[train], use adam optimizer, not muon for ci, because ci can not install muon related library.

Author: LiJunscs

requirements/cuda/train.txt

@@ -5,4 +5,3 @@ sentencepiece==0.2.1
 transformers==4.57.6
 tiktoken==0.12.0
 megatron_core @ https://resource.flagos.net/repository/flagos-pypi-hosted/packages/megatron-core/0.16.0rc0/megatron_core-0.16.0rc0-cp312-cp312-linux_x86_64.whl
-emerging_optimizers @ https://github.com/NVIDIA-NeMo/Emerging-Optimizers.git@v0.2.0

tests/functional_tests/train/deepseek/conf/train/tp1_pp2_ep2_v4.yaml

@@ -18,8 +18,7 @@ system:
   reset_attention_mask: true
   qk_layernorm: true
   sequence_parallel: true
-  use_distributed_optimizer: false
-  use_padded_layerwise_optimizer: false
+  use_distributed_optimizer: true
   num_distributed_optimizer_instances: 1
   overlap_grad_reduce: true
   overlap_param_gather: true
@@ -136,8 +135,6 @@ model:
   # pytorch_profiler_collect_callstack: true
 
   optimizer:
-    optimizer: dist_muon
-    muon_tp_mode: distributed
     clip_grad: 1.0
     weight_decay: 0.1
     adam_beta1: 0.9

tests/test_utils/runners/setup_training_test_env.sh

@@ -109,8 +109,7 @@ setup_cuda_training_env() {
         websockets==15.0.1 \
         msgpack==1.1.0 \
         datasets==4.5.0 \
-        fast-hadamard-transform==1.1.0 \
-        git+https://github.com/NVIDIA-NeMo/Emerging-Optimizers.git@v0.2.0
+        fast-hadamard-transform==1.1.0
 }
 
 setup_metax_training_env() {