Gabriele Berton ✓

@gabriberton · Mountain View, CA · joined 13 Dec 2021

Training VLMs at @GoogleDeepMind - ex @Amazon @CarnegieMellon @PoliTOnews @IITalk - I post honest non-AI-generated paper reviews

17 401Followers

1 206Following

3 757Posts total

952KViews on collected posts

Últimas publicaciones

Gabriele Berton ✓@gabriberton9 Sep

@gulvarol starting now at the Video4Real workshop! Rooms 504-505 at @eccvconf #ECCV https://t.co/esLyIGO5Eb

1.4K views · 13 likes · 1 reposts · 1 replies Open on X →

Gabriele Berton ✓@gabriberton9 Sep

Viorica Patraucean starting now at the Video4Real workshop! She's speaking about video understanding efforts at Google DeepMind Rooms 504-505 at @eccvconf #ECCV, the room is already packed! https://t.co/lTWJWw6mI8

1.6K views · 15 likes · 3 reposts · 1 replies Open on X →

Gabriele Berton ✓@gabriberton9 Sep

@doughty_hazel starting now at the Video4Real workshop! Rooms 504-505 at @eccvconf #ECCV https://t.co/NoinRJfy1O

1.4K views · 17 likes · 3 reposts · 1 replies Open on X →

Gabriele Berton ✓@gabriberton7 Sep

@gabriberton I think that's good? Ppl can go around in adhoc groups (and maybe make new friends) instead of desperately trying to get into the 'VIP' company events. 499 views · 2 likes · 0 reposts · 0 replies Open on X →

Gabriele Berton ✓@gabriberton7 Sep

@gabriberton Is there any?? I didn’t find them anywhere. 573 views · 2 likes · 0 reposts · 0 replies Open on X →

Gabriele Berton ✓@gabriberton7 Sep

@gabriberton recession indicator 712 views · 8 likes · 0 reposts · 0 replies Open on X →

Gabriele Berton ✓@gabriberton7 Sep

How come there are so few social events at ECCV? 8.6K views · 49 likes · 0 reposts · 9 replies Open on X →

Gabriele Berton ✓@gabriberton31 May 2024

@gabriberton I don't think you always reduce GPU usage by half. In your situation, only the memory (of backward) used by Branch1 will be freed. The common network (before branching) will still occupy the same memory for both versions of the code. https://t.co/KXxL9cFk5V

3.2K views · 10 likes · 1 reposts · 1 replies Open on X →

Gabriele Berton ✓@gabriberton31 May 2024

Ok I'll use this - I'm looking for computer vision jobs/internships from November, anywhere in the world. I have industry experience (Amazon and more) and good publications (first authored 6 CVPR/ICCV). Background in VPR/IR but expanding my horizons. ~1000 stars on GitHub 13.2K views · 63 likes · 2 reposts · 1 replies Open on X →

Gabriele Berton ✓@gabriberton31 May 2024

Wow this is getting traction. I will post more tricks on how to write efficient code then 11.7K views · 56 likes · 0 reposts · 0 replies Open on X →

Gabriele Berton ✓@gabriberton31 May 2024

Limitations: this only works when you have more than one loss (with disentagled computational graphs). Bonus: the more losses you have, the more memory you'll save 37.8K views · 169 likes · 3 reposts · 8 replies Open on X →

Gabriele Berton ✓@gabriberton31 May 2024

This simple pytorch trick will cut in half your GPU memory use / double your batch size (for real). Instead of adding losses and then computing backward, it's better to compute the backward on each loss (which frees the computational graph). Results will be exactly identical http

871.3K views · 3.2K likes · 348 reposts · 48 replies Open on X →

Frente a cuentas del mismo tamaño

7 publicaciones de los últimos 90 días, junto al rango de 10K–100K seguidores. llega a mucha gente, pero pocos de esos espectadores reaccionan.

Visualizaciones medianas1 353esta cuenta1 026mediana de 10K–100K

Alcance, %7.78%esta cuenta3.75%mediana de 10K–100K

Interacción, %1.10%esta cuenta1.68%mediana de 10K–100K

Métrica	Esta cuenta	Mediana de 10K–100K	Proporción
Visualizaciones medianas por publicación	1 353	1 026	1.32×
Alcance (visualizaciones ÷ seguidores)	7.78%	3.75%	2.07×
Tasa de interacción	1.10%	1.68%	0.66×

Otras cuentas de este rango → Comparar con otra cuenta → Cómo se construyen estas referencias →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

871.3K31 May

37.8K

11.7K

13.2K

3.2K

8.6K7 Sep

712

573

499

1.4K9 Sep

1.6K

1.4K

Last 12 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.42%31 May

0.48%

0.50%

0.37%

0.67%7 Sep

1.12%

0.35%

0.40%

1.55%9 Sep

1.19%

1.10%

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.68%.

What the audience does

Likes50.0%3 650 in total

Reposts4.9%361 in total

Replies1.0%70 in total

Bookmarks44.1%3 215 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Últimas publicaciones

Frente a cuentas del mismo tamaño

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Cuentas similares