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

Update finetune code to the latest version of Yi #222

Merged
merged 4 commits into from
Dec 2, 2023
Merged

Update finetune code to the latest version of Yi #222

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
1d28c16
adapt new version of Yi
jiangchengSilent Dec 1, 2023
acad836
adapt new version of Yi
jiangchengSilent Dec 1, 2023
47596e1
Merge branch 'main' into jiangcheng_dev
findmyway Dec 2, 2023
54ec43c
apply autoflake
findmyway Dec 2, 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
1 change: 1 addition & 0 deletions finetune/utils/model/model_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,7 @@ def create_hf_model(
from_tf=bool(".ckpt" in model_name_or_path),
config=model_config,
trust_remote_code=True,
use_flash_attention_2=True,
)
else:
model = model_class.from_pretrained(
Expand Down
35 changes: 0 additions & 35 deletions finetune/utils/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,5 @@
import json
import os
from shutil import copy

import deepspeed
import torch
Expand Down Expand Up @@ -90,12 +89,6 @@ def save_hf_format(model, tokenizer, args, sub_folder=""):
print(os.listdir(output_dir))
print(os.getcwd())

source = args.model_name_or_path
target = os.path.abspath(os.path.join(os.getcwd(), args.output_dir))
copy(os.path.join(source, "configuration_yi.py"), target)
copy(os.path.join(source, "modeling_yi.py"), target)
copy(os.path.join(source, "tokenization_yi.py"), target)


def get_all_reduce_mean(tensor):
torch.distributed.all_reduce(tensor, op=torch.distributed.ReduceOp.SUM)
Expand Down Expand Up @@ -258,31 +251,3 @@ def save_zero_three_model(model_ema, global_rank, save_dir, zero_stage=0):
if global_rank == 0:
torch.save(output_state_dict, output_model_file)
del output_state_dict


def save_zero_three_model(model_ema, global_rank, save_dir, zero_stage=0):
zero_stage_3 = zero_stage == 3
os.makedirs(save_dir, exist_ok=True)
WEIGHTS_NAME = "pytorch_model.bin"
output_model_file = os.path.join(save_dir, WEIGHTS_NAME)

model_to_save = model_ema.module if hasattr(model_ema, "module") else model_ema
if not zero_stage_3:
if global_rank == 0:
torch.save(model_to_save.state_dict(), output_model_file)
else:
output_state_dict = {}
for k, v in model_to_save.named_parameters():
if hasattr(v, "ds_id"):
with deepspeed.zero.GatheredParameters(
_z3_params_to_fetch([v]), enabled=zero_stage_3
):
v_p = v.data.cpu()
else:
v_p = v.cpu()
if global_rank == 0 and "lora" not in k:
output_state_dict[k] = v_p

if global_rank == 0:
torch.save(output_state_dict, output_model_file)
del output_state_dict