GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
EK-01
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
EK-01
Model Tree
HuggingFaceTB/SmolLM2-360M-Instruct
Input Modalities
Output Modalities
Supported Functionality
This model was merged using the DARE TIES merge method using SmolLM2-360M-Instruct structural layers as the foundational base.
The following models were included in the merge:
The following YAML configuration layout was used to produce this model:
yaml
merge_method: dare_tiesbase_model: HuggingFaceTB/SmolLM2-360M-Basemodels: - model: HuggingFaceTB/SmolLM2-360M-Instruct parameters: weight: 0.65 density: 1.0 # Keep 100% of the core grammar paths - model: HuggingFaceTB/SmolLM2-360M-Instruct parameters: weight: 0.35 density: 0.15 # Drops 85% of Instruct's facts, code, and safety bloat to isolate specific structural layersThis project is officially distributed under the Apache 2.0 License to ensure absolute compliance with upstream requirements. Huge credit to the Hugging Face Team for the exceptional SmolLM2 architecture.
merge_method: dare_tiesbase_model: HuggingFaceTB/SmolLM2-360M-Basemodels: - model: HuggingFaceTB/SmolLM2-360M-Instruct parameters: weight: 0.65 density: 1.0 # Keep 100% of the core grammar paths - model: HuggingFaceTB/SmolLM2-360M-Instruct parameters: weight: 0.35 density: 0.15 # Drops 85% of Instruct's facts, code, and safety bloat to isolate specific structural layers