File size: 2,122 Bytes
bc55b34 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 |
# Copyright 2024 the LlamaFactory team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import os
import pytest
from llamafactory.train.tuner import export_model, run_exp
DEMO_DATA = os.environ.get("DEMO_DATA", "llamafactory/demo_data")
TINY_LLAMA = os.environ.get("TINY_LLAMA", "llamafactory/tiny-random-Llama-3")
TINY_LLAMA_ADAPTER = os.environ.get("TINY_LLAMA_ADAPTER", "llamafactory/tiny-random-Llama-3-lora")
TRAIN_ARGS = {
"model_name_or_path": TINY_LLAMA,
"do_train": True,
"finetuning_type": "lora",
"dataset_dir": "REMOTE:" + DEMO_DATA,
"template": "llama3",
"cutoff_len": 1,
"overwrite_cache": False,
"overwrite_output_dir": True,
"per_device_train_batch_size": 1,
"max_steps": 1,
}
INFER_ARGS = {
"model_name_or_path": TINY_LLAMA,
"adapter_name_or_path": TINY_LLAMA_ADAPTER,
"finetuning_type": "lora",
"template": "llama3",
"infer_dtype": "float16",
"export_dir": "llama3_export",
}
OS_NAME = os.environ.get("OS_NAME", "")
@pytest.mark.parametrize(
"stage,dataset",
[
("pt", "c4_demo"),
("sft", "alpaca_en_demo"),
("dpo", "dpo_en_demo"),
("kto", "kto_en_demo"),
pytest.param("rm", "dpo_en_demo", marks=pytest.mark.xfail(OS_NAME.startswith("windows"), reason="OS error.")),
],
)
def test_run_exp(stage: str, dataset: str):
output_dir = "train_{}".format(stage)
run_exp({"stage": stage, "dataset": dataset, "output_dir": output_dir, **TRAIN_ARGS})
assert os.path.exists(output_dir)
def test_export():
export_model(INFER_ARGS)
assert os.path.exists("llama3_export")
|