Fine-tuning Gemma 4 for Hindi on a Free GPU — and shipping it to run on a CPU
pankajpandey-dev
• • 1NEW Articles from Team or Enterprise organizations will get promoted to the main section.
VLX-Seek: Improving VLM Fine-Grained Perception via Region Reference Instead of Coordinate Generation
omlab
• • 7Critique and Synthesis: The Military-Industrial State and the Hydraulic State
ajsbsd
• DukaanBench: Can AI Run an Indian Grocery Store for 30 Days?
77ethers
• • 2VLX-Flow: Continuous Video Understanding for Real-Time Multimodal Interaction
omlab
• • 9Prompt Phrasing Shifts Model Performance More Than Tier — and Which LLMs Will Get Your Jokes
Wayfinder6
• SportsBERT Small: Domain-specialized small models
NeuML
• • 2Here's the TTS-formatted version of your Cold War Erosion Simulation output:
ajsbsd
• OlmoLogic: Boosting Reasoning via RLVR with Inductive Logic Programming
LukasHug
• • 117 Models, $38.95, and One Paradox: What We Learned Training Compression Models in Public
PeetPedro
• SPROG-9M: how far a 9-million-parameter, LLM-free model gets on grade-school math
codelion
• • 1QHexRT Is Live: Full-Stack NPU Inference for Qualcomm Hexagon
runanywhere
• Performance Tuning for LLMs
EXDai
• LLM Inference Benchmarking
EXDai
• GB10 Workbench Setup
EXDai
• Subject Bucketing: Teaching a Diffusion Model New Prompt Languages Without Forgetting
AbstractPhil
• Fine-tuning DINOv2-Small on RSNA Pneumonia: the model-card decisions worth justifying
t22000t
• 🔁 LoopKit — build self-improving systems.
PeetPedro
• Giving Khasi a Digital Voice
toiar
• • 1How a 7B Model Beat a 70B Baseline by 46% — Building CodeReviewEnv
lucifer0077
•