Robert Fielding (@_rpf) header
Robert Fielding (@_rpf) profile picture

Robert Fielding

@_rpf

mostly technology, mostly RTs

 manchester to auckland
state-transition.com
Joined March 2010
  • Tweets761
  • Following524
  • Followers111
17Media
Photo by Robert Fielding (@_rpf)Photo by Robert Fielding (@_rpf)Photo by Robert Fielding (@_rpf)Photo by Robert Fielding (@_rpf)Photo by Robert Fielding (@_rpf)Photo by Robert Fielding (@_rpf)Photo by Robert Fielding (@_rpf)Photo by Robert Fielding (@_rpf)
Robert Fielding retweeted
loackme (@loackme_) profile picture
loackme@loackme_8 Sep 2023
🌆 SPACE\TIME #33
5
40
1
314
11,236
Robert Fielding retweeted
Jake Fried (@jakejfried) profile picture
Jake Fried
@jakejfried26 Aug 2023
One of my all-time favorite animations "The Thief and The Cobbler" (excerpt) Richard Williams, 1960s-90s
52
296
13
1,637
94,896
Robert Fielding retweeted
Justin Alvey (@justLV) profile picture
Justin Alvey
@justLV18 Jul 2023
I “jailbroke” a Google Nest Mini so that you can run your own LLM’s, agents and voice models. Here’s a demo using it to manage all my messages (with help from @onbeeper) 🔊 on, and wait for surprise guest! I thought hard about how to best tackle this and why, see 🧵
364
2,412
523
13,723
1,700,119
Robert Fielding retweeted
loackme (@loackme_) profile picture
loackme@loackme_11 Aug 2023
↺↻
8
27
1
187
9,604
Robert Fielding retweeted
loackme (@loackme_) profile picture
loackme@loackme_3 Aug 2023
Spindle teia.art/objkt/68001
8
57
465
15,893
Robert Fielding retweeted
Leander Herzog (@lennyjpg) profile picture
Leander Herzog
@lennyjpg1 Aug 2023
Zickzack, 2021 objkt.com/asset/hicetnunc/22… #glsl
2
16
117
4,755
Robert Fielding retweeted
Leander Herzog (@lennyjpg) profile picture
Leander Herzog
@lennyjpg31 Jul 2023
download all the things
9
101
2
744
37,716
Robert Fielding retweeted
Matt DesLauriers (@mattdesl) profile picture
Matt DesLauriers
@mattdesl31 Jul 2023
hidden line removal with signed distance fields—
5
26
353
17,972
Robert Fielding retweeted
loackme (@loackme_) profile picture
loackme@loackme_31 Jul 2023
SPACE\TIME coming to @verse_works on August 14th
5
14
1
120
4,208
Robert Fielding retweeted
Science girl (@sciencegirl) profile picture
Science girl
@sciencegirl1 Aug 2023
Pythagoras' theorem In a right-angled triangle, the square of the length of the longest side (called the hypotenuse) is equal to the sum of the squares of the other two sides (known as the legs). Watch that here
27
276
12
2,003
256,049
Robert Fielding retweeted
Oelhan (@oelhan_tv) profile picture
Oelhan@oelhan_tv1 Aug 2023
bave wucket
28
432
5
2,908
255,618
Robert Fielding retweeted
ʇʞz (@tuukzs) profile picture
ʇʞz
@tuukzs1 Aug 2023
A beautiful morning to you all fam ♥️ GM!
Image from ʇʞz (@tuukzs)
73
220
4
2,100
83,910
tunnelvision - wip 🎹🎛️🔊
25
109
4
607
29,816
Robert Fielding retweeted
Rikard Lindström (@rikard_io) profile picture
Rikard Lindström
@rikard_io26 Jul 2023
Grids. Fragment shader. #genartclub #creativecoding
Image from Rikard Lindström (@rikard_io)
Image from Rikard Lindström (@rikard_io)
Image from Rikard Lindström (@rikard_io)
7
16
184
9,978
Robert Fielding retweeted
Andrej Karpathy (@karpathy) profile picture
Andrej Karpathy
@karpathy25 Jul 2023
Yay, llama2.c can now load and inference the Meta released models! :) E.g. here inferencing the smallest 7B model at ~3 tokens/s on 96 OMP threads on a cloud Linux box. Still just CPU, fp32, one single .c file of 500 lines: github.com/karpathy/llama2.c expecting ~300 tok/s tomorrow :)
57
311
32
2,504
410,920
Robert Fielding retweeted
Andrej Karpathy (@karpathy) profile picture
Andrej Karpathy
@karpathy23 Jul 2023
My fun weekend hack: llama2.c 🦙🤠 github.com/karpathy/llama2.c Lets you train a baby Llama 2 model in PyTorch, then inference it with one 500-line file with no dependencies, in pure C. My pretrained model (on TinyStories) samples stories in fp32 at 18 tok/s on my MacBook Air M1 CPU.
Image from Andrej Karpathy (@karpathy)
89
680
102
4,953
1,082,517
Robert Fielding retweeted
Simon Willison (@simonw) profile picture
Simon Willison
@simonw19 Jul 2023
We have a winner! teapaw.org/AdrienBrault/status/16… The incantation I had missed: LLAMA_METAL=1 make
llama_new_context_with_model: max tensor size = 87.89 MB ggml_metal_add_buffer: allocated 'data ' buffer, size = 6984.06 MB, ( 6984.52 / 49152.00) ### Instruction: Ten great names for a pet lemur, with explanations \n### Response: Here are ten great names for a pet lemur, along with some information about each name and its meaning. 1. Luna - This name is perfect for a lemur because it evokes the moon and stars, which are often associated with nocturnal creatures like lemurs. Plus, "luna" means "moon" in Latin! ... llama_print_timings: load time = 14104.58 ms llama_print_timings: sample time = 371.20 ms / 548 runs ( 0.68 ms per token, 1476.31 tokens per second) llama_print_timings: prompt eval time = 2047.96 ms / 23 tokens ( 89.04 ms per token, 11.23 tokens per second) llama_print_timings: eval time = 18088.85 ms / 547 runs ( 33.07 ms per token, 30.24 tokens per second) llama_print_timings: total time = 20556.32 ms

ALT llama_new_context_with_model: max tensor size = 87.89 MB ggml_metal_add_buffer: allocated 'data ' buffer, size = 6984.06 MB, ( 6984.52 / 49152.00) ### Instruction: Ten great names for a pet lemur, with explanations \n### Response: Here are ten great names for a pet lemur, along with some information about each name and its meaning. 1. Luna - This name is perfect for a lemur because it evokes the moon and stars, which are often associated with nocturnal creatures like lemurs. Plus, "luna" means "moon" in Latin! ... llama_print_timings: load time = 14104.58 ms llama_print_timings: sample time = 371.20 ms / 548 runs ( 0.68 ms per token, 1476.31 tokens per second) llama_print_timings: prompt eval time = 2047.96 ms / 23 tokens ( 89.04 ms per token, 11.23 tokens per second) llama_print_timings: eval time = 18088.85 ms / 547 runs ( 33.07 ms per token, 30.24 tokens per second) llama_print_timings: total time = 20556.32 ms

Adrien Brault-Lesage (@AdrienBrault) profile pictureAdrien Brault-Lesage
@AdrienBrault19 Jul 2023
Replying to @simonw
Yes: gist.github.com/adrienbrault… Getting 20-25 token/s with huggingface.co/TheBloke/Llam… on M2 Max
Run Llama-2-13B-chat locally on your M1/M2 Mac with GPU inference. Uses 10GB RAM. UPDATE: see...

Run Llama-2-13B-chat locally on your M1/M2 Mac with GPU inference. Uses 10GB RAM. UPDATE: see...

Run Llama-2-13B-chat locally on your M1/M2 Mac with GPU inference. Uses 10GB RAM. UPDATE: see https://twitter.com/simonw/status/1691495807319674880?s=20 - llama2-mac-gpu.sh

gist.github.com
7
20
1
281
39,013
Robert Fielding retweeted
Pauric Freeman (@pfree05) profile picture
Pauric Freeman@pfree0513 Jul 2023
Looking forward to playing Complex Gallery, Dublin, next weekend #generativeart #audiovisual #touchdesigner #eurorack #synthesizer
11
54
8
499
36,419
Robert Fielding retweeted
Visakan Veerasamy (@visakanv) profile picture
Visakan Veerasamy
@visakanv11 Jul 2023
from time to time I think about this post, which I think remains in the top 0.01% of everything I have ever read
Image from Visakan Veerasamy (@visakanv)
375
20,027
654
169,516
10,897,306
Robert Fielding retweeted
おかず / Okazz (@okazz_) profile picture
おかず / Okazz
@okazz_9 Jul 2023
めっちゃかっこいいのできた #p5js #generative #creativecoding
2
30
2
265
19,377
Load more