Skip to content

[Model] NemotronH Support #22349

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Aug 11, 2025
Merged

[Model] NemotronH Support #22349

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 21 additions & 5 deletions vllm/model_executor/models/nemotron_h.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,20 +64,32 @@ class NemotronHMLP(nn.Module):
def __init__(
self,
config: NemotronHConfig,
layer_idx: int,
quant_config: Optional[QuantizationConfig] = None,
bias: bool = False,
prefix: str = "",
) -> None:
super().__init__()

hybrid_override_pattern = config.hybrid_override_pattern
mlp_index = hybrid_override_pattern[:layer_idx + 1].count("-") - 1
Copy link
Contributor

@honghanhh honghanhh Aug 6, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As layer_idx can be None, layer_idx + 1 will fail if it's None, adding a guard for layer_idx and ensure hybrid_override_pattern exists before using them would help.

Copy link
Contributor Author

@danielafrimi danielafrimi Aug 6, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thanks!
change it that now layer_idx cant be None.
for the config.hybrid_override_pattern we do assume that all nemtoronH will have this configuration.

if isinstance(config.intermediate_size, list):
if len(config.intermediate_size) == 1:
intermediate_size = config.intermediate_size[0]
else:
intermediate_size = config.intermediate_size[mlp_index]
else:
intermediate_size = config.intermediate_size

self.up_proj = ColumnParallelLinear(
input_size=config.hidden_size,
output_size=config.intermediate_size,
output_size=intermediate_size,
bias=bias,
quant_config=quant_config,
prefix=f"{prefix}.up_proj",
)
self.down_proj = RowParallelLinear(
input_size=config.intermediate_size,
input_size=intermediate_size,
output_size=config.hidden_size,
bias=bias,
quant_config=quant_config,
Expand Down Expand Up @@ -110,6 +122,7 @@ def __init__(
quant_config=quant_config,
bias=config.mlp_bias,
prefix=f"{prefix}.mixer",
layer_idx=layer_idx,
)

self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
Expand Down Expand Up @@ -146,7 +159,7 @@ def __init__(
hidden_size=config.hidden_size,
ssm_state_size=config.ssm_state_size,
conv_kernel_size=config.conv_kernel,
intermediate_size=config.expand * config.hidden_size,
intermediate_size=config.mamba_num_heads * config.mamba_head_dim,
use_conv_bias=config.use_conv_bias,
use_bias=config.use_bias,
n_groups=config.n_groups,
Expand Down Expand Up @@ -205,7 +218,10 @@ def __init__(
# the KV heads across multiple tensor parallel GPUs.
assert tp_size % self.total_num_kv_heads == 0
self.num_kv_heads = max(1, self.total_num_kv_heads // tp_size)
self.head_dim = config.hidden_size // self.total_num_heads
if hasattr(config, "head_dim") and config.head_dim is not None:
self.head_dim = config.head_dim
else:
self.head_dim = config.hidden_size // self.total_num_heads
self.q_size = self.num_heads * self.head_dim
self.kv_size = self.num_kv_heads * self.head_dim
self.scaling = self.head_dim**-0.5
Expand Down Expand Up @@ -481,7 +497,7 @@ def get_mamba_state_shape_from_config(
"""
parallel_config = vllm_config.parallel_config
hf_config = vllm_config.model_config.hf_config
intermediate_size = hf_config.expand * hf_config.hidden_size
intermediate_size = hf_config.mamba_num_heads * hf_config.mamba_head_dim

return MambaStateShapeCalculator.mamba2_state_shape(
intermediate_size=intermediate_size,
Expand Down
4 changes: 2 additions & 2 deletions vllm/transformers_utils/configs/nemotron_h.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,7 +151,7 @@ def __init__(
num_hidden_layers=52,
hybrid_override_pattern="M-M-M-M*-M-M-M-M-M*-M-M-M-M-M*-M-M-M-M-M*-M-M-M-M-M-",
num_attention_heads=32,
attention_head_dim=128,
head_dim=128,
num_key_value_heads=8, # nemo: num_query_groups
mlp_hidden_act="relu2",
attention_bias=False,
Expand Down Expand Up @@ -194,7 +194,7 @@ def __init__(
self.num_hidden_layers = num_hidden_layers
self.hybrid_override_pattern = hybrid_override_pattern
self.num_attention_heads = num_attention_heads
self.attention_head_dim = attention_head_dim
self.head_dim = head_dim
self.sliding_window = sliding_window
self.max_position_embeddings = max_position_embeddings
self.attention_dropout = attention_dropout
Expand Down