-
Notifications
You must be signed in to change notification settings - Fork 417
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.
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
Add bminf #238
Add bminf #238
Conversation
Signed-off-by: Anhforth <yanzaodong2021@163.com>
Signed-off-by: Anhforth <yanzaodong2021@163.com>
Signed-off-by: ftgreat <ftgreat@163.com>
Signed-off-by: ftgreat <ftgreat@163.com>
Signed-off-by: ftgreat <ftgreat@163.com>
Signed-off-by: ftgreat <ftgreat@163.com>
Signed-off-by: ftgreat <ftgreat@163.com>
Signed-off-by: ftgreat <ftgreat@163.com>
Signed-off-by: ftgreat <ftgreat@163.com>
Signed-off-by: ftgreat <ftgreat@163.com>
flagai/model/gpt2_model.py
Outdated
@@ -112,10 +112,15 @@ def __init__(self, config): | |||
self.drop = nn.Dropout(config.embd_pdrop) | |||
self.project_in = None | |||
self.project_out = None | |||
# self.h = bminf.TransformerBlockList([ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
rm
flagai/model/gpt2_model.py
Outdated
@@ -274,10 +279,14 @@ def __init__(self, config, **kwargs): | |||
self.parallel_output = True | |||
|
|||
self.transformer = GPT2Stack(config_gpt) | |||
# self.lm_head = bminf.QuantizedLinear(nn.Linear(config_gpt.n_embd, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
rm
setup.py
Outdated
@@ -5,7 +5,7 @@ | |||
|
|||
setup( | |||
name="flagai", | |||
version="v1.6.0", | |||
version="v1.6.2", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
revert
|
||
loader = AutoLoader(task_name="lm", | ||
model_name="galactica-6.7b-en", | ||
model_dir="/share/projset/baaishare/baai-mrnd/xingzhaohu/") |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
./checkpoints
@@ -12,7 +11,7 @@ | |||
# device = torch.device("cpu") | |||
# single gpu | |||
trainer = Trainer( | |||
env_type="pytorchDDP", | |||
env_type="deepspeed+mpu", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
建议恢复。 如果想新增 mpu 方式,可以单独文件。
@@ -31,15 +30,15 @@ | |||
num_nodes=1, | |||
num_gpus=2, | |||
checkpoint_activations=False, | |||
model_parallel_size=1, | |||
model_parallel_size=2, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
=1
name: Pull Request
title: '[PR]'
assignees: 'BAAI-OpenPlatform,ftgreat'
Description
Please describe here what the PR does.
Checklist