GPU? ??? ??? ???? 3D ???? ????? ? ???? ???? ?? ?? ?? ??? ???????. ??? GPU? AI ??? ??? ?? ??? ?????.
?? ??? ??? ??? ???? ? ?? ????? ????, ???? ??? ????? ?????? ???? ??? ? ????.
??? AI ?? ???? ??? ???? ?? ???? GPU? ?????? ?? ? ????? ????. ?? ?? ??? ?? ???? ???? ???? ? ?? ?? ?? ???????.
NVIDIA DGX Cloud
NVIDIA DGX Cloud? AI ??? ???? ???? ?? ???? ???? ?? ??, ??, ???? ???? ?? ???? ? ?? ??? ?? ??? ???? ?????. ?? ? ?? ?????? ??? ? ?? ??? ??? ????? AI ??? ??? ???? ???.
??? ?????? AI ?????? ????? ? ??? AI ??????? ??? ? ??? ?????? ??? ???? ?? ???? ????? ???. NVIDIA? ??? ??? ?? ???? ?? ?? ??? ???, ??, ???, ??? ???? AI ???????? ???? ????.
Flyte ?????? NVIDIA DGX Cloud? ??? ? ?? Union? NVIDIA DGX Agent? ?????. GPU ???? ?? ????? ???? ??? ????? ?? ?? ???? AI ?????? ????? ?? ??? ?????.
Union? ?? ???? ???? Flyte?? ?? ?? ?? ???????? ???? AI ??????? ?????, ?????, ?? ???? ?? ?? ?? ??? ??? ? ? ?? ???? ??? ??? ?? ??????.
?????? ??? ??
Union? ML ????? ??? ??????? ???? ??? ???? ? ??? ? ??? ????? AI ??????? ? ?? ??? ?? ??? ????? ????? ? ??? ???.
?? 1? ??? ??? Repl.it? ?? ??? ??? ?? LLM? ?? ???? ?? ???? ??? ?????.
(??: ?? ?? ??(LLM)? ?? ???? ??)
?? 1? LLM? ?? ???? ????? ?? ??? ????, ?? ?? ???? GPU? ??? ??? ?? ??? ?????. ???? ??? ?????? ???? ??? ??? ? ?????
AI ?? ???? ?????? ??? ? ?? ??? ???? ???? ?? ??? ?????.
Union ?? ???? Lyft?? Linux Foundation ?? ?? ????? ??? Flyte? ??? ? ?? ?? ?????? ?? ? ??? ?? ?? ??? ???. ?? ???? ?? ? ?? ?? ??? ??? ????.
- ??: ??? ??? ?? ?? ?? ?????? ?? ??? ?????. ? ???? ?? ?? ? ??? ????. ???? ??, ??? ??, ?? ?? ?????.
- ?????: ??? ??? ?? ???? ?????? ???. ?????? ? ??? ??? ????? ?? ??? ???? ????.
- ??? ??? ??: ? ? ???? ???? ? ???? ?? ??? ?? ??? ????. Flyte?? “? ??? ???? ? ? ??? ??? ??? ?????.”?? ??? ??? ????.
- ????: ????? Flyte? DGX Cloud? ?? ?? ???? ???? Flyte? ?? ?? ??? ???? ?? ??? ? ?? ??????.
?? ?? ????? ??? ??? ?? ?? ???? ?? ????? ?? ??? ?????.
from collections import NamedTuple
import pandas as pd
import torch.nn as nn
from flytekit import task, workflow, Resources
from flytekit.extras.accelerators import T4
from flytekitplugins.spark import Databricks
@task(task_config=Databricks(...))
def create_data() -> pd.DataFrame:
...
@task(requests=Resources(gpu="4"), accelerator=T4)
def train_model(data: pd.DataFrame) -> nn.Module:
...
@workflow
def model_training_pipeline() -> nn.Module:
train_data, test_data = get_data()
model = train_model(data=train_data)
return model? ????? Databricks ?? ??? ???? ?? ???? ???? ??? Flyte? ?? ??? ?? ??? ???? 4?? NVIDIA T4 Tensor ?? GPU? ???? Databricks ?????? create_data ??? ?????.
Flyte?? Repl.it ?? ?? ?????? ??? ????.?
Union? NVIDIA DGX Agent ???
Mixtral 8x7b ??? ?? ????? ??? ?????. Flyte? ????? ?? GPU? ?????? ? ???, Amazon Web Service, Google Cloud Platform ? Microsoft Azure? ?? ?? ???? ??????? ?? GPU ??? ???? ???? ??? ??? ?? ? ????.
DGX Cloud? ???? ? ??? ????? ??? ???? ???.
- ?? ?? ???? ?? IDE?? ?? ?? ?????? ?????.
- ?? ? ???? ????????.
- ???? NVIDIA ???? ?????? ?????.
- ?????? ??? ??? ???? ?? ?? ??? ??????.
- 2??? ??? ??? ???? ? ??? ?????.
Union? ???? ??? ??? ??? ?? ???? ? ? ????. NVIDIA? ? ? ?? ??? DGX Cloud? ??? ??? ???? ??? ??? ??? ????? ??? ? ?? ??? ?????.
Mixtral 8x7B ??? ?? ???? fine_tune??? Flyte ??? ??? ??? ?????.
from datasets import load_dataset
from transformers import (
AutoModelForCausalLM,
BitsAndBytesConfig,
TrainingArguments,
)
from trl import SFTTrainer
@task
def fine_tune(dataset_name: str, output_dir: str):
# load model, tokenizer, and dataset
model = AutoModelForCausalLM.from_pretrained(
"mistralai/Mixtral-8x7B-v0.1",
quantization_config=BitsAndBytesConfig(load_in_4bit=True),
device_map="auto",
torch_dtype=torch.bfloat16,
)
tokenizer = AutoTokenizer.from_pretrained(
"mistralai/Mixtral-8x7B-v0.1",
use_fast=True
)
dataset = load_dataset(dataset_name, split="train")
# configer training arguments and trainer
training_args = TrainingArguments(...)
trainer = SFTTrainer(...)
# train and save the model
trainer.train()
trainer.save_model(output_dir)Flyte ? Union? NVIDIA DGX Agent? ????? ???? ????? ?? ???? dgx_image_spec ??? ??? DGXConfig ?? ?? ? ???? ?????.?
from flytekitpligins.dgx import DGXConfig, dgx_image_spec
fine_tuning_image_spec = dgx_image_spec.with_packages([
"accelerate", "datasets", "torch", "transformers", "trl",
])
@task(
task_config=DGXConfig(instance="dgxa100.80g.8.norm"),
container_image=fine_tuning_image_spec,
)
def fine_tune(dataset_name: str, output_dir: str):
... ?? ????? ??? ??? ? ?? 8?? NVIDIA A100 Tensor ?? GPU? ??? ?? ??? dgxa100.80g.8.norm ????? ?????. ?? ?? ??? ??, ???? ????? ????? ?? ??? ? ?? DGXTorchElastic ?? ???? ?????.?
?? Union? NVIDIA DGX Agent? ???? ?? blob ????(Amazon S3, GCS, Azure Blob Storage)? ???? DGX Cloud? blob ???? ????? ???? ?? ????? ?????. Flyte? ?? ??? ?? ??? ???? ??? ???? ??? ??? ? ????. ??? ??? Union AI ???? ?? Union ??? ?????.
?? ??? ?? ???? ??????!
AI? ????? GPU ???? ?? ??? ???? ?? ??? Union? ?? ???? ???? ?? ? ??? ???? ?? ????. ?? ????? ?????? ????? ??? ???? ?? ???? ???? ???? ? ?? ??? ??? ??? ? ??? ? ????.
AI ????? ???? ??? ????? Union ?? ???? ??? ??? ???? ??? ??????. ??? ??? ???? ??? AI ??? ?? ?????.
