Aman Priyanshu PRO

AmanPriyanshu's activity

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

posted an update 4 months ago

view post

Post

458

Stratified LLM Subsets: Balanced Training Data at 100K-1M Scale

Released three training datasets using embedding-based k-means clustering to create balanced subsets from large-scale corpora:

Interactive cluster visualization:
https://amanpriyanshu.github.io/Stratified-LLM-Subsets-100K-1M-Scale/

Pre-Training (FineWeb-Edu + Proof-Pile-2)
AmanPriyanshu/stratified-kmeans-diverse-pretraining-100K-1M

Instruction-Following (Tulu-3 + Orca AgentInstruct)
AmanPriyanshu/stratified-kmeans-diverse-instruction-following-100K-1M

Reasoning (Llama-Nemotron with sqrt balancing)
AmanPriyanshu/stratified-kmeans-diverse-reasoning-100K-1M

Methodology: k-means clustering on Snowflake Arctic-embed-xs embeddings (100 iterations), selecting cluster centroids as representatives. Balancing applied to imbalanced datasets to reduce category dominance.

Available at 50k, 100k, 250k, 500k, and 1M scales.

See translation

• 

👍

1

1

+

Reply

posted an update 5 months ago

view post

Post

303

Interactive Procedural Terrain Generation

While most of us work with neural networks for generation, there's something fascinating about how mathematical noise functions can create realistic landscapes. Built interactive demos exploring alternatives to basic Perlin noise: domain warping, ridged multifractals, and exponential slope weighting.

Each technique transforms the same input noise into completely different terrain characteristics - from smooth rolling hills to dramatic alpine ridges. Includes real-time parameter controls and performance analysis for understanding the computational tradeoffs.

Live demos: https://amanpriyanshu.github.io/The-Mountains-of-Madness/

Source: https://github.com/AmanPriyanshu/The-Mountains-of-Madness"

See translation

• 

Reply