Update docs around mixing hf scheduler with deepspeed optimizer (#28223)

update docs around mixing hf scheduler with deepspeed optimizer

docs/source/en/main_classes/deepspeed.md

@@ -1221,12 +1221,7 @@ Therefore you have two ways to take advantage of this very beneficial feature:
 ### Optimizer and Scheduler
 
 As long as you don't enable `offload_optimizer` you can mix and match DeepSpeed and HuggingFace schedulers and
-optimizers, with the exception of using the combination of HuggingFace scheduler and DeepSpeed optimizer:
-
-| Combos       | HF Scheduler | DS Scheduler |
-|:-------------|:-------------|:-------------|
-| HF Optimizer | Yes          | Yes          |
-| DS Optimizer | No           | Yes          |
+optimizers.
 
 It is possible to use a non-DeepSpeed optimizer when `offload_optimizer` is enabled, as long as it has both CPU and
 GPU implementation (except LAMB).

src/transformers/integrations/deepspeed.py

@@ -275,14 +275,7 @@ def deepspeed_optim_sched(trainer, hf_deepspeed_config, args, num_training_steps
 
     config = hf_deepspeed_config.config
 
-    # Optimizer + Scheduler
-    # Currently supported combos:
-    # 1. DS scheduler + DS optimizer: Yes
-    # 2. HF scheduler + HF optimizer: Yes
-    # 3. DS scheduler + HF optimizer: Yes
-    # 4. HF scheduler + DS optimizer: No
-    #
-    # Unless Offload is enabled in which case it's:
+    # Mixing and matching DS schedulers and optimizers is supported unless Offload is enabled in which case it's:
     # 1. DS scheduler + DS optimizer: Yes
     # 2. HF scheduler + HF optimizer: Mostly*
     # 3. DS scheduler + HF optimizer: Mostly*