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 ddp example spec #554

Merged
merged 2 commits into from
Nov 27, 2023
Merged

update ddp example spec #554

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7fde8fd
update ddp example spec
akhileshh Nov 27, 2023
31633a1
Merge branch 'main' into akhilesh/update-ddp-example-spec
akhileshh Nov 27, 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
30 changes: 25 additions & 5 deletions specs/examples/training/ddp.cue
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
#EXP_NAME: "encoding_coarsener"
#EXP_VERSION: "remote_ddp_x52"
#EXP_VERSION: "single_spec_ddp_x0"

#TRAINING_ROOT: "gs://tmp_2w/examples/training_artifacts"

Expand All @@ -8,15 +8,35 @@
#ENCODER_CKPT: null
#DECODER_CKPT: null


#ENC_CV: "gs://fafb_v15_aligned/v0/img/img"

"@type": "lightning_train"

///////////////////////////////////////////////////////////////////
////////////////////////// DDP Config /////////////////////////////
///////////////////////////////////////////////////////////////////
local_run: false
cluster_name: "zutils-x3"
cluster_region: "us-east1"
cluster_project: "zetta-research"
image: "us.gcr.io/zetta-research/zetta_utils:single_spec_ddp_x0"
resource_requests: {
memory: "8192Mi"
"nvidia.com/gpu": "1"
}
resource_limits: {
memory: "8192Mi"
"nvidia.com/gpu": "1"
}
env_vars: {
"LOGLEVEL": "INFO"
"NCCL_SOCKET_IFNAME": "eth0"
}

///////////////////////////////////////////////////////////////////
//////////////////////// Training Spec ////////////////////////////
///////////////////////////////////////////////////////////////////

"@type": "lightning_train"
regime: {
"@type": "EncodingCoarsenerRegime"
lr: 4e-4
Expand All @@ -42,7 +62,7 @@ trainer: {
"@type": "ZettaDefaultTrainer"
accelerator: "cuda"
devices: 1
num_nodes: 32
num_nodes: 1
max_epochs: 100
default_root_dir: #TRAINING_ROOT
experiment_name: #EXP_NAME
Expand Down Expand Up @@ -121,7 +141,7 @@ trainer: {
{
"@type": "to_float32"
"@mode": "partial"
}
},
]
}
}
Expand Down
17 changes: 0 additions & 17 deletions specs/examples/training/ddp_config.cue
View file
Open in desktop

This file was deleted.

5 changes: 2 additions & 3 deletions zetta_utils/training/lightning/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -90,9 +90,8 @@ def lightning_train(
return

assert image is not None, "Must provide a container image for remote training."
assert (
resource_requests or resource_limits
), "Must provide at least one of resource requests or limits for remote training."
assert resource_limits is not None, "Must provide resource limits for remote training."

execution_id = mazepa.id_generation.get_unique_id(
prefix="exec", slug_len=4, add_uuid=False, max_len=50
)
Expand Down
Loading