Model CoverageLarge Language ModelsMoonshot AIKimi-K2-Base

Kimi-K2-Base

View as Markdown

Kimi-K2-Base has a checked-in NeMo AutoModel recipe for text generation. The Hugging Face configuration declares the DeepseekV3ForCausalLM architecture.

Fine-Tune Kimi-K2-Base

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

automodel examples/llm_benchmark/kimi/kimi_k2_te_deepep.yaml --nproc-per-node 8

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

Choose a Workflow

GoalStart Here
Run the primary recipe for this modelUse the recipe configuration.
Prepare your environmentFollow the installation instructions.

Configuration

SettingConfiguration
Hardware-
StrategyFSDP2; tp_size=1, pp_size=8, cp_size=1, ep_size=32
Nodes32
FeaturesActivation checkpointing (activation_checkpointing=true), Transformer Engine, DeepEP (dispatcher=deepep)
Advancedattention=te, linear=te, experts=gmm, dispatcher=deepep

Model Reference

Model Architecture

PropertyValue
TaskText generation
ArchitectureDeepseekV3ForCausalLM

Available Models