D
distributed-llm-pretraining-torchtitan
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.
Installation
Pick a client and clone the repository into its skills directory.
Installation
About this skill
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.
Related skills
D
data-storytelling
by wshobson
Data Science
40,44926105
I
infographic-creation
by antvis
Data Science
6,98760199
X
xlsx
by anthropics
Data Science
180,62740128
M
market-analysis
by xbklairith
Data Science
329144
W
web-artifacts-builder
by anthropics
Data Science
180,62737124
S
skill-creator
by anthropics
Data Science
180,62759147