Files

54 lines
2.1 KiB
Python

# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""DeepseekV4 rotary embedding initialization."""
from vllm.model_executor.layers.rotary_embedding import get_rope
from vllm.model_executor.layers.rotary_embedding.base import RotaryEmbedding
def build_deepseek_v4_rope(
config,
*,
head_dim: int,
rope_head_dim: int,
max_position_embeddings: int,
compress_ratio: int,
) -> RotaryEmbedding:
rope_parameters = config.rope_parameters
# Newer checkpoints nest per-layer-type rope dicts ({"main", "compress"});
# older ones ship a single flat dict shared by all layer types.
if isinstance(rope_parameters.get("main"), dict) and isinstance(
rope_parameters.get("compress"), dict
):
key = "compress" if compress_ratio > 0 else "main"
rope_parameters = dict(rope_parameters[key])
else:
rope_parameters = dict(rope_parameters)
rope_parameters["rope_theta"] = (
config.compress_rope_theta if compress_ratio > 0 else config.rope_theta
)
if compress_ratio > 0 and rope_parameters["rope_type"] != "default":
# YaRN applies to all compressed-KV layers (compress_ratio > 0); only
# pure sliding-window layers (compress_ratio == 0) use plain RoPE.
rope_parameters["rope_type"] = (
"deepseek_llama_scaling"
if rope_parameters.get("attention_factor") == 1.0
else "deepseek_yarn"
)
else:
# Sliding-window layers use plain RoPE (theta=rope_theta, no YaRN).
rope_parameters["rope_type"] = "deepseek_yarn"
rope_parameters["factor"] = 1.0
rope_parameters["original_max_position_embeddings"] = max_position_embeddings
rope_parameters["mscale"] = 0 # Disable mscale
rope_parameters["mscale_all_dim"] = 0 # Disable mscale
rope_parameters["is_deepseek_v4"] = True
rope_parameters["rope_dim"] = rope_head_dim
return get_rope(
head_dim,
max_position=max_position_embeddings,
rope_parameters=rope_parameters,
is_neox_style=False,
)