o
    ei4                     @   s  d dl Z d dlmZ d dlmZ d dlZd dlmZ d dl	m
Z
 d dlmZ d dlmZ d dlmZ d dlmZ d d	lmZ d d
lmZ d dlmZmZmZmZmZmZmZ d dlm Z m!Z! d dl"m#Z#m$Z$ d dl%m&Z& d dl'm(Z( d dl)m*Z*m+Z+m,Z, d dl-m.Z. d dl/m0Z0 d dl1m2Z2 d dl3m4Z4 e5e6e7ee8 dB ee8 f f Z9dgZ:d0de8de6de6fddZ;	d1dej<dB defddZ=dej>de?fd d!Z@	d0d"ed#ee8 de6dej>fd$d%ZAd&ede7e9ej<dB f fd'd(ZBG d)d* d*eZC	d1d+ed,e6d-e(d.e!dB def
d/dZDdS )2    N)Sequence)cast)_get_device_module)ShardedTensor)TensorProperties)Shard)ChunkShardingSpec)unflatten_state_dict)DefaultLoadPlanner)BytesStorageMetadataChunkStorageMetadataMetadataMetadataIndexSTATE_DICT_TYPEr   TensorStorageMetadata)LoadPlanLoadPlanner)_create_read_items create_read_items_for_chunk_list)load_state_dict)StorageReader)_element_wise_add_element_wise_sub_normalize_device_info)_get_default_group)_create_chunk_sharded_tensor)_remote_device)DTensor!load_sharded_optimizer_state_dictcudaglobal_rankdevice_typereturnc                 C   s2   |dkrdS t |}| rt|| |  S dS )Ncpu)r   is_availabler   device_count)r    r!   device_module r'   p/var/www/addictedbytheproject.nl/epg/venv/lib/python3.10/site-packages/torch/distributed/checkpoint/optimizer.py_gen_rank_device8   s   r)   pgc                    sh   t j j d u rfddtt  D }n fddt  D }tdtt	t
tB  |dS )Nc                    s"   g | ]}d | dt |  qS rank:/)r)   .0idx)pg_device_typer'   r(   
<listcomp>H   s    z(_create_colwise_spec.<locals>.<listcomp>c              
      s*   g | ]}d | dt t | qS r+   )r)   distget_global_rankr.   r*   r1   r'   r(   r2   M   s    r   dim
placements)r3   distributed_c10d_get_pg_default_devicetyperangeget_world_sizesizer   r   listr   str)r*   r8   r'   r5   r(   _create_colwise_specC   s   


rA   valc                 C   s   t | tu r.t|  dkrdS t |  d jtu rdS t |  d jtu r,tddS t | tu rFt | jtu sBt | jtu rFtddS )Nr   FTz1Cannot handle DTensor nested inside ShardedTensorzCannot handle nested DTensor)r;   r   lenlocal_shardstensorr   
ValueError_local_tensor)rB   r'   r'   r(   _is_nested_tensorW   s   rH   propsr>   c                 C   sP   |dkrt tjt| }n
t|t| }tj|| j| j| j| j	|dS )Nr#   )r>   dtypelayoutrequires_grad
pin_memorydevice)
r   torchrN   r   current_deviceemptyrJ   rK   rL   rM   )rI   r>   r!   rN   r'   r'   r(   _alloc_tensorf   s   rR   
state_dictc                 C   s   i }d}|   D ]9\}}d| f||< t|rAt| dks$tdt|ts-td| d }|jj	|jj
f||< |jj}q||fS )a+  
    Load the right TP slice of the optimizer state.

    This is not easy since the per-tensor slicing can't be inferred from checkpoint metadata.
    We take advantage of the model state_dict producing a sliced ST to figure out what we need to load.
    This is pretty fragile and it might be easier for FSDP to compute this info for us.
    Returns a dictionary where keys are the same of the state_dict and the value is a tuple of
    (offset, size) for the current rank TP slice.
    N.B. The state_dict *MUST* come from FSDP.sharded_state_dict.
    N   z%Cannot handle ST with multiple shardsz$Can only handle nested ShardedTensorr   )itemsr>   rH   rC   rD   AssertionError
isinstancer   metadatashard_offsetsshard_sizesrE   _process_group)rS   specsdp_pgkeyvalueshardr'   r'   r(   _get_state_dict_2d_layoutz   s$   
ra   c                       sz   e Zd ZU eeef ed< eed< eed< deee	e
 f ddf fddZdefd	d
Zdedejf fddZ  ZS )_ReaderWithOffsettranslationrS   rX   fqn_to_offsetr"   Nc                    s*   t    || _ti | _i | _i | _d S N)super__init__rd   r   rX   rS   rc   )selfrd   	__class__r'   r(   rg      s
   


z_ReaderWithOffset.__init__c                 C   s(  g }i | _ | j D ]\}}| jj| }t|ts"|t|||7 }q
|| jvr0|t|||7 }q
| j| }t	|
 dksAtd|
 d }ttt|jj|t|jjdg}t|tt||}|D ]$}	|	jjd u rrtdt|	jj|}
tj|	jt|
d}|| j |	j< qf||7 }q
t|S )NrT   z Expected exactly one local shardr   )offsetssizesz"dest_index.offset must not be None)offset)rc   rS   rU   rX   state_dict_metadatarW   r   r   rd   rC   rD   rV   r   rO   Sizer   rY   rZ   r   r   r   
dest_indexrm   r   dataclassesreplacer   )rh   requestsfqnobjmdrm   original_shardlocal_chunksreqsrioriginal_offsetoriginal_indexr'   r'   r(   create_local_plan   sD   


	
z#_ReaderWithOffset.create_local_planindexc                    s   t  | j||S re   )rf   lookup_tensorrc   get)rh   r~   ri   r'   r(   r      s   z_ReaderWithOffset.lookup_tensor)__name__
__module____qualname__dictr   __annotations__r   r   r@   r   intrg   r   r}   rO   Tensorr   __classcell__r'   r'   ri   r(   rb      s   
 " ,rb   model_state_dictoptimizer_keystorage_readerplannerc              	   C   sN  |  }t| \}}tj|j}t|}|du r?g }	tt D ]}
t	||
|
  }|	d|
 d|  q!td|	d}nt|}i }i }|j D ]\}}|j| }|d |kr\qLt|trfd||< qL|j dkrxt|j|j|||< qL|du rtt|j|j|t t |
 t d||< qL|d	 }||d|jfd }t|jj|jj|jj|jj|jj d
}|!t"#||}g }t|}|j$D ]}t%t&|j'( |krq|t)t|j|j*||d qt+j,|||d}||v r
|| d dur
t%t-t. || d ||< |||< qLt/|||durt0|n|d t1||j}|S )a  
    Load a state_dict in conjunction with FSDP sharded optimizer state.

    This is the current recommended way to checkpoint FSDP.
    >>> # xdoctest: +SKIP
    >>> import torch.distributed.checkpoint as dist_cp
    >>> # Save
    >>> model: torch.nn.Model
    >>> optim_params = model.parameters()
    >>> optim = torch.optim.SGD(optim_params, lr=0.01)
    >>> # Save
    >>> with FSDP.state_dict_type(model, StateDictType.SHARDED_STATE_DICT):
    >>>     state_dict = {
    >>>         "optimizer": FSDP.optim_state_dict(model, optim),
    >>>         "model": model.state_dict()
    >>>     }
    >>>     dist_cp.save_state_dict(
    >>>         state_dict=optim_state,
    >>>         storage_writer=dist_cp.FileSystemWriter("checkpoint"),
    >>>         planner=dist_cp.DefaultSavePlanner(),
    >>>     )
    >>>
    >>> # Load
    >>> with FSDP.state_dict_type(model_tp, StateDictType.SHARDED_STATE_DICT):
    >>>     model_state_dict = model_tp.state_dict()
    >>>     checkpoint = {
    >>>         "model": model_state_dict
    >>>     }
    >>>     dist_cp.load_state_dict(
    >>>         state_dict=checkpoint,
    >>>         storage_reader=dist_cp.FileSystemReader(checkpoint_file),
    >>>         planner=dist_cp.DefaultLoadPlanner(),
    >>>     )
    >>>     model.load_state_dict(checkpoint["model_state"])
    >>>
    >>>     optim_state = dist_cp.load_sharded_optimizer_state_dict(
    >>>         model_state_dict,
    >>>         optimizer_key="optimizer",
    >>>         storage_reader=dist_cp.FileSystemReader("checkpoint"),
    >>>     )
    >>>
    >>>     flattened_osd = FSDP.optim_state_dict_to_load(
    >>>        model, optim, optim_state["optimizer"]
    >>>     )
    >>>
    >>>     optim.load_state_dict(flattened_osd)
    Nr,   r-   r   r6   z
<bytes_io>rT   )rank
world_sizenum_devices_per_noder*      )rJ   rK   rL   memory_formatrM   )rE   rX   )process_group)rS   r   r   )2read_metadatara   r3   r9   r:   r;   r   r<   r=   r   r%   appendr   rA   rn   rU   planner_datarW   r   r>   numelrR   
propertiesr   get_rankr   r   ShardTensorPropertiesrJ   rK   rL   r   rM   build_metadatarO   ro   shards_metadatar   r   	placementr   r   rZ   r   +_init_from_local_shards_and_global_metadatar   r   r   rb   r	   )r   r   r   r   rX   layout_specsr]   dp_pg_device_typer&   r8   idevice_infosharding_specrS   rd   r^   r_   key_pathspec_key
alloc_sizer   st_mdrD   current_rankshard_mdstr'   r'   r(   r      s   5






	
)r   re   )Erq   collections.abcr   typingr   rO   torch.distributeddistributedr3   torch._utilsr   +torch.distributed._shard.sharded_tensor.apir   0torch.distributed._shard.sharded_tensor.metadatar   r   -torch.distributed._shard.sharded_tensor.shardr   :torch.distributed._shard.sharding_spec.chunk_sharding_specr   )torch.distributed.checkpoint._nested_dictr	   ,torch.distributed.checkpoint.default_plannerr
   %torch.distributed.checkpoint.metadatar   r   r   r   r   r   $torch.distributed.checkpoint.plannerr   r   ,torch.distributed.checkpoint.planner_helpersr   r   .torch.distributed.checkpoint.state_dict_loaderr   $torch.distributed.checkpoint.storager   "torch.distributed.checkpoint.utilsr   r   r   "torch.distributed.distributed_c10dr   #torch.distributed.fsdp._shard_utilsr   torch.distributed.remote_devicer   torch.distributed.tensorr   r   r@   tupler   STATE_DICT_2D_LAYOUT__all__r)   ProcessGrouprA   r   boolrH   rR   ra   rb   r   r'   r'   r'   r(   <module>   sz   $	 


#A