Abdullah Abdelrhim

abdullah

6 263 998

abodacs

AI & ML interests

None yet

Recent Activity

upvoted a collection 1 day ago

📝 FinePhrase

liked a Space 1 day ago

Omartificial-Intelligence-Space/llm-glossary

reacted to sergiopaniego's post with 🔥 1 day ago

New post-training example: SFT on agent traces with TRL → A large model works as a coding agent in OpenCode, run through OpenEnv → The tasks it solves become the training data → A 2B model goes from 5.5% to 34.4% on held-out tasks About 500 traces get most of the gain. Five times more add only 3 to 4 points, so SFT gets the agent working and RL has to do the rest Full guide, with the models, traces and every rollout open: https://github.com/huggingface/trl/tree/main/examples/sft_agent_traces

View all activity

Organizations

upvoted a collection 1 day ago

📝 FinePhrase

Collection

Synthetic rewrites of FineWeb-Edu documents for language model pre-training. • 4 items • Updated 3 days ago • 1

upvoted a paper 15 days ago

EmbeddingGemma 2: Unified and Efficient Multimodal Representations

Paper • 2610.37364 • Published 18 days ago • 37

upvoted an article about 2 months ago

Article

Know Who Spoke When: Build Real-Time, Multi-Speaker AI with NVIDIA Nemotron 3 Diarization

nvidia

•

Sep 23

• 73

upvoted a paper 2 months ago

Repo-To-Skill: Distilling GitHub Repositories Into AI4AI Skills

Paper • 2609.02749 • Published Sep 2 • 148

upvoted a paper 3 months ago

Inject, Align, Recover: Staged Post-Training for Retrieval-Free Document Knowledge Internalization

Paper • 2608.20281 • Published Aug 20 • 9

upvoted an article 3 months ago

Article

Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers

tomaarsen, NohTow, raphaelsty

•

Aug 18

• 120

upvoted a collection 3 months ago

Saudi Arabia dataset

Collection

7 items • Updated Dec 11, 2025 • 1

upvoted an article 3 months ago

Article

On the integrity of the Arabic TTS Arena leaderboard

Navid-AI

•

Aug 7

• 7

upvoted a paper 3 months ago

Wnuan: Staged Post-Training for Question Answering over Proprietary Enterprise Knowledge

Paper • 2608.01862 • Published Aug 3 • 6

upvoted an article 3 months ago

Article

Deploy local agents everywhere with LFM2.5-2.6B

LiquidAI

•

Aug 4

• 101

upvoted an article 4 months ago

Article

Distillation in 2026 (so far): which frontier models use it and how

sergiopaniego

•

Jul 8

• 25

upvoted a collection 4 months ago

Qari-OCR: A High-Accuracy Model for Arabic Optical Character

Collection

𝐵𝑢𝑖𝑙𝑡 𝑜𝑛 𝑡ℎ𝑒 𝑝𝑜𝑤𝑒𝑟𝑓𝑢𝑙 𝑄𝑤𝑒𝑛2 𝑉𝐿 2𝐵 𝑎𝑛𝑑 𝑓𝑖𝑛𝑒-𝑡𝑢𝑛𝑒𝑑 𝑜𝑛 𝑎𝑛 𝐴𝑟𝑎𝑏𝑖𝑐 𝑂𝐶𝑅 𝑑𝑎𝑡𝑎𝑠𝑒𝑡, 𝑄𝑎𝑟𝑖 𝑣0.1 𝑑𝑒 • 9 items • Updated Sep 7 • 22

upvoted 3 collections 5 months ago

upvoted 2 articles 5 months ago

Article

Agentic Resource Discovery: Let agents search

burtenshaw, evalstate

•

Jun 17

• 24

Article

There is no such thing as a tokenizer-free lunch

catherinearnett

•

Sep 25, 2025

• 103

upvoted an article 7 months ago

Article

How we OCR'ed 30,000 papers using Codex, open OCR models and Jobs

nielsr

•

Apr 7

• 62

upvoted 2 papers 8 months ago

OpenResearcher: A Fully Open Pipeline for Long-Horizon Deep Research Trajectory Synthesis

Paper • 2603.20278 • Published Mar 17 • 103

V_1: Unifying Generation and Self-Verification for Parallel Reasoners

Paper • 2603.04304 • Published Mar 4 • 14

Abdullah Abdelrhim

AI & ML interests

Recent Activity

Organizations

abdullah's activity

**Know Who Spoke When: Build Real-Time, Multi-Speaker AI with NVIDIA Nemotron 3 Diarization**

Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers

On the integrity of the Arabic TTS Arena leaderboard

Deploy local agents everywhere with LFM2.5-2.6B

Distillation in 2026 (so far): which frontier models use it and how

Agentic Resource Discovery: Let agents search

There is no such thing as a tokenizer-free lunch

How we OCR'ed 30,000 papers using Codex, open OCR models and Jobs

Know Who Spoke When: Build Real-Time, Multi-Speaker AI with NVIDIA Nemotron 3 Diarization