Skip to content

Commit

Permalink
Merge branch 'recipe_for_25.01' of github.com:guyueh1/NeMo into recip…
Browse files Browse the repository at this point in the history
…e_for_25.01
  • Loading branch information
guyueh1 committed Jan 9, 2025
2 parents 2b8114b + 5ca96db commit 93cb713
Showing 3 changed files with 3 additions and 7 deletions.
4 changes: 1 addition & 3 deletions nemo/collections/llm/recipes/nemotron4_15b.py
Original file line number Diff line number Diff line change
@@ -25,9 +25,7 @@
from nemo.collections.llm.recipes.log.default import default_log, default_resume, tensorboard_logger
from nemo.collections.llm.recipes.nemotron import nemotron_model, nemotron_trainer
from nemo.collections.llm.recipes.optim.adam import distributed_fused_adam_with_cosine_annealing
from nemo.collections.llm.recipes.tp_overlap_configs.userbuffers import (
userbuffers_bf16_h100_h8192_tp2_mbs1_seqlen8192
)
from nemo.collections.llm.recipes.tp_overlap_configs.userbuffers import userbuffers_bf16_h100_h8192_tp2_mbs1_seqlen8192
from nemo.lightning.pytorch.callbacks.megatron_comm_overlap import MegatronCommOverlapCallback
from nemo.utils.exp_manager import TimingCallback

2 changes: 1 addition & 1 deletion nemo/collections/llm/recipes/nemotron4_340b.py
Original file line number Diff line number Diff line change
@@ -26,7 +26,7 @@
from nemo.collections.llm.recipes.nemotron import nemotron_model, nemotron_trainer
from nemo.collections.llm.recipes.optim.adam import distributed_fused_adam_with_cosine_annealing
from nemo.collections.llm.recipes.tp_overlap_configs.userbuffers import (
userbuffers_bf16_h100_h18432_tp8_mbs1_seqlen4096
userbuffers_bf16_h100_h18432_tp8_mbs1_seqlen4096,
)
from nemo.lightning.pytorch.callbacks.megatron_comm_overlap import MegatronCommOverlapCallback
from nemo.utils.exp_manager import TimingCallback
4 changes: 1 addition & 3 deletions nemo/lightning/run/plugins.py
Original file line number Diff line number Diff line change
@@ -19,6 +19,7 @@
from typing import Callable, Optional

import nemo_run as run
import torch
import yaml
from lightning.pytorch import Callback
from lightning.pytorch.loggers import WandbLogger
@@ -27,11 +28,8 @@
from nemo.lightning.pytorch.callbacks import NsysCallback, PreemptionCallback
from nemo.lightning.pytorch.strategies.megatron_strategy import MegatronStrategy
from nemo.utils import logging

from nemo.utils.import_utils import safe_import

import torch

res_module, HAVE_RES = safe_import('nvidia_resiliency_ext.ptl_resiliency')

# This file contains plugins based on NeMo-Run's run.Plugin API.

0 comments on commit 93cb713

Please sign in to comment.