To create a public link, set `share=True` in `launch()`. 

Downloading Model to directory: /root/.cache/modelscope/hub/models/Qwen/Qwen2.5-0.5B 

2025-02-18 16:55:56,769 - modelscope - WARNING - Using branch: master as version is unstable, use with caution 

2025-02-18 16:55:57,005 - modelscope - INFO - Target directory already exists, skipping creation. 

[INFO|configuration_utils.py:697] 2025-02-18 16:55:57,006 >> loading configuration file /root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B/config.json 

[INFO|configuration_utils.py:771] 2025-02-18 16:55:57,008 >> Model config Qwen2Config { 

  "_name_or_path": "/root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B", 

  "architectures": [ 

    "Qwen2ForCausalLM" 

  ], 

  "attention_dropout": 0.0, 

  "bos_token_id": 151643, 

  "eos_token_id": 151643, 

  "hidden_act": "silu", 

  "hidden_size": 896, 

  "initializer_range": 0.02, 

  "intermediate_size": 4864, 

  "max_position_embeddings": 32768, 

  "max_window_layers": 24, 

  "model_type": "qwen2", 

  "num_attention_heads": 14, 

  "num_hidden_layers": 24, 

  "num_key_value_heads": 2, 

  "rms_norm_eps": 1e-06, 

  "rope_scaling": null, 

  "rope_theta": 1000000.0, 

  "sliding_window": 32768, 

  "tie_word_embeddings": true, 

  "torch_dtype": "bfloat16", 

  "transformers_version": "4.49.0", 

  "use_cache": true, 

  "use_mrope": false, 

  "use_sliding_window": false, 

  "vocab_size": 151936 

} 

 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,009 >> loading file vocab.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,009 >> loading file merges.txt 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,009 >> loading file tokenizer.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,009 >> loading file added_tokens.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,009 >> loading file special_tokens_map.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,009 >> loading file tokenizer_config.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,009 >> loading file chat_template.jinja 

[INFO|tokenization_utils_base.py:2313] 2025-02-18 16:55:57,396 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. 

[INFO|configuration_utils.py:697] 2025-02-18 16:55:57,398 >> loading configuration file /root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B/config.json 

[INFO|configuration_utils.py:771] 2025-02-18 16:55:57,400 >> Model config Qwen2Config { 

  "_name_or_path": "/root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B", 

  "architectures": [ 

    "Qwen2ForCausalLM" 

  ], 

  "attention_dropout": 0.0, 

  "bos_token_id": 151643, 

  "eos_token_id": 151643, 

  "hidden_act": "silu", 

  "hidden_size": 896, 

  "initializer_range": 0.02, 

  "intermediate_size": 4864, 

  "max_position_embeddings": 32768, 

  "max_window_layers": 24, 

  "model_type": "qwen2", 

  "num_attention_heads": 14, 

  "num_hidden_layers": 24, 

  "num_key_value_heads": 2, 

  "rms_norm_eps": 1e-06, 

  "rope_scaling": null, 

  "rope_theta": 1000000.0, 

  "sliding_window": 32768, 

  "tie_word_embeddings": true, 

  "torch_dtype": "bfloat16", 

  "transformers_version": "4.49.0", 

  "use_cache": true, 

  "use_mrope": false, 

  "use_sliding_window": false, 

  "vocab_size": 151936 

} 

 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,401 >> loading file vocab.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,401 >> loading file merges.txt 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,401 >> loading file tokenizer.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,401 >> loading file added_tokens.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,401 >> loading file special_tokens_map.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,401 >> loading file tokenizer_config.json 

[INFO|tokenization_utils_base.py:2048] 2025-02-18 16:55:57,401 >> loading file chat_template.jinja 

[INFO|tokenization_utils_base.py:2313] 2025-02-18 16:55:57,781 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. 

[INFO|configuration_utils.py:697] 2025-02-18 16:55:57,820 >> loading configuration file /root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B/config.json 

[INFO|configuration_utils.py:771] 2025-02-18 16:55:57,822 >> Model config Qwen2Config { 

  "_name_or_path": "/root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B", 

  "architectures": [ 

    "Qwen2ForCausalLM" 

  ], 

  "attention_dropout": 0.0, 

  "bos_token_id": 151643, 

  "eos_token_id": 151643, 

  "hidden_act": "silu", 

  "hidden_size": 896, 

  "initializer_range": 0.02, 

  "intermediate_size": 4864, 

  "max_position_embeddings": 32768, 

  "max_window_layers": 24, 

  "model_type": "qwen2", 

  "num_attention_heads": 14, 

  "num_hidden_layers": 24, 

  "num_key_value_heads": 2, 

  "rms_norm_eps": 1e-06, 

  "rope_scaling": null, 

  "rope_theta": 1000000.0, 

  "sliding_window": 32768, 

  "tie_word_embeddings": true, 

  "torch_dtype": "bfloat16", 

  "transformers_version": "4.49.0", 

  "use_cache": true, 

  "use_mrope": false, 

  "use_sliding_window": false, 

  "vocab_size": 151936 

} 

 

[INFO|2025-02-18 16:55:57] llamafactory.model.patcher:157 >> Using KV cache for faster generation. 

[INFO|modeling_utils.py:3979] 2025-02-18 16:55:57,887 >> loading weights file /root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B/model.safetensors 

[INFO|modeling_utils.py:1633] 2025-02-18 16:55:57,903 >> Instantiating Qwen2ForCausalLM model under default dtype torch.float16. 

[INFO|configuration_utils.py:1140] 2025-02-18 16:55:57,907 >> Generate config GenerationConfig { 

  "bos_token_id": 151643, 

  "eos_token_id": 151643 

} 

 

[WARNING|logging.py:329] 2025-02-18 16:55:57,919 >> Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered. 

[INFO|modeling_utils.py:4970] 2025-02-18 16:55:59,482 >> All model checkpoint weights were used when initializing Qwen2ForCausalLM. 

 

[INFO|modeling_utils.py:4978] 2025-02-18 16:55:59,482 >> All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B. 

If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. 

[INFO|configuration_utils.py:1093] 2025-02-18 16:55:59,493 >> loading configuration file /root/.cache/modelscope/hub/models/Qwen/Qwen2___5-0___5B/generation_config.json 

[INFO|configuration_utils.py:1140] 2025-02-18 16:55:59,494 >> Generate config GenerationConfig { 

  "bos_token_id": 151643, 

  "eos_token_id": 151643, 

  "max_new_tokens": 2048 

} 

 

[INFO|2025-02-18 16:55:59] llamafactory.model.model_utils.attention:157 >> Using torch SDPA for faster training and inference. 

[INFO|2025-02-18 16:55:59] llamafactory.model.loader:157 >> all params: 494,032,768 

[WARNING|2025-02-18 16:55:59] llamafactory.chat.hf_engine:168 >> There is no current event loop, creating a new one. 

[WARNING|logging.py:329] 2025-02-18 16:55:57,919 >> Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered. 

Segmentation fault (core dumped) 

root@4bd629244b70:/app/src# [WARNING|logging.py:329] 2025-02-18 16:55:57,919 >> Sliding Window Attention is enabled but not implemented for `sdpa`; unexpected results may be encountered. 

bash: [WARNING: command not found 

bash: sdpa: command not found 

bash: logging.py:329]: command not found 

bash: unexpected: command not found 

root@4bd629244b70:/app/src# [ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared! 

[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared! 

[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared! 

[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared! 

[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared! 

[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared! 

[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared! 

[ERROR] TBE Subprocess[task_distribute] raise error[], main process disappeared!