Adding support for multi-node DDP training#708
Merged
quic-akuruvil merged 2 commits intoquic:mainfrom Jan 13, 2026
Merged
Conversation
smedhe
requested review from
ochougul,
quic-amitraj,
quic-hemagnih and
quic-rishinr
as code owners
Contributor
quic-akuruvil
left a comment
quic-akuruvil
left a comment
There was a problem hiding this comment.
Put the sample command for reference interface, along with description
Contributor
|
For PP, device_map.py uses get_rank() which was earlier returning local rank but now returns rank. This will not break PP as of now, but it's a wrong practice. In case PP is used with multi node, it would break. Please call the get_local_rank() in device_map once it is defined. |
quic-swatia
reviewed
Jan 8, 2026
Contributor
After this change, we can also test if PP + DDP for multi node is working with this PR. It seems like it might work. It's okay if it doesn't, we can still go ahead with the PR. |
quic-meetkuma
approved these changes
Jan 9, 2026
Contributor
quic-meetkuma
left a comment
quic-meetkuma
left a comment
There was a problem hiding this comment.
Looks fine to me. Please update the finetune.md file. Imagine or other teams refer the commands from there. They will need to change command at their side of the code.
Contributor
quic-akuruvil
left a comment
quic-akuruvil
left a comment
There was a problem hiding this comment.
Start CI run also for this. So CI passes before merge
quic-swatia
approved these changes
Jan 12, 2026
Contributor
quic-swatia
left a comment
quic-swatia
left a comment
There was a problem hiding this comment.
PR is good to merge with the latest changes . Please make sure these commands are added in the finetuning documentation. Thanks
Contributor
Added as part of PR 717 |
Signed-off-by: Sharvari Medhe <smedhe@qti.qualcomm.com>
Signed-off-by: Sharvari Medhe <smedhe@qti.qualcomm.com>
smedhe
force-pushed
the
multi_node_ddp
branch
from
fd66951 to
1ebfdb3
Compare
tchawada
pushed a commit
to tchawada/QEff_tanisha
that referenced
this pull request
Feb 4, 2026
Add support for multi-node Distributed Data Parallel (DDP) training to the QEfficient finetuning pipeline. This enables scaling training across multiple nodes while keeping the existing single-node behavior unchanged. Commands for DDP across 2 servers: For the Master Addr or the Primary Machine, use node-rank as 0: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=0 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 For Node 1, use node-rank as 1: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=1 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 --------- Signed-off-by: Sharvari Medhe <smedhe@qti.qualcomm.com>
tchawada
pushed a commit
to tchawada/QEff_tanisha
that referenced
this pull request
Feb 4, 2026
Add support for multi-node Distributed Data Parallel (DDP) training to the QEfficient finetuning pipeline. This enables scaling training across multiple nodes while keeping the existing single-node behavior unchanged. Commands for DDP across 2 servers: For the Master Addr or the Primary Machine, use node-rank as 0: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=0 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 For Node 1, use node-rank as 1: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=1 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 --------- Signed-off-by: Sharvari Medhe <smedhe@qti.qualcomm.com>
tchawada
pushed a commit
to tchawada/QEff_tanisha
that referenced
this pull request
Feb 4, 2026
Add support for multi-node Distributed Data Parallel (DDP) training to the QEfficient finetuning pipeline. This enables scaling training across multiple nodes while keeping the existing single-node behavior unchanged. Commands for DDP across 2 servers: For the Master Addr or the Primary Machine, use node-rank as 0: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=0 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 For Node 1, use node-rank as 1: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=1 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 --------- Signed-off-by: Sharvari Medhe <smedhe@qti.qualcomm.com>
tchawada
pushed a commit
to tchawada/QEff_tanisha
that referenced
this pull request
Feb 5, 2026
Add support for multi-node Distributed Data Parallel (DDP) training to the QEfficient finetuning pipeline. This enables scaling training across multiple nodes while keeping the existing single-node behavior unchanged. Commands for DDP across 2 servers: For the Master Addr or the Primary Machine, use node-rank as 0: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=0 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 For Node 1, use node-rank as 1: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=1 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 --------- Signed-off-by: Sharvari Medhe <smedhe@qti.qualcomm.com>
qcdipankar
pushed a commit
to qcdipankar/efficient-transformers
that referenced
this pull request
Feb 8, 2026
Add support for multi-node Distributed Data Parallel (DDP) training to the QEfficient finetuning pipeline. This enables scaling training across multiple nodes while keeping the existing single-node behavior unchanged. Commands for DDP across 2 servers: For the Master Addr or the Primary Machine, use node-rank as 0: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=0 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 For Node 1, use node-rank as 1: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=1 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 --------- Signed-off-by: Sharvari Medhe <smedhe@qti.qualcomm.com> Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
tchawada
pushed a commit
to tchawada/QEff_tanisha
that referenced
this pull request
Feb 16, 2026
Add support for multi-node Distributed Data Parallel (DDP) training to the QEfficient finetuning pipeline. This enables scaling training across multiple nodes while keeping the existing single-node behavior unchanged. Commands for DDP across 2 servers: For the Master Addr or the Primary Machine, use node-rank as 0: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=0 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 For Node 1, use node-rank as 1: QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=1 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50 --------- Signed-off-by: Sharvari Medhe <smedhe@qti.qualcomm.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Add support for multi-node Distributed Data Parallel (DDP) training to the QEfficient finetuning pipeline. This enables scaling training across multiple nodes while keeping the existing single-node behavior unchanged.
Commands for DDP across 2 servers:
For the Master Addr or the Primary Machine, use node-rank as 0:
QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=0 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50
For Node 1, use node-rank as 1:
QAIC_VISIBLE_DEVICES=0,1,2,3 torchrun --nnodes=2 --nproc-per-node=4 --seed 0 --node-rank=1 --master_addr=<MASTER_NODE_IP> --master_port=8000 -m QEfficient.cloud.finetune --device qaic --enable_ddp --model_name "meta-llama/Llama-3.2-1B" --dataset alpaca_dataset --train_batch_size 1 --val_batch_size 1 --num_epochs 1 --max_train_step 200 --max_eval_step 50