[plugin] hybrid support zero bubble pipeline (hpcaitech#6060)

* hybrid support zbv

* fix

fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* Update zero_bubble_pp.py

* fix

* fix-ci

* fix

[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* fix

* fix

* fix

* [zerobubble]Support ZeroBubble Pipeline (hpcaitech#6034)

* [feat] add zerobubble pp (just a frame now); add POC test for dx_dw; add test for zerobubble;

* [feat] add dw test;

* [fix] fix weight not close;

* [update] update text;

* [feat] add test run_fwd_bwd automatic scheduling;

* [feat] split communication and calculation; fix pop empty send_bwd_buffer error;

* [feat] add test for p & p grad;

* [feat] add comments for ZBV func;

* [fix] rm useless assign and comments;

* [fix] fix ci test; add pytest;

* [feat] add run_fwd_bwd_with_microbatch  (replace input) & test; add p&p.grad assert close test & all pass;

* [feat] add apply v_schedule graph; p & p.grad assert err exist;

* [fix] update

* [feat] fix ci; add assert;

* [feat] fix poc format

* [feat] fix func name & ci; add comments;

* [fix] fix poc test; add comments in poc;

* [feat] add optim backward_b_by_grad

* [feat] fix optimizer bwd b & w; support return accum loss & output

* [feat] add fwd_bwd_step, run_fwd_only;

* [fix] fix optim bwd; add license for v_schedule; remove redundant attributes; fix schedule loop "while"--> "for"; add communication dict;

* [fix] fix communication_map;

* [feat] update test; rm comments;

* [fix] rm zbv in hybridplugin

* [fix] fix optim bwd;

* [fix] fix optim bwd;

* [fix] rm output.data after send fwd;

* [fix] fix bwd step if condition; remove useless comments and format info;

* [fix] fix detach output & release output;

* [fix] rm requir_grad for output;

* [fix] fix requir grad position and detach position and input&output local buffer append position;

* [feat] add memory assertation;

* [fix] fix mem check;

* [fix] mem assertation'

* [fix] fix mem assertation

* [fix] fix mem; use a new model shape; only assert mem less and equal than theo;

* [fix] fix model zoo import;

* [fix] fix redundant detach & clone; add buffer assertation in the end;

* [fix] add output_obj_grad assert None at bwd b step; replace input_obj.require_grad_ with treemap;

* [fix] update optim state dict assert (include param group & state); fix mem assert after add optim;

* [fix] add testcase with microbatch 4;

* hybrid support zbv

* fix

fix

* fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update zero_bubble_pp.py

* fix

* fix-ci

* fix

[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

fix

* fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix

* fix

* fix

* fix

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: duanjunwen <[email protected]>

colossalai/booster/plugin/hybrid_parallel_plugin.py

@@ -1166,6 +1166,14 @@ def __init__(
  num_microbatch=num_microbatches,
  microbatch_size=microbatch_size,
  )
+ elif pp_style == "zbv":
+ self.scheduler = ZeroBubbleVPipeScheduler(
+ stage_manager=self.stage_manager,
+ schedule=scheduler_nodes,
+ num_model_chunks=num_model_chunks,
+ num_microbatch=num_microbatches,
+ microbatch_size=microbatch_size,
+ )
  else:
  raise NotImplementedError()
  if sequence_parallelism_mode == "ring_attn":
@@ -1279,7 +1287,6 @@ def configure(
 
  # Replace with distributed implementation if exists
  optimizer = cast_to_distributed(optimizer)
-
  if isinstance(optimizer, DistGaloreAwamW) and zero_stage > 0 and self.dp_size > 0:
  self.logger.warning(
  "Galore is only supported for Tensor Parallel and vanilla Data Parallel yet. Disabling ZeRO.",

colossalai/zero/low_level/low_level_optim.py

@@ -433,7 +433,6 @@ def backward(self, loss, inputs=None, retain_graph=False):
 
  ctx = nullcontext() if self._backward_context is None else self._backward_context()
  with ctx:
- loss.backward(retain_graph=retain_graph)
  loss.backward(inputs=inputs, retain_graph=retain_graph)
 
  if not self.require_grad_sync: