fc4e37920b
* Prepare for transformers v5 upgrade * fix hf cli * update for hf hub changes * fix tokenizer apply_chat_template args * remap include_tokens_per_second * fix tps * handle migration for warmup * use latest hf hub * Fix scan -> ls * fix import * fix for renaming of mistral common tokenizer -> backend * update for fixed tokenziation for llama * Skip phi35 tests for now * remove mistral patch fixed upstream in huggingface/transformers#41439 * use namespacing for patch * don't rely on sdist for e2e tests for now * run modal ci without waiting too * Fix dep for ci * fix imports * Fix fp8 check * fsdp2 fixes * fix version handling * update fsdp version tests for new v5 behavior * Fail multigpu tests after 3 failures * skip known v5 broken tests for now and cleanup * bump deps * unmark skipped test * re-enable test_fsdp_qlora_prequant_packed test * increase multigpu ci timeout * skip broken gemma3 test * reduce timout back to original 120min now that the hanging test is skipped * fix for un-necessary collator for pretraining with bsz=1 * fix: safe_serialization deprecated in transformers v5 rc01 (#3318) * torch_dtype deprecated * load model in float32 for consistency with tests * revert some test fixtures back * use hf cache ls instead of scan * don't strip fsdp_version more fdsp_Version fixes for v5 fix version in fsdp_config fix aliasing fix fsdp_version check check fsdp_version is 2 in both places * Transformers v5 rc2 (#3347) * bump dep * use latest fbgemm, grab model config as part of fixture, un-skip test * import AutoConfig * don't need more problematic autoconfig when specifying config.json manually * add fixtures for argilla ultrafeedback datasets * download phi4-reasoning * fix arg * update tests for phi fast tokenizer changes * use explicit model types for gemma3 --------- Co-authored-by: Wing Lian <wing@axolotl.ai> * fix: AutoModelForVision2Seq -> AutoModelForImageTextToText * chore: remove duplicate * fix: attempt fix gemma3 text mode * chore: lint * ga release of v5 * need property setter for name_or_path for mistral tokenizer * vllm not compatible with transformers v5 * setter for chat_template w mistral too --------- Co-authored-by: NanoCode012 <nano@axolotl.ai> Co-authored-by: salman <salman.mohammadi@outlook.com>
99 lines
3.2 KiB
Python
99 lines
3.2 KiB
Python
"""
|
|
E2E tests for multigpu lora tinyllama
|
|
"""
|
|
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
import yaml
|
|
from accelerate.test_utils import execute_subprocess_async
|
|
from huggingface_hub import snapshot_download
|
|
from transformers.testing_utils import get_torch_dist_unique_port
|
|
|
|
from axolotl.utils.dict import DictDefault
|
|
|
|
from tests.e2e.utils import check_tensorboard
|
|
|
|
AXOLOTL_ROOT = Path(__file__).parent.parent.parent.parent
|
|
|
|
|
|
@pytest.fixture(scope="session", autouse=True)
|
|
def download_model():
|
|
# download the model
|
|
snapshot_download("axolotl-mirrors/gemma-3-4b-pt", repo_type="model")
|
|
|
|
|
|
@pytest.mark.skip(reason="FIXME")
|
|
class TestMultiGPUGemma3:
|
|
"""
|
|
Test case for Gemma3 models using LoRA
|
|
"""
|
|
|
|
def test_lora_ddp_packed(self, temp_dir):
|
|
cfg = DictDefault(
|
|
{
|
|
"base_model": "axolotl-mirrors/gemma-3-4b-pt",
|
|
"unfrozen_parameters": ["model.language_model.*", "lm_head"],
|
|
"sequence_len": 2048,
|
|
"ddp_find_unused_parameters": True,
|
|
"sample_packing": True,
|
|
"eval_sample_packing": False,
|
|
"pad_to_sequence_len": True,
|
|
"adapter": "lora",
|
|
"lora_r": 8,
|
|
"lora_alpha": 16,
|
|
"lora_dropout": 0.05,
|
|
"lora_target_linear": True,
|
|
"val_set_size": 0.0,
|
|
"chat_template": "gemma3",
|
|
"datasets": [
|
|
{
|
|
"path": "mlabonne/FineTome-100k",
|
|
"type": "chat_template",
|
|
"split": "train[:10%]",
|
|
"field_messages": "conversations",
|
|
"message_field_role": "from",
|
|
"message_field_content": "value",
|
|
},
|
|
],
|
|
"num_epochs": 1,
|
|
"max_steps": 2,
|
|
"micro_batch_size": 4,
|
|
"gradient_checkpointing": True,
|
|
"gradient_checkpointing_kwargs": {
|
|
"use_reentrant": False,
|
|
},
|
|
"gradient_accumulation_steps": 2,
|
|
"output_dir": temp_dir,
|
|
"dataset_prepared_path": temp_dir + "/last_run_prepared",
|
|
"learning_rate": 0.0001,
|
|
"optimizer": "adamw_8bit",
|
|
"lr_scheduler": "cosine",
|
|
"flash_attention": True,
|
|
"use_tensorboard": True,
|
|
"bf16": True,
|
|
"save_first_step": False,
|
|
}
|
|
)
|
|
|
|
# write cfg to yaml file
|
|
Path(temp_dir).mkdir(parents=True, exist_ok=True)
|
|
with open(Path(temp_dir) / "config.yaml", "w", encoding="utf-8") as fout:
|
|
fout.write(yaml.dump(cfg.to_dict(), Dumper=yaml.Dumper))
|
|
|
|
execute_subprocess_async(
|
|
[
|
|
"axolotl",
|
|
"train",
|
|
str(Path(temp_dir) / "config.yaml"),
|
|
"--num-processes",
|
|
"2",
|
|
"--main-process-port",
|
|
f"{get_torch_dist_unique_port()}",
|
|
]
|
|
)
|
|
|
|
check_tensorboard(
|
|
temp_dir + "/runs", "train/train_loss", 1.8, "Train Loss (%s) is too high"
|
|
)
|