Update base for Update on "[cp] apply fsdp to model when CP is enable…

…d without DP for correct loss and lower mem usage"


**Summary**
Previously CP forgot to shard the model via `apply_fsdp` when DP is not combined with CP. This leads to high peak memory usage and diverging loss.

**Test**
1. modify `train_configs/llama3_8b.toml`
```
steps = 20
context_parallel_degree = 8
```
2.  run training on 8xH100 GPUs
`CONFIG_FILE="./train_configs/llama3_8b.toml" NGPU=8 LOG_RANK=0,1,2,3,4,5,6,7 ./run_llama_train.sh`
Before: CUDA OutOfMemory 
After: successful 20-steps training


[ghstack-poisoned]

torchtitan/parallelisms/parallel_dims.py

@@ -83,7 +83,7 @@ def build_mesh(self, device_type):
             if self.dp_replicate > 1 and self.dp_shard > 1:  # HSDP
                 mesh["dp_replicate", "dp_shard", "cp"]._flatten(mesh_dim_name="dp_cp")
             elif self.dp_shard > 1:  # FSDP
-                mesh["dp", "cp"]._flatten(mesh_dim_name="dp_cp")
+                mesh["dp_shard", "cp"]._flatten(mesh_dim_name="dp_cp")
 
         return mesh