forked from vllm-project/vllm
-
Notifications
You must be signed in to change notification settings - Fork 78
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Update compile scenarios to match with lazy test suite
- Loading branch information
1 parent
489a526
commit 391f101
Showing
4 changed files
with
112 additions
and
59 deletions.
There are no files selected for viewing
This file was deleted.
Oops, something went wrong.
This file was deleted.
Oops, something went wrong.
This file contains 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
This file contains 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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,33 +1,101 @@ | ||
| # test_config_t_compile.yaml | ||
| # test_config.yaml | ||
| stages: | ||
| - name: test_gsm8k_small_models_tcompile | ||
| - name: test_gsm8k_small_models | ||
| steps: | ||
| - name: gsm8k_small_g3_tp1_tc | ||
| - name: gsm8k_small_g3_tp1 | ||
| flavor: g3 | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-small.txt -t 1 | ||
| - name: gsm8k_small_g3_tp2_tc | ||
| - name: gsm8k_small_g3_tp2 | ||
| flavor: g3.s | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-small.txt -t 2 | ||
| - name: gsm8k_small_g2_tp1_tc | ||
| - name: gsm8k_small_g2_tp1 | ||
| flavor: g2 | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-small.txt -t 1 | ||
| - name: gsm8k_small_g2_tp2_tc | ||
| - name: gsm8k_small_g2_tp2 | ||
| flavor: g2.s | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-small.txt -t 2 | ||
| - name: gsm8k_llama2_g3_tp1_tc | ||
| - name: test_gsm8k_large_models | ||
| steps: | ||
| - name: gsm8k_large_g3_tp2 | ||
| flavor: g3.s | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-large.txt -t 2 | ||
| - name: gsm8k_large_g2_tp4 | ||
| flavor: g2.m | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-large.txt -t 4 | ||
| - name: test_gsm8k_fp8 | ||
| steps: | ||
| - name: gsm8k_small_g3_tp1_fp8 | ||
| flavor: g3 | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-llama2.txt -t 1 | ||
| - name: gsm8k_llama2_g3_tp2_tc | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-fp8.txt -t 1 | ||
| - name: gsm8k_small_g3_tp2_fp8 | ||
| flavor: g3.s | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-llama2.txt -t 2 | ||
| - name: gsm8k_lama2_g2_tp1_tc | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-fp8.txt -t 2 | ||
| - name: test_gsm8k_mss | ||
| steps: | ||
| - name: gsm8k_small_g3_tp1_mss | ||
| flavor: g3 | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-mss.txt -t 1 | ||
| - name: gsm8k_small_g2_tp1_mss | ||
| flavor: g2 | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-llama2.txt -t 1 | ||
| - name: gsm8k_lama2_g2_tp2_tc | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-mss.txt -t 1 | ||
| - name: gsm8k_small_g3_tp2_mss | ||
| flavor: g3.s | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-mss.txt -t 2 | ||
| - name: gsm8k_small_g2_tp2_mss | ||
| flavor: g2.s | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-llama2.txt -t 2 | ||
| - name: tests_lora_compile | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-mss.txt -t 2 | ||
| - name: gsm8k_small_g2_tp1_spec_decode | ||
| flavor: g2 | ||
| command: cd .jenkins/lm-eval-harness && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-mss.txt -t 1 | ||
| - name: test_gsm8k_spec_decode | ||
| steps: | ||
| - name: gsm8k_small_g2_tp1_mlp_spec_decode | ||
| flavor: g2 | ||
| command: PT_HPU_LAZY_MODE=0 TORCH_COMPILE_DISABLE=true VLLM_CONTIGUOUS_PA=false VLLM_SKIP_WARMUP=True pytest -v tests/spec_decode/e2e/test_mlp_correctness.py::test_mlp_e2e_greedy_correctness | ||
| - name: gsm8k_small_g2_tp1_medusa_spec_decode | ||
| flavor: g2 | ||
| command: PT_HPU_LAZY_MODE=0 TORCH_COMPILE_DISABLE=true VLLM_CONTIGUOUS_PA=false VLLM_SKIP_WARMUP=True pytest -v tests/spec_decode/e2e/test_medusa_correctness.py::test_medusa_e2e_greedy_correctness | ||
| - name: gsm8k_small_g2_tp1_eagle_spec_decode | ||
| flavor: g2 | ||
| command: PT_HPU_LAZY_MODE=0 VLLM_COS_SIN_RECOMPUTE=true TORCH_COMPILE_DISABLE=true VLLM_CONTIGUOUS_PA=false VLLM_SKIP_WARMUP=True pytest -v tests/spec_decode/e2e/test_eagle_correctness.py::test_eagle_e2e_greedy_correctness | ||
| - name: tests_lora | ||
| steps: | ||
| - name: test_multilora_compile | ||
| - name: test_llama_lora | ||
| flavor: g2 | ||
| command: PT_HPU_LAZY_MODE=0 VLLM_SKIP_WARMUP=true pytest -v tests/lora/test_multilora_hpu.py::test_llama_multilora_1x | ||
| command: PT_HPU_LAZY_MODE=0 VLLM_SKIP_WARMUP=true pytest -v tests/lora/test_llama_hpu.py::test_llama_lora_1x | ||
| - name: test_multilora | ||
| flavor: g2 | ||
| command: PT_HPU_LAZY_MODE=0 VLLM_SKIP_WARMUP=true pytest -v tests/lora/test_multilora_hpu.py::test_llama_multilora_1x | ||
| - name: test_long_context | ||
| flavor: g2 | ||
| command: PT_HPU_LAZY_MODE=0 VLLM_SKIP_WARMUP=true pytest -v tests/lora/test_long_context_hpu.py::test_quality | ||
| - name: tests_multimodal | ||
| steps: | ||
| - name: multimodal_small_g3_tp1 | ||
| flavor: g3 | ||
| command: cd .jenkins/vision && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-small.txt -t 1 | ||
| - name: multimodal_small_g3_tp2 | ||
| flavor: g3.s | ||
| command: cd .jenkins/vision && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-small.txt -t 2 | ||
| - name: multimodal_small_g3_tp1_mss | ||
| flavor: g3 | ||
| command: cd .jenkins/vision && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-mss.txt -t 1 | ||
| - name: multimodal_small_g3_tp2_mss | ||
| flavor: g3.s | ||
| command: cd .jenkins/vision && PT_HPU_LAZY_MODE=0 bash run-tests.sh -c configs/models-mss.txt -t 2 | ||
| - name: tests_int4_quantization | ||
| steps: | ||
| - name: test_awq | ||
| flavor: g2 | ||
| command: PT_HPU_LAZY_MODE=0 VLLM_SKIP_WARMUP=true pytest -v tests/quantization/test_awq.py::test_awq | ||
| - name: test_gptq | ||
| flavor: g2 | ||
| command: PT_HPU_LAZY_MODE=0 VLLM_SKIP_WARMUP=true pytest -v tests/quantization/test_gptq.py::test_gptq | ||
| - name: tests_guided_decode | ||
| steps: | ||
| - name: test_lazy_outlines | ||
| flavor: g2 | ||
| command: export VLLM_SKIP_WARMUP=true && pip install -e tests/vllm_test_utils && PT_HPU_LAZY_MODE=0 pytest -v tests/entrypoints/llm/test_lazy_outlines.py -s -vvv --log-cli-level=INFO | ||
| - name: test_guided_generate | ||
| flavor: g2 | ||
| command: export VLLM_SKIP_WARMUP=true && pip install -e tests/vllm_test_utils && PT_HPU_LAZY_MODE=0 pytest -v tests/entrypoints/llm/test_guided_generate.py -s -vvv --log-cli-level=INFO |