qwen3.5-amharic-4b-thinking-uncensored
abdukuzi45
Language model for Amharic. Fine-tuned from Qwen/Qwen3.5-4B.
abdukuzi45
Builder · part of the Ethiopian AI ecosystem
🇪🇹 Amharic CPT Packed Text 2048 v9 High-quality packed and pre-tokenized Amharic dataset for Continued Pre-training (CPT) of Large Language Models. This dataset contains 4,753,149 chunks of exactly
2
models
5
datasets
0
repos
0
spaces
1,408
downloads/mo
Footprint shape — each axis normalized against the ecosystem-wide maximum. A fuller polygon means this org publishes across every channel.
