Struct DistributedMetadata 

pub struct DistributedMetadata {
    pub world_size: usize,
    pub rank: usize,
    pub process_group_info: HashMap<String, String>,
    pub dp_size: usize,
    pub tp_size: usize,
    pub pp_size: usize,
    pub fsdp_sharding: HashMap<String, Vec<usize>>,
}

Metadata about the distributed training state

Fields

world_size: usize

World size when checkpoint was created

rank: usize

Rank of the process that created the checkpoint

process_group_info: HashMap<String, String>

Process group information

dp_size: usize

Data parallel group size

tp_size: usize

Tensor parallel group size

pp_size: usize

Pipeline parallel group size

fsdp_sharding: HashMap<String, Vec<usize>>

FSDP sharding information

Trait Implementations

impl Clone for DistributedMetadata

fn clone(&self) -> DistributedMetadata

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for DistributedMetadata

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for DistributedMetadata

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for DistributedMetadata

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.