Alsa ⚡️ (@alsaai_eth) header
Alsa ⚡️ (@alsaai_eth) profile picture

Optimizing Machine Learning.

Earth
Joined August 2021
Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)Photo by Alsa ⚡️ (@alsaai_eth)
Alsa ⚡️ retweeted
We need AI regulations more than ever 😂
5
5
3
15
1,000
Alsa ⚡️ retweeted
Latent Navigation Exploring CLIP text space with FLUX.1 schnell 🪐 *FREE* *Open Source* huggingface.co/spaces/latent…
1
6
22
4,185
Alsa ⚡️ retweeted
🧹 Room Cleaner Upload an image and use the pencil tool (✏️ icon at the bottom) to mark the areas you want to remove. For best results, include the shadows and reflections of the objects you want to remove. You can remove multiple objects at once. *Open Source* huggingface.co/spaces/Hedro/…
1
5
12
1,708
Alsa ⚡️ retweeted
Yi-Coder-9B-Chat Demo! 9B parameter model fine-tuned for coding tasks. This demo showcases its ability to generate code based on your prompts. Yi is a series of open-source code language models that delivers state-of-the-art coding performance. huggingface.co/spaces/Tonic/…
1
11
18
5,352
Alsa ⚡️ retweeted
I've been thinking about this question. Why does M365 Copilot block links, but the go-to method successfully prompts injection to render an image or link? It's quite an interesting puzzle, don't you think?
the go-to method for data exfil after a successful prompt injection is rendering an image or a clickable link that's why m365 copilot refuses to print links no matter what unless of course..
1
5
13
3,603
Alsa ⚡️ retweeted
🚨 Please be careful when you download any LLM model open source. This model Llama-3-8B-Web has 2 files that have been marked as unsafe by @huggingface This files are unsafe in this model training_args.bin , pytorch_model.bin
2
5
12
1,225
Alsa ⚡️ retweeted
Fashion Photography AWPortrait-FL is finetuned on FLUX.1-dev using the training set of AWPortrait-XL and nearly 2,000 fashion photography photos with extremely high aesthetic quality. *Open Source *FREE huggingface.co/spaces/vilari…
3
10
33
6,666
Alsa ⚡️ retweeted
AI models depend on the quality of data they’re given. Bad data leads to bad results, including bias and errors. For accurate, fair, and real world ready AI, good and diverse data is essential.
3
7
338
Alsa ⚡️ retweeted
Normalize not supporting scammers Normalize not supporting scammers Normalize not supporting scammers Normalize not supporting scammers Normalize not supporting scammers Normalize not supporting scammers Normalize not supporting scammers
9
20
3
101
2,464
Alsa ⚡️ retweeted
Cybersecurity assistant This cybersecurity assistant is a Mistral Fine-tune model with 22,000 cybersecurity and hacking-related data pairs. It's trained in various topics like APT Management, Digital Forensics, Network Security, Malware Analysis, and more. huggingface.co/spaces/alsaet…
3
11
25
6,252
Alsa ⚡️ retweeted
Flux Realism LoRa *FREE* This repository provides a checkpoint with trained LoRA photorealism for FLUX.1-dev model by Black Forest Labs huggingface.co/spaces/alsaet…
4
11
27
5,898
Alsa ⚡️ retweeted
Text-to-Image *FREE* Phi-3.5-MoE is a lightweight, state-of-the-art open model built upon datasets used for Phi-3 - synthetic data and filtered publicly available documents - with a focus on very high-quality, reasoning dense data. huggingface.co/spaces/alsaet…
2
9
24
4,834
Alsa ⚡️ retweeted
Generative Cartoon Interpolation Open Source : *FREE huggingface.co/spaces/Doubii…
2
5
12
2,399
Alsa ⚡️ retweeted
In 10 years, you must have this line of code executed in your LLM model world_status = "AI Dominance" if ai_intelligence > human_intelligence else "Human Control"
1
4
7
268
Alsa ⚡️ retweeted
#AI is like digital mirror, reflects our faults
1
2
7
259
Alsa ⚡️ retweeted
This is crazy Open Source tool *FREE, high quality insane result. GO and test its SUPER FUN. Cancel all the paid AI you have. huggingface.co/spaces/KwaiVG…
3
1
4
242
Weekly Insight #8 Gary “Vee” Vaynerchuck Exposed The Manipulation and Deception King of NFTs and Web3 mariolongtin.substack.com/p/…
12
9
3
27
8,257
Alsa ⚡️ retweeted
LLM jailbreaking
1
4
1
7
295