hfm8tr/smollm2-135m-smoltalk-lora
Adapter vs HuggingFaceTB/SmolLM2-135M. hfm8tr/smollm2-135m-smoltalk-lora is a LoRA adapter, not full weights: rank 8, on q_proj, v_proj in 30 layers (60 adapted matrices). It modifies HuggingFaceTB/SmolLM2-135M at load time.