Mat (npub1lx3…ck4m)
Be prepared that Intel and AMD CPUs are frustratingly slower (up to 40 or even 100 times) than Huggingface servers with Nvidia Tesla T4 or Nvidia A10. Even the M1 or M2 in iPads with its 16 NPU (neural processing unit) is faster than many Intel CPUs.
The more complex your prompt text, the more computationally intensive.
#AI #ChatGPT #StableDiffusion and Intel? You won't enjoy this for long. Let a cloud or server render all. AI Render at home are far too weak and too expensive (electricity).
MichaelSwiss / Michael Swiss
npub1ka…yj790
2023-03-29 05:01:39
in reply to nevent1q…dhfy
Author Public Key
npub1karme3jc630epyn9eu7rr3mqx4h0z7c0wt52xw0954vkl78jfmqs2yj790Published at
2023-03-29 05:01:39Event JSON
{
"id": "a54daa0dbaf3b34ec0b2ecd70c09a53c1977bf3f71a30e7598db9f98858c0072",
"pubkey": "b747bcc658d45f909265cf3c31c760356ef17b0f72e8a339e5a5596ff8f24ec1",
"created_at": 1680066099,
"kind": 1,
"tags": [
[
"e",
"ff8ce312549035981d1067141d3a02dd87c8b689483cd57a7e4fbc68748cf670",
""
],
[
"e",
"ca8c18ff3c4c445e3dfd23d6267ce80c3a854b343dc7cca80cf6009ff0fd119d"
],
[
"p",
"d3af49435bb13c695232fd8ca8c9e73db8c65630553baef83f9544e81cef9be7",
""
],
[
"p",
"f9a352db4aa115ec5d330540dda37b71e2460cc0f65e3318fa3b244945dc8eb8"
],
[
"t",
"ai"
],
[
"t",
"chatgpt"
],
[
"t",
"stablediffusion"
]
],
"content": "#[3] \n\nBe prepared that Intel and AMD CPUs are frustratingly slower (up to 40 or even 100 times) than Huggingface servers with Nvidia Tesla T4 or Nvidia A10. Even the M1 or M2 in iPads with its 16 NPU (neural processing unit) is faster than many Intel CPUs.\n\nThe more complex your prompt text, the more computationally intensive.\n\n#AI #ChatGPT #StableDiffusion and Intel? You won't enjoy this for long. Let a cloud or server render all. AI Render at home are far too weak and too expensive (electricity).",
"sig": "412a87e3fc04d5f981630a0b4ea2d3b387399238e09c5ebb5ef44cc96cf2352b4b2af52ad7c9e9d76159fbc6b291d643e0ff8c1582f0563890e4c5dbe6ac7735"
}