ab4b32187d
* need to update deepspeed version in extras too * fix patch import * fix monkeypatch reloading in tests and deepspeed patch * remove duplicated functionality fixture * reset LlamaForCausalLM too in fixtures for cce patch * reset llama attn too * disable xformers patch for cce * skip problematic test on low usage functionality
90 lines
2.8 KiB
Python
90 lines
2.8 KiB
Python
"""
|
|
E2E tests for lora llama
|
|
"""
|
|
|
|
import logging
|
|
import os
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
from transformers.utils import is_torch_bf16_gpu_available
|
|
|
|
from axolotl.cli import load_datasets
|
|
from axolotl.common.cli import TrainerCliArgs
|
|
from axolotl.train import train
|
|
from axolotl.utils.config import normalize_config
|
|
from axolotl.utils.dict import DictDefault
|
|
|
|
from ..utils import check_tensorboard
|
|
|
|
LOG = logging.getLogger("axolotl.tests.e2e")
|
|
os.environ["WANDB_DISABLED"] = "true"
|
|
|
|
|
|
class TestFAXentropyLlama:
|
|
"""
|
|
Test case for Llama models using LoRA w multipack
|
|
"""
|
|
|
|
@pytest.mark.parametrize(
|
|
"gradient_accumulation_steps",
|
|
[1, 4],
|
|
)
|
|
def test_lora_packing_fa_cross_entropy(self, temp_dir, gradient_accumulation_steps):
|
|
# pylint: disable=duplicate-code
|
|
cfg = DictDefault(
|
|
{
|
|
"base_model": "HuggingFaceTB/SmolLM2-135M",
|
|
"sequence_len": 1024,
|
|
"sample_packing": True,
|
|
"flash_attention": True,
|
|
"flash_attn_cross_entropy": True,
|
|
"load_in_8bit": True,
|
|
"adapter": "lora",
|
|
"lora_r": 8,
|
|
"lora_alpha": 16,
|
|
"lora_dropout": 0.05,
|
|
"lora_target_linear": True,
|
|
"val_set_size": 0.05,
|
|
"special_tokens": {
|
|
"pad_token": "<|endoftext|>",
|
|
},
|
|
"chat_template": "chatml",
|
|
"datasets": [
|
|
{
|
|
"path": "mlabonne/FineTome-100k",
|
|
"field_messages": "conversations",
|
|
"message_field_content": "value",
|
|
"message_field_role": "from",
|
|
"type": "chat_template",
|
|
"split": "train[:2%]",
|
|
},
|
|
],
|
|
"num_epochs": 1,
|
|
"max_steps": 5,
|
|
"save_steps": 5,
|
|
"micro_batch_size": 2,
|
|
"gradient_accumulation_steps": gradient_accumulation_steps,
|
|
"output_dir": temp_dir,
|
|
"learning_rate": 0.00001,
|
|
"optimizer": "adamw_8bit",
|
|
"lr_scheduler": "cosine",
|
|
"use_tensorboard": True,
|
|
}
|
|
)
|
|
if is_torch_bf16_gpu_available():
|
|
cfg.bf16 = True
|
|
else:
|
|
cfg.fp16 = True
|
|
|
|
normalize_config(cfg)
|
|
cli_args = TrainerCliArgs()
|
|
dataset_meta = load_datasets(cfg=cfg, cli_args=cli_args)
|
|
|
|
train(cfg=cfg, cli_args=cli_args, dataset_meta=dataset_meta)
|
|
assert (Path(temp_dir) / "adapter_model.bin").exists()
|
|
|
|
check_tensorboard(
|
|
temp_dir + "/runs", "train/train_loss", 1.5, "Train Loss is too high"
|
|
)
|