Jonathan Fischoff
@jfischoff
Followers
3,551
Following
2,476
Media
1,160
Statuses
22,997
Computer person. AI @fal
San Diego, CA
Joined April 2008
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
Leyen
• 318645 Tweets
Ursula
• 248840 Tweets
#BamBam
• 184875 Tweets
DAOUOFFROAD COL EP4
• 128738 Tweets
Kamari
• 73446 Tweets
Ali Koç
• 71870 Tweets
#BAMESIS
• 70934 Tweets
岸田総理
• 47843 Tweets
Mehmet Büyükekşi
• 45561 Tweets
डिब्रूगढ़ एक्सप्रेस
• 41589 Tweets
Gill
• 39489 Tweets
国立公園
• 37000 Tweets
T20I
• 34828 Tweets
AMIA
• 34735 Tweets
Hardik
• 33612 Tweets
マンション爆発
• 33021 Tweets
高級リゾートホテル誘致
• 24736 Tweets
西川口の爆発
• 23690 Tweets
TFF Başkanı
• 20634 Tweets
#MaskShadesCap
• 19492 Tweets
Sanju
• 17944 Tweets
ODIs
• 17584 Tweets
Kimiko
• 15336 Tweets
What's your color
• 15204 Tweets
テレ東BIZ
• 13163 Tweets
Hulk Hogan
• 10857 Tweets
Fantastic result by
@kijaidesign
using
@CiaraRowles1
's SVD ControlNet.
This is using
@Kijaidesign
's new SVD CN Comfy node:
3
68
331
@datepsych
Probably more women understand this than men understanding how the older they are the more likely their sperm will have genetic defects.
7
7
303
Tencent released the model and code for their text2video system. There model is very similar to SD but with a 3D U-Net. I was most interested in the design of their U-Net, but there were a ton of details, however, I found this section interesting in light of more recent work.
Latent Video Diffusion Models for High-Fidelity Long Video Generation
github:
project page:
7
88
467
5
25
153
I just pushed a new AnimateDiff fine-tune that uses an alternative loss function I'm calling `diff_loss`. Get it here:
Why might you want to use a different loss function?
Read the🧵or check out the details in this blog post
11
20
148
@HappyFelineMeow
@SnarkySteff
@fmanjoo
@SopanDeb
At this point, considering our laws and I doubt what the cop did was against the law, and punishing him would be out of line with other acts we have sanctioned. The fake report should be punished, full stop yes. Cops at your house should not be a death sentence.
4
8
134
@Attractiveincel
@CartoonsHateHer
that's just what they said my man. you are as surprised as I was 🤷♂️
1
0
139
@nntaleb
Just be glad you weren’t stabbed.
The big plot twist is gonna be when we find out the attacker thought Salman was you.
0
0
121
Reading “FreeU” ().
The find that scaling the backbone and skip connect features to boost the backbone contribution and decrease the skip-connection low freq contribution has a large improvement on quality.
Pretty compelling examples.
5
18
111
Reading through the new Stability video model paper ().
I'm just going to comment on the things I find interesting in a 🧵 1/
2
19
109
@dieworkwear
He would get coffee at my coffeeshop in NYC and I’ve never seen a better dressed man in person.
0
0
99
I'm not a fan of the "Haskell is faster than C" blog posts.
It is almost always optimized Haskell against unoptimized C.
Haskell is a very good choice for many problems even if you can't optimize to the level you can with C/C++/Rust etc.
7
9
86
@GarettJones
I have a vague recollection of story where Rutherford, who he worked for, pointing out in public that the model was inherently unstable and Bohr had no counter to how the orbits where maintained (because they are not)
2
1
85
@nntaleb
Exercise is the stimulus. Recovery generates the benefit. All stimuli are dose dependent.
0
6
80
@RadishHarmers
Curious, do you think Israel should attempt to dismantle Hamas? They seem to care more about killing jews more than peace, not exactly sure how that can be tolerated.
Things were improving before this, so why attack now? They don't seem like rational actors to me.
60
2
74
Tried a quick little SVD "Motion Brush" experiment via inpainting.
Pipeline:
Masked just the middle portion, so there is no camera movement anymore. You can make out the coloration change if you look closely.
4
17
72
@jbarro
"If we lose this election, US democracy will cease to exist."
but also:
"If we lose, hey, at least we let an old man give it his best try. No biggie either way."
3
4
74
@zachweinberg
Flying once a month is a pretty big deal for parents. Lot’s of stress for the parent staying home.
6
0
72
I like the idea behind "Upsampling Diffusion Probabilistic Models" ().
Instead using a fixed sized noised sample, use a smaller sized sample for noiser timesteps
2
15
64
@moultano
This is like all the discourse around the declining birth rates and then you see it is just a straight line going back to 1800 so the actual factors are unrelated to whatever culture war story of the last month everyone is on about.
3
1
68
@AGHamilton29
I think the other question is whether it will be successful militarily.
It is not clear to me how the ground invasion will help destroy the tunnels when they are so extensive and sophisticated.
40
6
65
How to make custom AI APIs:
1. Make your workflow in Comfy.
2. Drag into
@fal
.
3. Label your inputs
profit?
all-in-one comfyUI experience
create workflows and turn them into scalable APIs, side by side
11
28
165
2
8
64
🚨New Model Alert!
I just released a new motion model for AnimateDiff.
It utilizes InstructPix2Pix and is designed to take in an image to animate.
model:
code:
3
16
62
Doesn't surprise me at all that the inventor of Javascript got equality all wrong again.
4
66
58
This paper buries the lead. If you use a negative prompt for only steps 5-15 out of 30 you get the best results.
5
7
57
@deech
Random relative: I’ve heard of it.
Me: What?!
Random relative: Yeah pascal. I’ve programmed in it.
Me: Right. So here’s the thing ...
2
2
53
Some improvements to the pipeline
6
5
53
Digging into the paper and I’ll give my take on how I think it works. I’m confused on a few details, so maybe you can help work them out.
First it uses IP-Adapter with the first frame, that is the Content Adapter in the diagram. Let’s move onto I2V Adapter. 🧵1/
2
4
49
@eigenrobot
@bryan_caplan
’s argument for open borders was coupled with providing zero resources to those that cross, while taxing the new entrants at significantly higher rates. He understood the tradeoffs a little better
3
2
47
@waitbutwhy
Body composition determines your health, not the food. All foods are unhealthy if you are consuming too many calories.
If you are eating maintenance calories, get your essential fats, protein, minerals and vitamins, then all foods are healthy.
16
0
46
@tanger1ng
@GarettJones
I don't think it makes sense at all because the electrons can't orbit indefinitely in a stable way.
2
0
44
I want to tweet something about 10x engineers but I’m not exactly sure how to say it.
Most businesses workflows are so onerous that the vast majority of engineers are operating well below their maximum efficiency.
1
13
47
@AliceFromQueens
@keithgessen
Really? The author feeling a need to air such private matters by itself is a huge red flag to me. Seems incredibly narcissistic. Why is this something their kids classmates need to read?
2
0
44
@tolstoybb
“he only needs to pump the price for 4% and dump on retail to make a profit. He’s a genius!”
Truly inspired investing in the future.
0
0
45
@Freezintweets
@quantian1
Yes, in the sense that the judge could order Elon to pay damages instead of completing the deal. The damages could the diff between merger mk and current.
3
0
44
@growing_daniel
That’s nothing.
I, a mancoder, routinely crush my keyboard with my man strength. Every bug sends me into a violent rage. Plus, I barely have time to code, since the entirety of my day is spent belittling my coworkers through PR review to assert dominance.
1
3
42
@jonst0kes
CDC is communicating the risk if you know how to listen for it.
They are saying “mitigation is likely” and “we will probably need to rely on social distancing” but most people don’t know what that means.
They should say a disruptive pandemic is coming and the to prepare is now
5
13
37
You'll have to translate the post, but they optimized the LCM img2img workflow to get 45fps a second with 4090.
apparently this is some of the code but not all of the improvements:
3
9
41