Superseded by zenlm/zen3-nano — canonical name.
Compact language model for fast inference and edge deployment.
Repackaged from Qwen/Qwen3-0.6B (apache-2.0, Alibaba Qwen). Not trained from scratch — a permissively-licensed redistribution for the OSS-clean Zen model line.
| Property | Value |
|---|---|
| Parameters | 0.6B (dense) |
| Architecture | Qwen3 (Qwen3ForCausalLM) |
apache-2.0. Upstream: Qwen/Qwen3-0.6B by Alibaba Qwen (apache-2.0).
checkpoint-216/optimizer.pt 2.4 GBcheckpoint-240/optimizer.pt 2.4 GBoptimizer.pt 2.4 GBcheckpoint-216/model.safetensors 1.2 GBcheckpoint-240/model.safetensors 1.2 GBmodel.safetensors 1.2 GBcheckpoint-216/rng_state.pth 14.4 kBcheckpoint-240/rng_state.pth 14.4 kBrng_state.pth 14.4 kBcheckpoint-216/training_args.bin 5.7 kBcheckpoint-240/training_args.bin 5.7 kBtraining_args.bin 5.7 kBcheckpoint-216/scheduler.pt 1.5 kBcheckpoint-240/scheduler.pt 1.5 kBscheduler.pt 1.5 kBQwen/Qwen3-0.6B