biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Models] Add remaining model PP support (#7168)

Browse Source 
Signed-off-by: Muralidhar Andoorveedu <muralidhar.andoorveedu@centml.ai>
Signed-off-by: Murali Andoorveedu <muralidhar.andoorveedu@centml.ai>
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Murali Andoorveedu
2024-10-03 19:56:58 -07:00
committed by GitHub
parent 303d44790a
commit 0f6d7a9a34
69 changed files with 2585 additions and 1344 deletions
Download Patch File Download Diff File Expand all files Collapse all files

72
vllm/model_executor/models/phimoe.py
View File

@@ -21,7 +21,7 @@
# See the License for the specific language governing permissions and
# limitations under the License.
"""Inference-only PhiMoE model."""
from typing import Iterable, List, Optional, Tuple
from typing import Iterable, List, Optional, Tuple, Union
import torch
from torch import nn
@@ -29,7 +29,7 @@ from transformers.configuration_utils import PretrainedConfig
from vllm.attention import Attention, AttentionMetadata
from vllm.config import CacheConfig, LoRAConfig
from vllm.distributed import get_tensor_model_parallel_world_size
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
from vllm.model_executor.layers.fused_moe import FusedMoE
from vllm.model_executor.layers.linear import (QKVParallelLinear,
ReplicatedLinear,
@@ -46,7 +46,9 @@ from vllm.model_executor.model_loader.weight_utils import (
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.sequence import IntermediateTensors
from .interfaces import SupportsLoRA
from .interfaces import SupportsLoRA, SupportsPP
from .utils import (is_pp_missing_parameter,
make_empty_intermediate_tensors_factory, make_layers)
class PhiMoEConfig(PretrainedConfig):
@@ -435,6 +437,7 @@ class PhiMoEModel(nn.Module):
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
lora_config: Optional[LoRAConfig] = None,
prefix: str = "",
) -> None:
super().__init__()
self.padding_idx = config.pad_token_id
@@ -448,33 +451,56 @@ class PhiMoEModel(nn.Module):
config.hidden_size,
org_num_embeddings=config.vocab_size,
)
self.layers = nn.ModuleList([
PhiMoEDecoderLayer(config, cache_config, quant_config=quant_config)
for _ in range(config.num_hidden_layers)
])
self.start_layer, self.end_layer, self.layers = make_layers(
config.num_hidden_layers,
lambda prefix: PhiMoEDecoderLayer(config, cache_config,
quant_config),
prefix=f"{prefix}.layers")
self.norm = nn.LayerNorm(config.hidden_size,
eps=config.rms_norm_eps,
elementwise_affine=True)
self.make_empty_intermediate_tensors = (
make_empty_intermediate_tensors_factory(
["hidden_states", "residual"], config.hidden_size))
def forward(
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
kv_caches: List[torch.Tensor],
attn_metadata: AttentionMetadata,
) -> torch.Tensor:
hidden_states = self.embed_tokens(input_ids)
residual = None
for i in range(len(self.layers)):
intermediate_tensors: Optional[IntermediateTensors],
) -> Union[torch.Tensor, IntermediateTensors]:
if get_pp_group().is_first_rank:
hidden_states = self.embed_tokens(input_ids)
residual = None
else:
assert intermediate_tensors is not None
hidden_states = intermediate_tensors["hidden_states"]
residual = intermediate_tensors["residual"]
for i in range(self.start_layer, self.end_layer):
layer = self.layers[i]
hidden_states, residual = layer(positions, hidden_states,
kv_caches[i], attn_metadata,
residual)
hidden_states, residual = layer(
positions,
hidden_states,
kv_caches[i - self.start_layer],
attn_metadata,
residual,
)
if not get_pp_group().is_last_rank:
return IntermediateTensors({
"hidden_states": hidden_states,
"residual": residual
})
hidden_states = self.norm(hidden_states)
return hidden_states
class PhiMoEForCausalLM(nn.Module, SupportsLoRA):
class PhiMoEForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
fall_back_to_pt_during_load = False
packed_modules_mapping = {
@@ -537,6 +563,9 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA):
config.vocab_size)
self.sampler = Sampler()
self.make_empty_intermediate_tensors = (
self.model.make_empty_intermediate_tensors)
def forward(
self,
input_ids: torch.Tensor,
@@ -544,9 +573,9 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA):
kv_caches: List[torch.Tensor],
attn_metadata: AttentionMetadata,
intermediate_tensors: Optional[IntermediateTensors] = None,
) -> torch.Tensor:
) -> Union[torch.Tensor, IntermediateTensors]:
hidden_states = self.model(input_ids, positions, kv_caches,
attn_metadata)
attn_metadata, intermediate_tensors)
return hidden_states
def compute_logits(self, hidden_states: torch.Tensor,
@@ -589,6 +618,9 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA):
# Skip loading extra bias for GPTQ models.
if name.endswith(".bias") and name not in params_dict:
continue
# Skip layers on other devices.
if is_pp_missing_parameter(name, self):
continue
param = params_dict[name]
weight_loader = param.weight_loader
weight_loader(param, loaded_weight, shard_id)
@@ -599,6 +631,9 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA):
if weight_name not in name:
continue
name = name.replace(weight_name, param_name)
# Skip layers on other devices.
if is_pp_missing_parameter(name, self):
continue
param = params_dict[name]
weight_loader = param.weight_loader
weight_loader(
@@ -613,6 +648,9 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA):
# Skip loading extra bias for GPTQ models.
if name.endswith(".bias") and name not in params_dict:
continue
# Skip layers on other devices.
if is_pp_missing_parameter(name, self):
continue
# Remapping the name of FP8 kv-scale.
name = maybe_remap_kv_scale_name(name, params_dict)
if name is None: