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

[tnx] add mixtral test to neuron ci #1837

Merged
merged 1 commit into from
Apr 29, 2024
Merged

[tnx] add mixtral test to neuron ci #1837

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
22 changes: 11 additions & 11 deletions .github/workflows/llm_inf2_integration.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -262,17 +262,6 @@ jobs:
python3 llm/client.py neuron-stable-diffusion stable-diffusion-xl-neuron
docker rm -f $(docker ps -aq)
sudo rm -rf models
- name: Test load split model with llama2 7B in handler
working-directory: tests/integration
run: |
rm -rf models
python3 llm/prepare.py transformers_neuronx llama-7b-split
./launch_container.sh deepjavalibrary/djl-serving:$DJLSERVING_DOCKER_TAG $PWD/models pytorch-inf2-2 \
serve
curl http://127.0.0.1:8080/models
python3 llm/client.py transformers_neuronx llama-7b-split
docker rm -f $(docker ps -aq)
sudo rm -rf models
- name: Test mistral 7B with handler
working-directory: tests/integration
run: |
Expand Down Expand Up @@ -337,6 +326,17 @@ jobs:
python3 llm/client.py transformers_neuronx_rolling_batch llama-7b-rb
docker rm -f $(docker ps -aq)
sudo rm -rf models
- name: Test transformers-neuronx mixtral-8x-7b rolling batch
working-directory: tests/integration
run: |
rm -rf models
python3 llm/prepare.py transformers_neuronx mixtral-8x7b-rb
./launch_container.sh deepjavalibrary/djl-serving:$DJLSERVING_DOCKER_TAG $PWD/models pytorch-inf2-4 \
serve
curl http://127.0.0.1:8080/models
python3 llm/client.py transformers_neuronx_rolling_batch mixtral-8x7b-rb
docker rm -f $(docker ps -aq)
sudo rm -rf models
- name: On fail step
if: ${{ failure() }}
working-directory: tests/integration
Expand Down
2 changes: 1 addition & 1 deletion tests/integration/launch_container.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -115,7 +115,7 @@ total_retries=24
if $is_llm; then
total_retries=60
if [[ "$platform" == *"inf2"* ]]; then
total_retries=80
total_retries=160
fi
if [[ "$platform" == *"trtllm"* || "$platform" == *"tensorrt-llm"* ]]; then
total_retries=150
Expand Down
5 changes: 5 additions & 0 deletions tests/integration/llm/client.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -182,6 +182,11 @@ def get_model_name():
"seq_length": [256],
"tokenizer": "mistralai/Mistral-7B-v0.1"
},
"mixtral-8x7b-rb": {
"batch_size": [4],
"seq_length": [256],
"tokenizer": "mistralai/Mixtral-8x7B-v0.1"
},
"llama-7b-rb": {
"batch_size": [1, 4],
"seq_length": [256],
Expand Down
27 changes: 9 additions & 18 deletions tests/integration/llm/prepare.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -184,15 +184,6 @@
"option.dtype": "fp16",
"option.model_loading_timeout": 900
},
"open-llama-7b": {
"option.model_id": "s3://djl-llm/open-llama-7b/",
"batch_size": 4,
"option.tensor_parallel_degree": 4,
"option.n_positions": 512,
"option.dtype": "fp16",
"option.neuron_optimize_level": 1,
"option.model_loading_timeout": 1200
},
"bloom-7b1": {
"option.model_id": "s3://djl-llm/bloom-7b1/",
"batch_size": 4,
Expand All @@ -201,15 +192,6 @@
"option.dtype": "fp16",
"option.model_loading_timeout": 1200
},
"llama-7b-split": {
"option.model_id": "s3://djl-llm/llama-2-7b-split-inf2/split-model/",
"batch_size": 1,
"option.tensor_parallel_degree": 4,
"option.n_positions": 512,
"option.model_loading_timeout": 2400,
"option.load_split_model": True,
"option.context_length_estimate": '256, 512, 1024'
},
"mistral-7b": {
"option.model_id": "s3://djl-llm/mistral-7b/",
"batch_size": 4,
Expand Down Expand Up @@ -277,6 +259,15 @@
"option.rolling_batch": 'auto',
"option.model_loading_timeout": 2400,
"option.output_formatter": "jsonlines"
},
"mixtral-8x7b-rb": {
"option.model_id": "s3://djl-llm/mixtral-8x7b/",
"option.tensor_parallel_degree": 8,
"option.n_positions": 1024,
"option.max_rolling_batch_size": 4,
"option.rolling_batch": 'auto',
"option.model_loading_timeout": 3600,
"option.output_formatter": "jsonlines"
}
}

Expand Down
Loading