Mohamed Anwar

mootje's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

liked 2 Spaces 30 days ago

Sleeping

MCP

1

JASR v1.2 – Multi-dialectal Arabic ASR

☕

1

Robust speech-to-text for Arabic dialects

Disabled

Agents

1.7k

MiniMax H3 Turbo LoRA

🎬

1.7k

Video generation with a synchronized soundtrack

liked a Space 4 months ago

Running on Zero

Agents

Featured

58

Dots.tts

👁

58

Generate speech audio from text

liked 2 Spaces 9 months ago

Sleeping

Agents

3

Drag And Drop Kanban Board

🏢

3

📋 Drag-&-Drop Kanban Board for Interactive task management

Sleeping

Agents

3

AI Storyboard Creator

🎬

3

Create a storyboard from a text prompt

reacted to BramVanroy's post with 👀 over 2 years ago

view post

Post

2,232

The InstructGPT paper mentions that they insert 10% pretraining data during SFT, which they find improves the effect of PPO (IIUC). Has anyone else done later ablations on this? I've only seen the inverse suggested, mixing in SFT data during pretraining.

• 
• 
• 2 replies

·

👀

3

3

+

Reply

liked a Space almost 3 years ago

Running on Zero

MCP

82

LLM Agent from an Image

🤖

82

Get a LLM Assistant personality idea from an image

liked a Space about 3 years ago

Runtime error

MCP

Featured

611

Image To Story

👁

611

Create a story from any uploaded image