[INFO|2025-05-29 19:59:18] tokenization_utils_base.py:2034 >> loading file vocab.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/vocab.json [INFO|2025-05-29 19:59:18] tokenization_utils_base.py:2034 >> loading file merges.txt from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/merges.txt [INFO|2025-05-29 19:59:18] tokenization_utils_base.py:2034 >> loading file tokenizer.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/tokenizer.json [INFO|2025-05-29 19:59:18] tokenization_utils_base.py:2034 >> loading file added_tokens.json from cache at None [INFO|2025-05-29 19:59:18] tokenization_utils_base.py:2034 >> loading file special_tokens_map.json from cache at None [INFO|2025-05-29 19:59:18] tokenization_utils_base.py:2034 >> loading file tokenizer_config.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/tokenizer_config.json [INFO|2025-05-29 19:59:18] tokenization_utils_base.py:2034 >> loading file chat_template.jinja from cache at None [INFO|2025-05-29 19:59:19] tokenization_utils_base.py:2304 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. [INFO|2025-05-29 19:59:20] configuration_utils.py:696 >> loading configuration file config.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/config.json [INFO|2025-05-29 19:59:20] configuration_utils.py:768 >> Model config Qwen2Config { "_name_or_path": "Qwen/Qwen2.5-Coder-7B-Instruct", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 151643, "eos_token_id": 151645, "hidden_act": "silu", "hidden_size": 3584, "initializer_range": 0.02, "intermediate_size": 18944, "max_position_embeddings": 32768, "max_window_layers": 28, "model_type": "qwen2", "num_attention_heads": 28, "num_hidden_layers": 28, "num_key_value_heads": 4, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.48.2", "use_cache": true, "use_sliding_window": false, "vocab_size": 152064 } [INFO|2025-05-29 19:59:20] tokenization_utils_base.py:2034 >> loading file vocab.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/vocab.json [INFO|2025-05-29 19:59:20] tokenization_utils_base.py:2034 >> loading file merges.txt from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/merges.txt [INFO|2025-05-29 19:59:20] tokenization_utils_base.py:2034 >> loading file tokenizer.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/tokenizer.json [INFO|2025-05-29 19:59:20] tokenization_utils_base.py:2034 >> loading file added_tokens.json from cache at None [INFO|2025-05-29 19:59:20] tokenization_utils_base.py:2034 >> loading file special_tokens_map.json from cache at None [INFO|2025-05-29 19:59:20] tokenization_utils_base.py:2034 >> loading file tokenizer_config.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/tokenizer_config.json [INFO|2025-05-29 19:59:20] tokenization_utils_base.py:2034 >> loading file chat_template.jinja from cache at None [INFO|2025-05-29 19:59:21] tokenization_utils_base.py:2304 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. [INFO|2025-05-29 19:59:21] logging.py:157 >> Add <|im_end|> to stop words. [INFO|2025-05-29 19:59:21] logging.py:157 >> Loading dataset Codes_query_filtered_330k_ns_over8_1.json... [INFO|2025-05-29 19:59:33] configuration_utils.py:696 >> loading configuration file config.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/config.json [INFO|2025-05-29 19:59:33] configuration_utils.py:768 >> Model config Qwen2Config { "_name_or_path": "Qwen/Qwen2.5-Coder-7B-Instruct", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 151643, "eos_token_id": 151645, "hidden_act": "silu", "hidden_size": 3584, "initializer_range": 0.02, "intermediate_size": 18944, "max_position_embeddings": 32768, "max_window_layers": 28, "model_type": "qwen2", "num_attention_heads": 28, "num_hidden_layers": 28, "num_key_value_heads": 4, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.48.2", "use_cache": true, "use_sliding_window": false, "vocab_size": 152064 } [WARNING|2025-05-29 19:59:33] logging.py:162 >> Input length is smaller than max length. Consider increase input length. [INFO|2025-05-29 19:59:33] logging.py:157 >> Using llama3 scaling strategy and setting scaling factor to 1.0. [INFO|2025-05-29 19:59:33] logging.py:157 >> Using block diagonal attention for sequence packing without cross-attention. [INFO|2025-05-29 19:59:33] logging.py:157 >> Liger kernel has been applied to the model. [INFO|2025-05-29 19:59:33] modeling_utils.py:3904 >> loading weights file model.safetensors from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/model.safetensors.index.json [INFO|2025-05-29 19:59:33] modeling_utils.py:1582 >> Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. [INFO|2025-05-29 19:59:33] configuration_utils.py:1140 >> Generate config GenerationConfig { "bos_token_id": 151643, "eos_token_id": 151645 } [INFO|2025-05-29 19:59:37] modeling_utils.py:4888 >> All model checkpoint weights were used when initializing Qwen2ForCausalLM. [INFO|2025-05-29 19:59:37] modeling_utils.py:4896 >> All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at Qwen/Qwen2.5-Coder-7B-Instruct. 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|2025-05-29 19:59:37] configuration_utils.py:1095 >> loading configuration file generation_config.json from cache at /home/kiho/.cache/huggingface/hub/models--Qwen--Qwen2.5-Coder-7B-Instruct/snapshots/c03e6d358207e414f1eca0bb1891e29f1db0e242/generation_config.json [INFO|2025-05-29 19:59:37] configuration_utils.py:1140 >> Generate config GenerationConfig { "bos_token_id": 151643, "do_sample": true, "eos_token_id": [ 151645, 151643 ], "pad_token_id": 151643, "repetition_penalty": 1.1, "temperature": 0.7, "top_k": 20, "top_p": 0.8 } [INFO|2025-05-29 19:59:37] logging.py:157 >> Gradient checkpointing enabled. [INFO|2025-05-29 19:59:37] logging.py:157 >> Using torch SDPA for faster training and inference. [INFO|2025-05-29 19:59:37] logging.py:157 >> Upcasting trainable params to float32. [INFO|2025-05-29 19:59:37] logging.py:157 >> Fine-tuning method: Freeze [INFO|2025-05-29 19:59:37] logging.py:157 >> Set trainable layers: .26.,.27. [INFO|2025-05-29 19:59:37] logging.py:157 >> trainable params: 466,115,584 || all params: 7,615,616,512 || trainable%: 6.1205 [INFO|2025-05-29 19:59:37] trainer.py:741 >> Using auto half precision backend [INFO|2025-05-29 19:59:37] logging.py:157 >> Found linear modules: q_proj,gate_proj,up_proj,o_proj,v_proj,k_proj,down_proj [INFO|2025-05-29 19:59:37] logging.py:157 >> Using APOLLO optimizer with args: {'rank': 256, 'proj': 'random', 'proj_type': 'std', 'update_proj_gap': 200, 'scale': 1, 'scale_type': 'channel', 'scale_front': False}. [INFO|2025-05-29 19:59:38] trainer.py:2369 >> ***** Running training ***** [INFO|2025-05-29 19:59:38] trainer.py:2370 >> Num examples = 6,260 [INFO|2025-05-29 19:59:38] trainer.py:2371 >> Num Epochs = 1 [INFO|2025-05-29 19:59:38] trainer.py:2372 >> Instantaneous batch size per device = 16 [INFO|2025-05-29 19:59:38] trainer.py:2375 >> Total train batch size (w. parallel, distributed & accumulation) = 512 [INFO|2025-05-29 19:59:38] trainer.py:2376 >> Gradient Accumulation steps = 8 [INFO|2025-05-29 19:59:38] trainer.py:2377 >> Total optimization steps = 12 [INFO|2025-05-29 19:59:38] trainer.py:2378 >> Number of trainable parameters = 466,115,584 [INFO|2025-05-29 20:01:27] logging.py:157 >> {'loss': 0.6759, 'learning_rate': 4.9148e-05, 'epoch': 0.08, 'throughput': 19284.83} [INFO|2025-05-29 20:03:09] logging.py:157 >> {'loss': 0.6698, 'learning_rate': 4.6651e-05, 'epoch': 0.16, 'throughput': 19910.87} [INFO|2025-05-29 20:04:51] logging.py:157 >> {'loss': 0.6228, 'learning_rate': 4.2678e-05, 'epoch': 0.24, 'throughput': 20123.38} [INFO|2025-05-29 20:06:33] logging.py:157 >> {'loss': 0.6228, 'learning_rate': 3.7500e-05, 'epoch': 0.33, 'throughput': 20244.72} [INFO|2025-05-29 20:08:15] logging.py:157 >> {'loss': 0.6104, 'learning_rate': 3.1470e-05, 'epoch': 0.41, 'throughput': 20320.48} [INFO|2025-05-29 20:09:57] logging.py:157 >> {'loss': 0.6063, 'learning_rate': 2.5000e-05, 'epoch': 0.49, 'throughput': 20365.11} [INFO|2025-05-29 20:11:38] logging.py:157 >> {'loss': 0.6236, 'learning_rate': 1.8530e-05, 'epoch': 0.57, 'throughput': 20397.20} [INFO|2025-05-29 20:13:20] logging.py:157 >> {'loss': 0.6225, 'learning_rate': 1.2500e-05, 'epoch': 0.65, 'throughput': 20419.47} [INFO|2025-05-29 20:15:02] logging.py:157 >> {'loss': 0.5952, 'learning_rate': 7.3223e-06, 'epoch': 0.73, 'throughput': 20439.93} [INFO|2025-05-29 20:16:45] logging.py:157 >> {'loss': 0.6016, 'learning_rate': 3.3494e-06, 'epoch': 0.82, 'throughput': 20429.94} [INFO|2025-05-29 20:18:27] logging.py:157 >> {'loss': 0.5926, 'learning_rate': 8.5185e-07, 'epoch': 0.90, 'throughput': 20439.47} [INFO|2025-05-29 20:20:09] logging.py:157 >> {'loss': 0.6137, 'learning_rate': 0.0000e+00, 'epoch': 0.98, 'throughput': 20449.96} [INFO|2025-05-29 20:20:09] trainer.py:3910 >> Saving model checkpoint to saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/checkpoint-12 [INFO|2025-05-29 20:20:09] configuration_utils.py:420 >> Configuration saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/checkpoint-12/config.json [INFO|2025-05-29 20:20:09] configuration_utils.py:909 >> Configuration saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/checkpoint-12/generation_config.json [INFO|2025-05-29 20:20:32] modeling_utils.py:2996 >> The model is bigger than the maximum size per checkpoint (5GB) and is going to be split in 4 checkpoint shards. You can find where each parameters has been saved in the index located at saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/checkpoint-12/model.safetensors.index.json. [INFO|2025-05-29 20:20:32] tokenization_utils_base.py:2491 >> tokenizer config file saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/checkpoint-12/tokenizer_config.json [INFO|2025-05-29 20:20:32] tokenization_utils_base.py:2500 >> Special tokens file saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/checkpoint-12/special_tokens_map.json [INFO|2025-05-29 20:20:33] trainer.py:2643 >> Training completed. Do not forget to share your model on huggingface.co/models =) [INFO|2025-05-29 20:20:33] trainer.py:3910 >> Saving model checkpoint to saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1 [INFO|2025-05-29 20:20:33] configuration_utils.py:420 >> Configuration saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/config.json [INFO|2025-05-29 20:20:33] configuration_utils.py:909 >> Configuration saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/generation_config.json [INFO|2025-05-29 20:20:56] modeling_utils.py:2996 >> The model is bigger than the maximum size per checkpoint (5GB) and is going to be split in 4 checkpoint shards. You can find where each parameters has been saved in the index located at saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/model.safetensors.index.json. [INFO|2025-05-29 20:20:56] tokenization_utils_base.py:2491 >> tokenizer config file saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/tokenizer_config.json [INFO|2025-05-29 20:20:56] tokenization_utils_base.py:2500 >> Special tokens file saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nlx_8_1/special_tokens_map.json [WARNING|2025-05-29 20:20:56] logging.py:162 >> No metric eval_loss to plot. [WARNING|2025-05-29 20:20:56] logging.py:162 >> No metric eval_accuracy to plot. [INFO|2025-05-29 20:20:56] modelcard.py:449 >> Dropping the following result as it does not have all the necessary fields: {'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}