GLM-5.2 | NVIDIA NeMo AutoModel

GLM-5.2 has a checked-in NeMo AutoModel recipe for text generation. The Hugging Face configuration declares the GlmMoeDsaForCausalLM architecture.

Fine-Tune GLM-5.2

Follow the installation instructions, then run the recipe from the repository root:

automodel examples/llm_finetune/glm/glm_5.2_lora.yaml --nproc-per-node 8

Use the Slurm launcher guide for the multi-node run.

Goal	Start Here
Run the primary recipe for this model	Use the recipe configuration.
Try another checked-in recipe for this model	Use the alternate recipe.

Setting	Configuration
Hardware	-
Strategy	FSDP2; tp_size=1, pp_size=1, cp_size=1, ep_size=128
Nodes	16
Features	Activation checkpointing (activation_checkpointing=true), Packed sequences (packed_sequence_size=4096), LoRA, HybridEP (dispatcher=hybridep)
Advanced	attention=tilelang, linear=torch, experts=torch_mm, dispatcher=hybridep, sequence length 4096

Property	Value
Task	Text generation
Architecture	`GlmMoeDsaForCausalLM`