Skip to content
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

ENH: auto detect device in pytorch model #322

Merged
merged 4 commits into from
Aug 8, 2023
Merged

ENH: auto detect device in pytorch model #322

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
7ddfac2
auto detect device in pytorch model
pangyoki Aug 7, 2023
73e0906
Merge branch 'main' of https://github.com/xorbitsai/inference into se…
pangyoki Aug 8, 2023
7050b45
fix comment
pangyoki Aug 8, 2023
ee81d31
fix
pangyoki Aug 8, 2023
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
8 changes: 8 additions & 0 deletions xinference/model/llm/core.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,6 +53,14 @@ def _is_darwin_and_apple_silicon():
def _is_linux():
return platform.system() == "Linux"

@staticmethod
def _is_darwin():
return platform.system() == "Darwin"

@staticmethod
def _is_arm():
return platform.processor() == "arm"

@abstractmethod
def load(self):
raise NotImplementedError
Expand Down
10 changes: 4 additions & 6 deletions xinference/model/llm/llm_family.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -160,13 +160,11 @@ def _is_linux():


def _has_cuda_device():
cuda_visible_devices = os.environ.get("CUDA_VISIBLE_DEVICES")
if cuda_visible_devices:
return True
else:
from xorbits._mars.resource import cuda_count
# `cuda_count` method already contains the logic for the
# number of GPUs specified by `CUDA_VISIBLE_DEVICES`.
from xorbits._mars.resource import cuda_count

return cuda_count() > 0
return cuda_count() > 0


def get_user_defined_llm_families():
Expand Down
83 changes: 45 additions & 38 deletions xinference/model/llm/pytorch/core.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@
EmbeddingUsage,
)
from ..core import LLM
from ..llm_family import LLMFamilyV1, LLMSpecV1
from ..llm_family import LLMFamilyV1, LLMSpecV1, _has_cuda_device
from ..utils import ChatModelMixin

logger = logging.getLogger(__name__)
Expand Down Expand Up @@ -73,6 +73,7 @@ def __init__(
self._pytorch_model_config: PytorchModelConfig = self._sanitize_model_config(
pytorch_model_config
)
self._device = self._pytorch_model_config["device"]

def _sanitize_model_config(
self, pytorch_model_config: Optional[PytorchModelConfig]
Expand All @@ -86,10 +87,10 @@ def _sanitize_model_config(
pytorch_model_config.setdefault("gptq_wbits", 16)
pytorch_model_config.setdefault("gptq_groupsize", -1)
pytorch_model_config.setdefault("gptq_act_order", False)
if self._is_darwin_and_apple_silicon():
pytorch_model_config.setdefault("device", "mps")
else:
pytorch_model_config.setdefault("device", "cuda")
pytorch_model_config.setdefault("device", "auto")
pytorch_model_config["device"] = self._select_device(
pytorch_model_config["device"]
)
return pytorch_model_config

def _sanitize_generate_config(
Expand All @@ -105,6 +106,25 @@ def _sanitize_generate_config(
pytorch_generate_config["model"] = self.model_uid
return pytorch_generate_config

def _select_device(self, device):
if device == "auto":
if self._is_darwin():
return "mps" if self._is_arm() else "cpu"
return "cuda" if _has_cuda_device() else "cpu"
elif device == "cuda":
if not _has_cuda_device():
raise ValueError("No cuda device is detected in your environment")
elif device == "mps":
if not self._is_darwin_and_apple_silicon():
raise ValueError(
"mps is only available for Mac computers with Apple silicon"
)
elif device == "cpu":
pass
else:
raise ValueError(f"Device {device} is not supported in temporary")
return device

def _load_model(self, kwargs: dict):
try:
from transformers import AutoModelForCausalLM, AutoTokenizer
Expand Down Expand Up @@ -142,23 +162,19 @@ def load(self):

quantization = self.quantization
num_gpus = self._pytorch_model_config.get("num_gpus", 1)
if self._is_darwin_and_apple_silicon():
device = self._pytorch_model_config.get("device", "mps")
else:
device = self._pytorch_model_config.get("device", "cuda")

if device == "cpu":
if self._device == "cpu":
kwargs = {"torch_dtype": torch.float32}
elif device == "cuda":
elif self._device == "cuda":
kwargs = {"torch_dtype": torch.float16}
elif device == "mps":
elif self._device == "mps":
kwargs = {"torch_dtype": torch.float16}
else:
raise ValueError(f"Device {device} is not supported in temporary")
raise ValueError(f"Device {self._device} is not supported in temporary")
kwargs["revision"] = self._pytorch_model_config.get("revision", "main")

if quantization != "none":
if device == "cuda" and self._is_linux():
if self._device == "cuda" and self._is_linux():
kwargs["device_map"] = "auto"
if quantization == "4-bit":
kwargs["load_in_4bit"] = True
Expand All @@ -185,7 +201,7 @@ def load(self):
else:
self._model, self._tokenizer = load_compress_model(
model_path=self.model_path,
device=device,
device=self._device,
torch_dtype=kwargs["torch_dtype"],
use_fast=self._use_fast_tokenizer,
revision=kwargs["revision"],
Expand All @@ -196,9 +212,9 @@ def load(self):
self._model, self._tokenizer = self._load_model(kwargs)

if (
device == "cuda" and num_gpus == 1 and quantization == "none"
) or device == "mps":
self._model.to(device)
self._device == "cuda" and num_gpus == 1 and quantization == "none"
) or self._device == "mps":
self._model.to(self._device)
logger.debug(f"Model Memory: {self._model.get_memory_footprint()}")

@classmethod
Expand Down Expand Up @@ -229,21 +245,21 @@ def generate(
)

def generator_wrapper(
prompt: str, device: str, generate_config: PytorchGenerateConfig
prompt: str, generate_config: PytorchGenerateConfig
) -> Iterator[CompletionChunk]:
if "falcon" in self.model_family.model_name:
for completion_chunk, _ in generate_stream_falcon(
self._model, self._tokenizer, prompt, device, generate_config
self._model, self._tokenizer, prompt, self._device, generate_config
):
yield completion_chunk
elif "chatglm" in self.model_family.model_name:
for completion_chunk, _ in generate_stream_chatglm(
self._model, self._tokenizer, prompt, device, generate_config
self._model, self._tokenizer, prompt, self._device, generate_config
):
yield completion_chunk
else:
for completion_chunk, _ in generate_stream(
self._model, self._tokenizer, prompt, device, generate_config
self._model, self._tokenizer, prompt, self._device, generate_config
):
yield completion_chunk

Expand All @@ -257,24 +273,20 @@ def generator_wrapper(
assert self._tokenizer is not None

stream = generate_config.get("stream", False)
if self._is_darwin_and_apple_silicon():
device = self._pytorch_model_config.get("device", "mps")
else:
device = self._pytorch_model_config.get("device", "cuda")
if not stream:
if "falcon" in self.model_family.model_name:
for completion_chunk, completion_usage in generate_stream_falcon(
self._model, self._tokenizer, prompt, device, generate_config
self._model, self._tokenizer, prompt, self._device, generate_config
):
pass
elif "chatglm" in self.model_family.model_name:
for completion_chunk, completion_usage in generate_stream_chatglm(
self._model, self._tokenizer, prompt, device, generate_config
self._model, self._tokenizer, prompt, self._device, generate_config
):
pass
else:
for completion_chunk, completion_usage in generate_stream(
self._model, self._tokenizer, prompt, device, generate_config
self._model, self._tokenizer, prompt, self._device, generate_config
):
pass
completion = Completion(
Expand All @@ -287,7 +299,7 @@ def generator_wrapper(
)
return completion
else:
return generator_wrapper(prompt, device, generate_config)
return generator_wrapper(prompt, generate_config)

def create_embedding(self, input: Union[str, List[str]]) -> Embedding:
try:
Expand All @@ -298,11 +310,6 @@ def create_embedding(self, input: Union[str, List[str]]) -> Embedding:
"Could not import torch. Please install it with `pip install torch`."
) from e

if self._is_darwin_and_apple_silicon():
device = self._pytorch_model_config.get("device", "mps")
else:
device = self._pytorch_model_config.get("device", "cuda")

if isinstance(input, str):
inputs = [input]
else:
Expand All @@ -315,8 +322,8 @@ def create_embedding(self, input: Union[str, List[str]]) -> Embedding:
encoding = tokenizer.batch_encode_plus(
inputs, padding=True, return_tensors="pt"
)
input_ids = encoding["input_ids"].to(device)
attention_mask = encoding["attention_mask"].to(device)
input_ids = encoding["input_ids"].to(self._device)
attention_mask = encoding["attention_mask"].to(self._device)
model_output = self._model(
input_ids, attention_mask, output_hidden_states=True
)
Expand Down Expand Up @@ -349,7 +356,7 @@ def create_embedding(self, input: Union[str, List[str]]) -> Embedding:
embedding = []
token_num = 0
for index, text in enumerate(inputs):
input_ids = tokenizer.encode(text, return_tensors="pt").to(device)
input_ids = tokenizer.encode(text, return_tensors="pt").to(self._device)
model_output = self._model(input_ids, output_hidden_states=True)
if is_chatglm:
data = (model_output.hidden_states[-1].transpose(0, 1))[0]
Expand Down