423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-16 20:00:36 +08:00
Code Issues Packages Projects Releases Wiki Activity

[test] add npu test yaml and add ascend a3 docker file (#9547)

Browse Source 
Co-authored-by: jiaqiw09 <jiaqiw960714@gmail.com>
This commit is contained in:
Username_Full
2025-11-30 09:37:08 +08:00
committed by GitHub
parent 22be45c78c
commit e43a972b25
33 changed files with 322 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
tests/data/processor/test_feedback.py
View File

@@ -42,6 +42,7 @@ TRAIN_ARGS = {
}
@pytest.mark.runs_on(["cpu","npu"])
@pytest.mark.parametrize("num_samples", [16])
def test_feedback_data(num_samples: int):
train_dataset = load_dataset_module(**TRAIN_ARGS)["train_dataset"]

1
tests/data/processor/test_pairwise.py
View File

@@ -51,6 +51,7 @@ def _convert_sharegpt_to_openai(messages: list[dict[str, str]]) -> list[dict[str
return new_messages
@pytest.mark.runs_on(["cpu"])
@pytest.mark.parametrize("num_samples", [16])
def test_pairwise_data(num_samples: int):
train_dataset = load_dataset_module(**TRAIN_ARGS)["train_dataset"]

1
tests/data/processor/test_processor_utils.py
View File

@@ -18,6 +18,7 @@ import pytest
from llamafactory.data.processor.processor_utils import infer_seqlen
@pytest.mark.runs_on(["cpu"])
@pytest.mark.parametrize(
"test_input,test_output",
[

4
tests/data/processor/test_supervised.py
View File

@@ -42,6 +42,7 @@ TRAIN_ARGS = {
}
@pytest.mark.runs_on(["cpu"])
@pytest.mark.parametrize("num_samples", [16])
def test_supervised_single_turn(num_samples: int):
train_dataset = load_dataset_module(dataset_dir="ONLINE", dataset=TINY_DATA, **TRAIN_ARGS)["train_dataset"]
@@ -61,6 +62,7 @@ def test_supervised_single_turn(num_samples: int):
assert train_dataset["input_ids"][index] == ref_input_ids
@pytest.mark.runs_on(["cpu"])
@pytest.mark.parametrize("num_samples", [8])
def test_supervised_multi_turn(num_samples: int):
train_dataset = load_dataset_module(dataset_dir="REMOTE:" + DEMO_DATA, dataset="system_chat", **TRAIN_ARGS)[
@@ -74,6 +76,7 @@ def test_supervised_multi_turn(num_samples: int):
assert train_dataset["input_ids"][index] == ref_input_ids
@pytest.mark.runs_on(["cpu"])
@pytest.mark.parametrize("num_samples", [4])
def test_supervised_train_on_prompt(num_samples: int):
train_dataset = load_dataset_module(
@@ -88,6 +91,7 @@ def test_supervised_train_on_prompt(num_samples: int):
assert train_dataset["labels"][index] == ref_ids
@pytest.mark.runs_on(["cpu"])
@pytest.mark.parametrize("num_samples", [4])
def test_supervised_mask_history(num_samples: int):
train_dataset = load_dataset_module(

1
tests/data/processor/test_unsupervised.py
View File

@@ -45,6 +45,7 @@ TRAIN_ARGS = {
}
@pytest.mark.runs_on(["cpu"])
@pytest.mark.parametrize("num_samples", [16])
def test_unsupervised_data(num_samples: int):
train_dataset = load_dataset_module(**TRAIN_ARGS)["train_dataset"]