Remove the HivemindStrategy (#16407)

Remove the collaborative strategy
Lightning-AI · Jan 19, 2023 · 3b7b82e · 3b7b82e
1 parent b60c75e
commit 3b7b82e
Show file tree

Hide file tree

Showing 15 changed files with 0 additions and 942 deletions.
diff --git a/docs/source-pytorch/api_references.rst b/docs/source-pytorch/api_references.rst
@@ -283,7 +283,6 @@ strategies
     DDPStrategy
     DataParallelStrategy
     DeepSpeedStrategy
-    HivemindStrategy
     HPUParallelStrategy
     IPUStrategy
     ParallelStrategy

diff --git a/docs/source-pytorch/common_usecases.rst b/docs/source-pytorch/common_usecases.rst
@@ -123,13 +123,6 @@ Customize and extend Lightning for things like custom hardware or distributed st
    :button_link: clouds/cloud_training.html
    :height: 100
 
-.. displayitem::
-   :header: Train on multiple machines over the internet
-   :description: Train on local machines or unreliable GPUs across the internet.
-   :col_css: col-md-12
-   :button_link: strategies/hivemind
-   :height: 100
-
 .. displayitem::
    :header: Train on single or multiple GPUs
    :description: Train models faster with GPUs.

diff --git a/docs/source-pytorch/extensions/strategy.rst b/docs/source-pytorch/extensions/strategy.rst
@@ -72,9 +72,6 @@ The below table lists all relevant strategies available in Lightning with their
    * - bagua
      - :class:`~pytorch_lightning.strategies.BaguaStrategy`
      - Strategy for training using the Bagua library, with advanced distributed training algorithms and system optimizations. :ref:`Learn more. <accelerators/gpu_intermediate:Bagua>`
-   * - collaborative
-     - :class:`~pytorch_lightning.strategies.HivemindStrategy`
-     - Strategy for training collaboratively on local machines or unreliable GPUs across the internet. :ref:`Learn more. <strategies/hivemind:Training on unreliable mixed GPUs across the internet>`
    * - colossalai
      - :class:`~pytorch_lightning.strategies.ColossalAIStrategy`
      - Colossal-AI provides a collection of parallel components for you. It aims to support you to write your distributed deep learning models just like how you write your model on your laptop. `Learn more. <https://www.colossalai.org/>`__

diff --git a/docs/source-pytorch/index.rst b/docs/source-pytorch/index.rst
@@ -200,7 +200,6 @@ Current Lightning Users
    clouds/cluster
    Save and load model progress <common/checkpointing>
    Save memory with half-precision <common/precision>
-   Training over the internet <strategies/hivemind>
    advanced/model_parallel
    clouds/cloud_training
    Train on single or multiple GPUs <accelerators/gpu>
@@ -246,7 +245,6 @@ Current Lightning Users
    Metrics <https://torchmetrics.readthedocs.io/en/stable/>
    Model <model/build_model.rst>
    Model Parallel <advanced/model_parallel>
-   Collaborative Training <strategies/hivemind>
    Plugins <extensions/plugins>
    Progress bar <common/progress_bar>
    Production <deploy/production_advanced>

diff --git a/docs/source-pytorch/strategies/hivemind.rst b/docs/source-pytorch/strategies/hivemind.rst
diff --git a/docs/source-pytorch/strategies/hivemind_basic.rst b/docs/source-pytorch/strategies/hivemind_basic.rst
diff --git a/docs/source-pytorch/strategies/hivemind_expert.rst b/docs/source-pytorch/strategies/hivemind_expert.rst
diff --git a/docs/source-pytorch/strategies/hivemind_intermediate.rst b/docs/source-pytorch/strategies/hivemind_intermediate.rst
diff --git a/requirements/pytorch/strategies.txt b/requirements/pytorch/strategies.txt
@@ -4,4 +4,3 @@
 # colossalai>=0.1.10  # TODO: uncomment when there's a stable version released
 fairscale>=0.4.5, <0.4.13
 deepspeed>=0.6.0, <=0.7.0
-hivemind==1.1.5; sys_platform == 'linux'
diff --git a/src/pytorch_lightning/strategies/__init__.py b/src/pytorch_lightning/strategies/__init__.py
@@ -20,7 +20,6 @@
 from pytorch_lightning.strategies.dp import DataParallelStrategy  # noqa: F401
 from pytorch_lightning.strategies.fully_sharded import DDPFullyShardedStrategy  # noqa: F401
 from pytorch_lightning.strategies.fully_sharded_native import DDPFullyShardedNativeStrategy  # noqa: F401
-from pytorch_lightning.strategies.hivemind import HivemindStrategy  # noqa: F401
 from pytorch_lightning.strategies.hpu_parallel import HPUParallelStrategy  # noqa: F401
 from pytorch_lightning.strategies.ipu import IPUStrategy  # noqa: F401
 from pytorch_lightning.strategies.parallel import ParallelStrategy  # noqa: F401