https://huggingface.co/tiiuae/falcon-40b
6
Fine tuning an LLM (large scale)"64 A100 40GB GPUs"H100 cluster
https://huggingface.co/TheBloke/falcon-40b-instruct-GGML
7
Fine tuning an LLM (small scale)"4x A100 80gb"H100 cluster
https://erichartford.com/uncensored-models
8
Stable Diffusion image generation"12GB+" or "16GB+"
See SD GPU table
https://timdettmers.com/2023/01/30/which-gpu-for-deep-learning/#GPU_Deep_Learning_Performance_per_Dollar
9
Whisper transcriptionMinimal, can run on CPU
Can run on a GPU. If you want it faster, see the Stable Diffusion GPU table
10
11
12
https://llm-utils.org/Nvidia+H100+and+A100+GPUs+-+comparing+available+capacity+at+GPU+cloud+providers
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100