Y Hacker News new | ask | show | jobs

user: hurrycane created: 2010-10-27 karma: 1715 [ my public key: https://keybase.io/hurrycane; my proof: https://keybase.io/hurrycane/sigs/tWmkUlyu3vIEZY7yIxQInC0U03kfYidKygli6Db98yQ ] submissions: Insights from Multilingual Curation for a 20T-Token Dataset 1 points | 0 comments DatBench fixes VLM evals: 70% blindly solvable, 42% mislabeled, 35% prod gap 5 points | 0 comments DatBench: Cut VLM eval compute by >10× while INCREASING signal 4 points | 0 comments Luxical: Lexical-Dense Embeddings for Web-Scale Data Curation (3×–100× Faster) 3 points | 0 comments Arcee Trinity Mini: US-Trained Moe Model 70 points | 15 comments BeyondWeb: Lessons from Scaling Synthetic Data for Trillion-Scale Pretraining 1 points | 0 comments BeyondWeb: Lessons from Scaling Synthetic Data for Trillion-Scale Pretraining 3 points | 0 comments Image-Text Curation for 1B+ Data: Faster, Better, Smaller Clip Models 12 points | 0 comments Zico Kolter Joins OpenAI's Board of Directors 3 points | 0 comments 0 points | 0 comments Augmenting Segment customer data with behavioral signals using the Moonsense SDK 1 points | 0 comments Moonsense Recorder – Build live prototypes using mobile device sensor data 2 points | 0 comments From the Gym to a Jupyter Notebook – Building a Squats Counter App in a Day 7 points | 0 comments WeWork to Become Publicly Traded via SPAC 2 points | 0 comments Reducing indexing latency of Twitter Search to one second 3 points | 0 comments Twitter meets TensorFlow 206 points | 68 comments 5 steps to better GCP network performance 2 points | 0 comments Feature Visualization 2 points | 0 comments AVA: A Finely Labeled Video Dataset for Human Action Understanding 44 points | 6 comments 0 points | 0 comments Nestlé Targets High-End Coffee by Taking Majority Stake in Blue Bottle 2 points | 0 comments CNTK version 2.0 released Featuring Keras 1 points | 0 comments Deep Learning with Emojis (not Math) 5 points | 1 comments TensorFlow v1.1.0 has been released 2 points | 0 comments A TensorFlow implementation of the Differentiable Neural Computer 157 points | 29 comments Tf-Seq2seq: An Open Source Sequence-To-Sequence Framework in TensorFlow 3 points | 0 comments Cloudera S-1 4 points | 0 comments Google Cloud Functions: a serverless environment to connect cloud services 1 points | 0 comments