New mcore transformer block spec (#8925)

* update package info (#8793)

Signed-off-by: eharper <[email protected]>

* update mcore (#8917)

Signed-off-by: Jan Baczek <[email protected]>

* Use new mcore transformer block config handling

Signed-off-by: Jan Baczek <[email protected]>

* API fixes

Signed-off-by: Jan Baczek <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Revert chages to CI and Dockerfile

Signed-off-by: Jan Baczek <[email protected]>

---------

Signed-off-by: eharper <[email protected]>
Signed-off-by: Jan Baczek <[email protected]>
Co-authored-by: Eric Harper <[email protected]>
Co-authored-by: Pablo Garay <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

nemo/collections/nlp/models/language_modeling/megatron/gpt_full_te_layer_autocast_spec.py

@@ -35,7 +35,9 @@
 
 try:
     from megatron.core import parallel_state, tensor_parallel
+    from megatron.core.fusions.fused_layer_norm import FusedLayerNorm
     from megatron.core.transformer.spec_utils import ModuleSpec
+    from megatron.core.transformer.transformer_block import TransformerBlockSubmodules, get_num_layers_to_build
     from megatron.core.transformer.transformer_layer import BaseTransformerLayer
     from megatron.core.transformer.utils import make_sharded_tensors_for_checkpoint
 
@@ -322,8 +324,10 @@ def sharded_state_dict(self, prefix: str = '', sharded_offsets: tuple = (), meta
 
 
 # Use this spec to use the full Transformer layer from Transformer Engine
-def get_gpt_full_te_layer_autocast_spec() -> ModuleSpec:
+def get_gpt_full_te_layer_autocast_spec(transformer_config) -> ModuleSpec:
     if not HAVE_MEGATRON_CORE or not HAVE_TE:
         raise ImportError(IMPORT_ERROR)
-
-    return ModuleSpec(module=TETransformerLayerAutocast)
+    num_layers = get_num_layers_to_build(transformer_config)
+    return TransformerBlockSubmodules(
+        layer_specs=[ModuleSpec(module=TETransformerLayerAutocast)] * num_layers, layer_norm=FusedLayerNorm
+    )

nemo/collections/nlp/models/language_modeling/megatron_gpt_model.py

@@ -135,7 +135,11 @@ def mcore_supports_moe() -> bool:
         return False
 
 
-def get_specs(spec_name, num_experts=None, moe_grouped_gemm=False, use_te=True):
+def get_specs(spec_name, transformer_config=None, use_te=True):
+    # else cases for backwards compatibility with neva
+    num_experts = transformer_config.num_moe_experts if transformer_config else None
+    moe_grouped_gemm = transformer_config.moe_grouped_gemm if transformer_config else False
+
     if num_experts is not None:
         assert mcore_supports_moe(), "Megatron-core >= v0.5.0 is required for MoE"
 
@@ -145,7 +149,7 @@ def get_specs(spec_name, num_experts=None, moe_grouped_gemm=False, use_te=True):
         "": get_gpt_layer_local_spec(num_experts, moe_grouped_gemm),
         "te_gpt": get_gpt_layer_with_transformer_engine_spec(num_experts, moe_grouped_gemm),
         "megatron_falcon_gpt": get_falcon_layer_spec(),
-        "megatron_gpt_full_te_layer_autocast": get_gpt_full_te_layer_autocast_spec(),
+        "megatron_gpt_full_te_layer_autocast": get_gpt_full_te_layer_autocast_spec(transformer_config),
         "ammo": get_gpt_layer_ammo_spec(),
     }
     if spec_name not in name_spec_dict:
@@ -391,12 +395,7 @@ def model_provider_func(self, pre_process, post_process):
         if self.mcore_gpt:
             model = MCoreGPTModel(
                 config=self.transformer_config,
-                transformer_layer_spec=get_specs(
-                    self.spec_name,
-                    self.transformer_config.num_moe_experts,
-                    self.transformer_config.moe_grouped_gemm,
-                    self.transformer_engine,
-                ),
+                transformer_layer_spec=get_specs(self.spec_name, self.transformer_config, self.transformer_engine,),
                 vocab_size=self.cfg.get('override_vocab_size', self.padded_vocab_size),
                 max_sequence_length=self.cfg.get('encoder_seq_length', 512),
                 pre_process=pre_process,