Fix PyTorch training benchmark doc template (#5357)

* fix template

* update wordlist

.wordlist.txt

@@ -622,6 +622,7 @@ coalescable
 codename
 collater
 comgr
+compat
 completers
 composable
 concretization
@@ -783,6 +784,7 @@ lossy
 macOS
 matchers
 maxtext
+megablocks
 megatron
 microarchitecture
 migraphx
@@ -941,6 +943,7 @@ softmax
 spack
 spmm
 src
+stanford
 stochastically
 strided
 subcommand

docs/how-to/rocm-for-ai/training/benchmark-docker/pytorch-training.rst

@@ -164,6 +164,8 @@ Run training
    {% for model_group in model_groups %}
       {% for model in model_group.models %}
 
+         .. container:: model-doc {{ model.mad_tag }}
+
             The following run command is tailored to {{ model.model }}.
             See :ref:`amd-pytorch-training-model-support` to switch to another available model.
 
@@ -176,8 +178,6 @@ Run training
                   cd MAD
                   pip install -r requirements.txt
 
-         .. container:: model-doc {{ model.mad_tag }}
-
             2. For example, use this command to run the performance benchmark test on the {{ model.model }} model
                using one node with the {{ model.precision }} data type on the host machine.
 
@@ -202,6 +202,8 @@ Run training
    {% for model_group in model_groups %}
       {% for model in model_group.models %}
 
+         .. container:: model-doc {{ model.mad_tag }}
+
             The following commands are tailored to {{ model.model }}.
             See :ref:`amd-pytorch-training-model-support` to switch to another available model.