AttributeError: 'Qwen2Attention' object has no attribute 'max_position_embeddings'

[chengchengpei]

Reminder

• [X] I have read the README and searched the existing issues.

System Info

python3.10
-e git+https://github.com/hiyouga/LLaMA-Factory.git@bdde35fd2e4a919c1d63ebfc9a0ea8ba0c97e14c#egg=llamafactory

==((====))==  Unsloth 2024.8: Fast Qwen2 patching. Transformers = 4.45.0.dev0.
   \\   /|    GPU: NVIDIA L4. Max memory: 22.161 GB. Platform = Linux.
O^O/ \_/ \    Pytorch: 2.3.0+cu121. CUDA = 8.9. CUDA Toolkit = 12.1.
\        /    Bfloat16 = TRUE. FA [Xformers = 0.0.26.post1. FA2 = True]
 "-____-"     Free Apache license: http://github.com/unslothai/unsloth

Reproduction

I used the following config

{
    "model_name_or_path": "Qwen/Qwen2-1.5B-Instruct",
    "use_unsloth": True,
    "stage": "sft",
    "do_train": True,
    "finetuning_type": "full",
    "lora_target": "all",
    "use_galore": True,
    # "enable_liger_kernel": True,
    "galore_layerwise": True,
    "galore_target": "mlp,self_attn",
    "galore_rank": 128,
    "galore_scale": 2.0,
    "dataset": "identity,alpaca_en_demo",
    "template": "llama3",
    "cutoff_len": 1024,
    "max_samples": 1000,
    "overwrite_cache": True,
    "preprocessing_num_workers": 4,
    "output_dir": "saves/llama3-8b/lora/sft",
    "logging_steps": 10,
    "save_steps": 500,
    "plot_loss": True,
    "overwrite_output_dir": True,
    "per_device_train_batch_size": 1,
    # "gradient_accumulation_steps": 2,
    "learning_rate": 0.0001,
    "num_train_epochs": 3,
    "lr_scheduler_type": "cosine",
    "warmup_ratio": 0.1,
    "bf16": True,
    "ddp_timeout": 180000000,
    "val_size": 0.1,
    "per_device_eval_batch_size": 1,
    "eval_strategy": "steps",
    "eval_steps": 500
}

but i got errors:

Traceback (most recent call last):
  File "/usr/local/bin/llamafactory-cli", line 8, in <module>
    sys.exit(main())
  File "/workspace/LLaMA-Factory/src/llamafactory/cli.py", line 111, in main
    run_exp()
  File "/workspace/LLaMA-Factory/src/llamafactory/train/tuner.py", line 50, in run_exp
    run_sft(model_args, data_args, training_args, finetuning_args, generating_args, callbacks)
  File "/workspace/LLaMA-Factory/src/llamafactory/train/sft/workflow.py", line 48, in run_sft
    model = load_model(tokenizer, model_args, finetuning_args, training_args.do_train)
  File "/workspace/LLaMA-Factory/src/llamafactory/model/loader.py", line 145, in load_model
    model = load_unsloth_pretrained_model(config, model_args)
  File "/workspace/LLaMA-Factory/src/llamafactory/model/model_utils/unsloth.py", line 57, in load_unsloth_pretrained_model
    model, _ = FastLanguageModel.from_pretrained(**unsloth_kwargs)
  File "/usr/local/lib/python3.10/dist-packages/unsloth/models/loader.py", line 301, in from_pretrained
    model, tokenizer = dispatch_model.from_pretrained(
  File "/usr/local/lib/python3.10/dist-packages/unsloth/models/qwen2.py", line 87, in from_pretrained
    return FastLlamaModel.from_pretrained(
  File "/usr/local/lib/python3.10/dist-packages/unsloth/models/llama.py", line 1580, in from_pretrained
    model = AutoModelForCausalLM.from_pretrained(
  File "/usr/local/lib/python3.10/dist-packages/transformers/models/auto/auto_factory.py", line 557, in from_pretrained
    return model_class.from_pretrained(
  File "/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py", line 3848, in from_pretrained
    model = cls(config, *model_args, **model_kwargs)
  File "/usr/local/lib/python3.10/dist-packages/transformers/models/qwen2/modeling_qwen2.py", line 1080, in __init__
    self.model = Qwen2Model(config)
  File "/usr/local/lib/python3.10/dist-packages/transformers/models/qwen2/modeling_qwen2.py", line 868, in __init__
    [Qwen2DecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
  File "/usr/local/lib/python3.10/dist-packages/transformers/models/qwen2/modeling_qwen2.py", line 868, in <listcomp>
    [Qwen2DecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
  File "/usr/local/lib/python3.10/dist-packages/transformers/models/qwen2/modeling_qwen2.py", line 655, in __init__
    self.self_attn = QWEN2_ATTENTION_CLASSES[config._attn_implementation](config, layer_idx)
  File "<string>", line 39, in Qwen2Attention__init__
  File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1709, in __getattr__
    raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'")
AttributeError: 'Qwen2Attention' object has no attribute 'max_position_embeddings'

Expected behavior

No response

Others

No response

[chengchengpei]

https://github.com/huggingface/transformers/pull/33550