[feat] add microbatch forwarding (#6251)

* add microbatch forwarding

* fix forward microbatch

* fix producer OOM

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

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

* change project name

* fix temperature annealing

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

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

* address conversation

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
This commit is contained in:
YeAnbang
2025-03-28 10:24:58 +08:00
committed by GitHub
parent 489f215ad9
commit 50153005b4
5 changed files with 112 additions and 72 deletions

View File

@@ -66,12 +66,7 @@ class BaseConsumer:
cc.init_collective_group(self.num_producers + 1, self.num_producers, group_name="sync_model")
launch(self.rank, self.world_size, self.master_addr, self.master_port, local_rank=0)
plugin_config = dict(
tp_size=1,
pp_size=1,
precision="bf16",
zero_stage=1,
)
plugin_config = dict(tp_size=1, pp_size=1, precision="bf16", zero_stage=2)
if self.plugin_config.get("pp_size", 1) > 1 and "num_microbatches" not in self.plugin_config:
plugin_config["microbatch_size"] = self.microbatch_size
plugin_config.update(self.plugin_config)