running_log.txt

[INFO|2025-04-28 17:07:05] 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|2025-04-28 17:07:05] 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|2025-04-28 17:07:05] 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|2025-04-28 17:07:05] 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|2025-04-28 17:07:05] 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|2025-04-28 17:07:05] tokenization_utils_base.py:2034 >> loading file added_tokens.json from cache at None

[INFO|2025-04-28 17:07:05] tokenization_utils_base.py:2034 >> loading file special_tokens_map.json from cache at None

[INFO|2025-04-28 17:07:05] 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|2025-04-28 17:07:05] tokenization_utils_base.py:2034 >> loading file chat_template.jinja from cache at None

[INFO|2025-04-28 17:07:06] 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|2025-04-28 17:07:06] logging.py:157 >> Add <|im_end|> to stop words.

[INFO|2025-04-28 17:07:06] logging.py:157 >> Loading dataset Codes_query_filtered_330k_ns.json...

[INFO|2025-04-28 17:08:07] 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|2025-04-28 17:08:07] 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|2025-04-28 17:08:07] logging.py:162 >> Input length is smaller than max length. Consider increase input length.

[INFO|2025-04-28 17:08:07] logging.py:157 >> Using llama3 scaling strategy and setting scaling factor to 1.0.

[INFO|2025-04-28 17:08:07] logging.py:157 >> Using block diagonal attention for sequence packing without cross-attention.

[INFO|2025-04-28 17:08:07] logging.py:157 >> Liger kernel has been applied to the model.

[INFO|2025-04-28 17:08:07] 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|2025-04-28 17:08:07] modeling_utils.py:1582 >> Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16.

[INFO|2025-04-28 17:08:07] configuration_utils.py:1140 >> Generate config GenerationConfig {
  "bos_token_id": 151643,
  "eos_token_id": 151645
}


[INFO|2025-04-28 17:08:12] modeling_utils.py:4888 >> All model checkpoint weights were used when initializing Qwen2ForCausalLM.


[INFO|2025-04-28 17:08:12] 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|2025-04-28 17:08:12] 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|2025-04-28 17:08:12] 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|2025-04-28 17:08:13] logging.py:157 >> Gradient checkpointing enabled.

[INFO|2025-04-28 17:08:13] logging.py:157 >> Using torch SDPA for faster training and inference.

[INFO|2025-04-28 17:08:13] logging.py:157 >> Upcasting trainable params to float32.

[INFO|2025-04-28 17:08:13] logging.py:157 >> Fine-tuning method: Freeze

[INFO|2025-04-28 17:08:13] logging.py:157 >> Set trainable layers: .13.,.27.

[INFO|2025-04-28 17:08:13] logging.py:157 >> trainable params: 466,115,584 || all params: 7,615,616,512 || trainable%: 6.1205

[INFO|2025-04-28 17:08:13] trainer.py:741 >> Using auto half precision backend

[INFO|2025-04-28 17:08:13] logging.py:157 >> Found linear modules: k_proj,up_proj,gate_proj,o_proj,down_proj,q_proj,v_proj

[INFO|2025-04-28 17:08:13] 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|2025-04-28 17:08:14] trainer.py:2369 >> ***** Running training *****

[INFO|2025-04-28 17:08:14] trainer.py:2370 >>   Num examples = 56,766

[INFO|2025-04-28 17:08:14] trainer.py:2371 >>   Num Epochs = 1

[INFO|2025-04-28 17:08:14] trainer.py:2372 >>   Instantaneous batch size per device = 16

[INFO|2025-04-28 17:08:14] trainer.py:2375 >>   Total train batch size (w. parallel, distributed & accumulation) = 512

[INFO|2025-04-28 17:08:14] trainer.py:2376 >>   Gradient Accumulation steps = 8

[INFO|2025-04-28 17:08:14] trainer.py:2377 >>   Total optimization steps = 110

[INFO|2025-04-28 17:08:14] trainer.py:2378 >>   Number of trainable parameters = 466,115,584

[INFO|2025-04-28 17:11:00] logging.py:157 >> {'loss': 0.8968, 'learning_rate': 4.9990e-05, 'epoch': 0.01, 'throughput': 12763.83}

[INFO|2025-04-28 17:13:36] logging.py:157 >> {'loss': 0.8202, 'learning_rate': 4.9959e-05, 'epoch': 0.02, 'throughput': 13091.94}

[INFO|2025-04-28 17:16:11] logging.py:157 >> {'loss': 0.7783, 'learning_rate': 4.9908e-05, 'epoch': 0.03, 'throughput': 13222.77}

[INFO|2025-04-28 17:18:47] logging.py:157 >> {'loss': 0.7642, 'learning_rate': 4.9837e-05, 'epoch': 0.04, 'throughput': 13287.21}

[INFO|2025-04-28 17:21:22] logging.py:157 >> {'loss': 0.7475, 'learning_rate': 4.9746e-05, 'epoch': 0.05, 'throughput': 13330.27}

[INFO|2025-04-28 17:23:57] logging.py:157 >> {'loss': 0.7319, 'learning_rate': 4.9634e-05, 'epoch': 0.05, 'throughput': 13369.02}

[INFO|2025-04-28 17:26:31] logging.py:157 >> {'loss': 0.7125, 'learning_rate': 4.9502e-05, 'epoch': 0.06, 'throughput': 13397.69}

[INFO|2025-04-28 17:29:06] logging.py:157 >> {'loss': 0.7117, 'learning_rate': 4.9350e-05, 'epoch': 0.07, 'throughput': 13417.34}

[INFO|2025-04-28 17:31:41] logging.py:157 >> {'loss': 0.7240, 'learning_rate': 4.9179e-05, 'epoch': 0.08, 'throughput': 13429.19}

[INFO|2025-04-28 17:34:16] logging.py:157 >> {'loss': 0.7144, 'learning_rate': 4.8987e-05, 'epoch': 0.09, 'throughput': 13440.96}

[INFO|2025-04-28 17:36:51] logging.py:157 >> {'loss': 0.6970, 'learning_rate': 4.8776e-05, 'epoch': 0.10, 'throughput': 13448.05}

[INFO|2025-04-28 17:39:26] logging.py:157 >> {'loss': 0.7117, 'learning_rate': 4.8546e-05, 'epoch': 0.11, 'throughput': 13452.88}

[INFO|2025-04-28 17:42:01] logging.py:157 >> {'loss': 0.6996, 'learning_rate': 4.8297e-05, 'epoch': 0.12, 'throughput': 13458.10}

[INFO|2025-04-28 17:44:36] logging.py:157 >> {'loss': 0.7172, 'learning_rate': 4.8028e-05, 'epoch': 0.13, 'throughput': 13461.87}

[INFO|2025-04-28 17:47:12] logging.py:157 >> {'loss': 0.7250, 'learning_rate': 4.7741e-05, 'epoch': 0.14, 'throughput': 13465.50}

[INFO|2025-04-28 17:49:47] logging.py:157 >> {'loss': 0.7073, 'learning_rate': 4.7435e-05, 'epoch': 0.14, 'throughput': 13469.44}

[INFO|2025-04-28 17:52:22] logging.py:157 >> {'loss': 0.7021, 'learning_rate': 4.7111e-05, 'epoch': 0.15, 'throughput': 13473.17}

[INFO|2025-04-28 17:54:57] logging.py:157 >> {'loss': 0.6894, 'learning_rate': 4.6769e-05, 'epoch': 0.16, 'throughput': 13473.99}

[INFO|2025-04-28 17:57:33] logging.py:157 >> {'loss': 0.7031, 'learning_rate': 4.6409e-05, 'epoch': 0.17, 'throughput': 13471.10}

[INFO|2025-04-28 18:00:08] logging.py:157 >> {'loss': 0.6629, 'learning_rate': 4.6031e-05, 'epoch': 0.18, 'throughput': 13474.21}

[INFO|2025-04-28 18:02:43] logging.py:157 >> {'loss': 0.6774, 'learning_rate': 4.5637e-05, 'epoch': 0.19, 'throughput': 13476.95}

[INFO|2025-04-28 18:05:18] logging.py:157 >> {'loss': 0.7041, 'learning_rate': 4.5225e-05, 'epoch': 0.20, 'throughput': 13481.25}

[INFO|2025-04-28 18:07:52] logging.py:157 >> {'loss': 0.6894, 'learning_rate': 4.4798e-05, 'epoch': 0.21, 'throughput': 13484.61}

[INFO|2025-04-28 18:10:27] logging.py:157 >> {'loss': 0.6730, 'learning_rate': 4.4354e-05, 'epoch': 0.22, 'throughput': 13488.41}

[INFO|2025-04-28 18:13:02] logging.py:157 >> {'loss': 0.6691, 'learning_rate': 4.3894e-05, 'epoch': 0.23, 'throughput': 13490.63}

[INFO|2025-04-28 18:15:36] logging.py:157 >> {'loss': 0.6794, 'learning_rate': 4.3419e-05, 'epoch': 0.23, 'throughput': 13493.84}

[INFO|2025-04-28 18:18:12] logging.py:157 >> {'loss': 0.6946, 'learning_rate': 4.2928e-05, 'epoch': 0.24, 'throughput': 13493.72}

[INFO|2025-04-28 18:20:47] logging.py:157 >> {'loss': 0.6939, 'learning_rate': 4.2423e-05, 'epoch': 0.25, 'throughput': 13495.64}

[INFO|2025-04-28 18:23:21] logging.py:157 >> {'loss': 0.6759, 'learning_rate': 4.1904e-05, 'epoch': 0.26, 'throughput': 13497.79}

[INFO|2025-04-28 18:26:01] logging.py:157 >> {'loss': 0.6856, 'learning_rate': 4.1372e-05, 'epoch': 0.27, 'throughput': 13485.48}

[INFO|2025-04-28 18:28:39] logging.py:157 >> {'loss': 0.6753, 'learning_rate': 4.0825e-05, 'epoch': 0.28, 'throughput': 13476.76}

[INFO|2025-04-28 18:31:18] logging.py:157 >> {'loss': 0.6769, 'learning_rate': 4.0266e-05, 'epoch': 0.29, 'throughput': 13467.64}

[INFO|2025-04-28 18:33:57] logging.py:157 >> {'loss': 0.6664, 'learning_rate': 3.9695e-05, 'epoch': 0.30, 'throughput': 13460.91}

[INFO|2025-04-28 18:36:34] logging.py:157 >> {'loss': 0.6383, 'learning_rate': 3.9111e-05, 'epoch': 0.31, 'throughput': 13456.16}

[INFO|2025-04-28 18:39:13] logging.py:157 >> {'loss': 0.6726, 'learning_rate': 3.8516e-05, 'epoch': 0.32, 'throughput': 13449.18}

[INFO|2025-04-28 18:41:51] logging.py:157 >> {'loss': 0.6612, 'learning_rate': 3.7910e-05, 'epoch': 0.32, 'throughput': 13443.43}

[INFO|2025-04-28 18:44:30] logging.py:157 >> {'loss': 0.6763, 'learning_rate': 3.7293e-05, 'epoch': 0.33, 'throughput': 13438.40}

[INFO|2025-04-28 18:47:08] logging.py:157 >> {'loss': 0.6533, 'learning_rate': 3.6667e-05, 'epoch': 0.34, 'throughput': 13433.21}

[INFO|2025-04-28 18:49:46] logging.py:157 >> {'loss': 0.6598, 'learning_rate': 3.6031e-05, 'epoch': 0.35, 'throughput': 13429.21}

[INFO|2025-04-28 18:52:24] logging.py:157 >> {'loss': 0.6801, 'learning_rate': 3.5385e-05, 'epoch': 0.36, 'throughput': 13424.38}

[INFO|2025-04-28 18:55:03] logging.py:157 >> {'loss': 0.6668, 'learning_rate': 3.4732e-05, 'epoch': 0.37, 'throughput': 13420.03}

[INFO|2025-04-28 18:57:41] logging.py:157 >> {'loss': 0.6756, 'learning_rate': 3.4070e-05, 'epoch': 0.38, 'throughput': 13414.75}

[INFO|2025-04-28 19:00:19] logging.py:157 >> {'loss': 0.6589, 'learning_rate': 3.3401e-05, 'epoch': 0.39, 'throughput': 13411.37}

[INFO|2025-04-28 19:02:57] logging.py:157 >> {'loss': 0.6435, 'learning_rate': 3.2725e-05, 'epoch': 0.40, 'throughput': 13408.63}

[INFO|2025-04-28 19:05:35] logging.py:157 >> {'loss': 0.6602, 'learning_rate': 3.2043e-05, 'epoch': 0.41, 'throughput': 13405.22}

[INFO|2025-04-28 19:08:14] logging.py:157 >> {'loss': 0.6480, 'learning_rate': 3.1355e-05, 'epoch': 0.41, 'throughput': 13401.83}

[INFO|2025-04-28 19:10:52] logging.py:157 >> {'loss': 0.6493, 'learning_rate': 3.0662e-05, 'epoch': 0.42, 'throughput': 13398.84}

[INFO|2025-04-28 19:13:30] logging.py:157 >> {'loss': 0.6701, 'learning_rate': 2.9965e-05, 'epoch': 0.43, 'throughput': 13395.76}

[INFO|2025-04-28 19:16:08] logging.py:157 >> {'loss': 0.6674, 'learning_rate': 2.9263e-05, 'epoch': 0.44, 'throughput': 13392.56}

[INFO|2025-04-28 19:18:47] logging.py:157 >> {'loss': 0.6635, 'learning_rate': 2.8558e-05, 'epoch': 0.45, 'throughput': 13389.71}

[INFO|2025-04-28 19:21:25] logging.py:157 >> {'loss': 0.6598, 'learning_rate': 2.7850e-05, 'epoch': 0.46, 'throughput': 13386.60}

[INFO|2025-04-28 19:24:04] logging.py:157 >> {'loss': 0.6732, 'learning_rate': 2.7139e-05, 'epoch': 0.47, 'throughput': 13383.73}

[INFO|2025-04-28 19:26:42] logging.py:157 >> {'loss': 0.6546, 'learning_rate': 2.6427e-05, 'epoch': 0.48, 'throughput': 13380.78}

[INFO|2025-04-28 19:29:20] logging.py:157 >> {'loss': 0.6705, 'learning_rate': 2.5714e-05, 'epoch': 0.49, 'throughput': 13379.10}

[INFO|2025-04-28 19:31:58] logging.py:157 >> {'loss': 0.6463, 'learning_rate': 2.5000e-05, 'epoch': 0.50, 'throughput': 13376.64}

[INFO|2025-04-28 19:34:36] logging.py:157 >> {'loss': 0.6505, 'learning_rate': 2.4286e-05, 'epoch': 0.51, 'throughput': 13375.59}

[INFO|2025-04-28 19:37:14] logging.py:157 >> {'loss': 0.6524, 'learning_rate': 2.3573e-05, 'epoch': 0.51, 'throughput': 13373.95}

[INFO|2025-04-28 19:39:52] logging.py:157 >> {'loss': 0.6559, 'learning_rate': 2.2861e-05, 'epoch': 0.52, 'throughput': 13371.57}

[INFO|2025-04-28 19:42:31] logging.py:157 >> {'loss': 0.6511, 'learning_rate': 2.2150e-05, 'epoch': 0.53, 'throughput': 13368.89}

[INFO|2025-04-28 19:45:09] logging.py:157 >> {'loss': 0.6457, 'learning_rate': 2.1442e-05, 'epoch': 0.54, 'throughput': 13366.30}

[INFO|2025-04-28 19:47:48] logging.py:157 >> {'loss': 0.6492, 'learning_rate': 2.0737e-05, 'epoch': 0.55, 'throughput': 13364.48}

[INFO|2025-04-28 19:50:25] logging.py:157 >> {'loss': 0.6539, 'learning_rate': 2.0035e-05, 'epoch': 0.56, 'throughput': 13363.30}

[INFO|2025-04-28 19:53:03] logging.py:157 >> {'loss': 0.6417, 'learning_rate': 1.9338e-05, 'epoch': 0.57, 'throughput': 13362.41}

[INFO|2025-04-28 19:55:41] logging.py:157 >> {'loss': 0.6663, 'learning_rate': 1.8645e-05, 'epoch': 0.58, 'throughput': 13360.92}

[INFO|2025-04-28 19:58:19] logging.py:157 >> {'loss': 0.6572, 'learning_rate': 1.7957e-05, 'epoch': 0.59, 'throughput': 13360.07}

[INFO|2025-04-28 20:00:57] logging.py:157 >> {'loss': 0.6825, 'learning_rate': 1.7275e-05, 'epoch': 0.60, 'throughput': 13358.01}

[INFO|2025-04-28 20:03:35] logging.py:157 >> {'loss': 0.6509, 'learning_rate': 1.6599e-05, 'epoch': 0.60, 'throughput': 13356.51}

[INFO|2025-04-28 20:06:13] logging.py:157 >> {'loss': 0.6619, 'learning_rate': 1.5930e-05, 'epoch': 0.61, 'throughput': 13355.52}

[INFO|2025-04-28 20:08:52] logging.py:157 >> {'loss': 0.6759, 'learning_rate': 1.5268e-05, 'epoch': 0.62, 'throughput': 13352.79}

[INFO|2025-04-28 20:11:30] logging.py:157 >> {'loss': 0.6568, 'learning_rate': 1.4615e-05, 'epoch': 0.63, 'throughput': 13352.37}

[INFO|2025-04-28 20:14:08] logging.py:157 >> {'loss': 0.6472, 'learning_rate': 1.3969e-05, 'epoch': 0.64, 'throughput': 13351.50}

[INFO|2025-04-28 20:16:47] logging.py:157 >> {'loss': 0.6473, 'learning_rate': 1.3333e-05, 'epoch': 0.65, 'throughput': 13349.03}

[INFO|2025-04-28 20:19:25] logging.py:157 >> {'loss': 0.6485, 'learning_rate': 1.2707e-05, 'epoch': 0.66, 'throughput': 13347.43}

[INFO|2025-04-28 20:22:04] logging.py:157 >> {'loss': 0.6544, 'learning_rate': 1.2090e-05, 'epoch': 0.67, 'throughput': 13345.97}

[INFO|2025-04-28 20:24:42] logging.py:157 >> {'loss': 0.6763, 'learning_rate': 1.1484e-05, 'epoch': 0.68, 'throughput': 13344.93}

[INFO|2025-04-28 20:27:21] logging.py:157 >> {'loss': 0.6406, 'learning_rate': 1.0889e-05, 'epoch': 0.69, 'throughput': 13342.47}

[INFO|2025-04-28 20:30:00] logging.py:157 >> {'loss': 0.6502, 'learning_rate': 1.0305e-05, 'epoch': 0.69, 'throughput': 13341.02}

[INFO|2025-04-28 20:32:38] logging.py:157 >> {'loss': 0.6495, 'learning_rate': 9.7338e-06, 'epoch': 0.70, 'throughput': 13339.92}

[INFO|2025-04-28 20:35:16] logging.py:157 >> {'loss': 0.6469, 'learning_rate': 9.1747e-06, 'epoch': 0.71, 'throughput': 13339.19}

[INFO|2025-04-28 20:37:54] logging.py:157 >> {'loss': 0.6642, 'learning_rate': 8.6285e-06, 'epoch': 0.72, 'throughput': 13337.79}

[INFO|2025-04-28 20:40:32] logging.py:157 >> {'loss': 0.6598, 'learning_rate': 8.0956e-06, 'epoch': 0.73, 'throughput': 13336.77}

[INFO|2025-04-28 20:43:11] logging.py:157 >> {'loss': 0.6461, 'learning_rate': 7.5766e-06, 'epoch': 0.74, 'throughput': 13335.06}

[INFO|2025-04-28 20:45:50] logging.py:157 >> {'loss': 0.6706, 'learning_rate': 7.0717e-06, 'epoch': 0.75, 'throughput': 13333.27}

[INFO|2025-04-28 20:48:28] logging.py:157 >> {'loss': 0.6768, 'learning_rate': 6.5815e-06, 'epoch': 0.76, 'throughput': 13332.46}

[INFO|2025-04-28 20:51:06] logging.py:157 >> {'loss': 0.6446, 'learning_rate': 6.1063e-06, 'epoch': 0.77, 'throughput': 13331.91}

[INFO|2025-04-28 20:53:44] logging.py:157 >> {'loss': 0.6393, 'learning_rate': 5.6465e-06, 'epoch': 0.78, 'throughput': 13331.08}

[INFO|2025-04-28 20:56:22] logging.py:157 >> {'loss': 0.6585, 'learning_rate': 5.2024e-06, 'epoch': 0.78, 'throughput': 13330.39}

[INFO|2025-04-28 20:59:01] logging.py:157 >> {'loss': 0.6393, 'learning_rate': 4.7746e-06, 'epoch': 0.79, 'throughput': 13329.42}

[INFO|2025-04-28 21:01:38] logging.py:157 >> {'loss': 0.6613, 'learning_rate': 4.3632e-06, 'epoch': 0.80, 'throughput': 13329.14}

[INFO|2025-04-28 21:04:18] logging.py:157 >> {'loss': 0.6631, 'learning_rate': 3.9687e-06, 'epoch': 0.81, 'throughput': 13327.42}

[INFO|2025-04-28 21:06:56] logging.py:157 >> {'loss': 0.6356, 'learning_rate': 3.5913e-06, 'epoch': 0.82, 'throughput': 13326.22}

[INFO|2025-04-28 21:09:35] logging.py:157 >> {'loss': 0.6475, 'learning_rate': 3.2313e-06, 'epoch': 0.83, 'throughput': 13325.12}

[INFO|2025-04-28 21:12:13] logging.py:157 >> {'loss': 0.6387, 'learning_rate': 2.8892e-06, 'epoch': 0.84, 'throughput': 13324.10}

[INFO|2025-04-28 21:14:52] logging.py:157 >> {'loss': 0.6533, 'learning_rate': 2.5650e-06, 'epoch': 0.85, 'throughput': 13323.33}

[INFO|2025-04-28 21:17:30] logging.py:157 >> {'loss': 0.6606, 'learning_rate': 2.2592e-06, 'epoch': 0.86, 'throughput': 13322.67}

[INFO|2025-04-28 21:20:08] logging.py:157 >> {'loss': 0.6667, 'learning_rate': 1.9719e-06, 'epoch': 0.87, 'throughput': 13322.10}

[INFO|2025-04-28 21:22:46] logging.py:157 >> {'loss': 0.6599, 'learning_rate': 1.7034e-06, 'epoch': 0.87, 'throughput': 13321.16}

[INFO|2025-04-28 21:25:25] logging.py:157 >> {'loss': 0.6499, 'learning_rate': 1.4539e-06, 'epoch': 0.88, 'throughput': 13319.93}

[INFO|2025-04-28 21:28:03] logging.py:157 >> {'loss': 0.6490, 'learning_rate': 1.2236e-06, 'epoch': 0.89, 'throughput': 13319.17}

[INFO|2025-04-28 21:30:42] logging.py:157 >> {'loss': 0.6551, 'learning_rate': 1.0127e-06, 'epoch': 0.90, 'throughput': 13318.33}

[INFO|2025-04-28 21:33:20] logging.py:157 >> {'loss': 0.6680, 'learning_rate': 8.2133e-07, 'epoch': 0.91, 'throughput': 13317.49}

[INFO|2025-04-28 21:35:59] logging.py:157 >> {'loss': 0.6505, 'learning_rate': 6.4970e-07, 'epoch': 0.92, 'throughput': 13316.86}

[INFO|2025-04-28 21:38:36] logging.py:157 >> {'loss': 0.6504, 'learning_rate': 4.9794e-07, 'epoch': 0.93, 'throughput': 13317.27}

[INFO|2025-04-28 21:41:13] logging.py:157 >> {'loss': 0.6504, 'learning_rate': 3.6615e-07, 'epoch': 0.94, 'throughput': 13317.01}

[INFO|2025-04-28 21:43:52] logging.py:157 >> {'loss': 0.6491, 'learning_rate': 2.5446e-07, 'epoch': 0.95, 'throughput': 13316.43}

[INFO|2025-04-28 21:46:29] logging.py:157 >> {'loss': 0.6449, 'learning_rate': 1.6296e-07, 'epoch': 0.96, 'throughput': 13316.07}

[INFO|2025-04-28 21:49:07] logging.py:157 >> {'loss': 0.6729, 'learning_rate': 9.1707e-08, 'epoch': 0.97, 'throughput': 13315.82}

[INFO|2025-04-28 21:51:45] logging.py:157 >> {'loss': 0.6308, 'learning_rate': 4.0772e-08, 'epoch': 0.97, 'throughput': 13315.41}

[INFO|2025-04-28 21:54:24] logging.py:157 >> {'loss': 0.6695, 'learning_rate': 1.0195e-08, 'epoch': 0.98, 'throughput': 13314.76}

[INFO|2025-04-28 21:57:02] logging.py:157 >> {'loss': 0.6467, 'learning_rate': 0.0000e+00, 'epoch': 0.99, 'throughput': 13314.02}

[INFO|2025-04-28 21:57:02] trainer.py:3910 >> Saving model checkpoint to saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/checkpoint-110

[INFO|2025-04-28 21:57:02] configuration_utils.py:420 >> Configuration saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/checkpoint-110/config.json

[INFO|2025-04-28 21:57:02] configuration_utils.py:909 >> Configuration saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/checkpoint-110/generation_config.json

[INFO|2025-04-28 21:57:27] 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_nsx/checkpoint-110/model.safetensors.index.json.

[INFO|2025-04-28 21:57:27] tokenization_utils_base.py:2491 >> tokenizer config file saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/checkpoint-110/tokenizer_config.json

[INFO|2025-04-28 21:57:27] tokenization_utils_base.py:2500 >> Special tokens file saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/checkpoint-110/special_tokens_map.json

[INFO|2025-04-28 21:57:27] trainer.py:2643 >> 

Training completed. Do not forget to share your model on huggingface.co/models =)



[INFO|2025-04-28 21:57:27] trainer.py:3910 >> Saving model checkpoint to saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx

[INFO|2025-04-28 21:57:27] configuration_utils.py:420 >> Configuration saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/config.json

[INFO|2025-04-28 21:57:27] configuration_utils.py:909 >> Configuration saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/generation_config.json

[INFO|2025-04-28 21:57:52] 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_nsx/model.safetensors.index.json.

[INFO|2025-04-28 21:57:52] tokenization_utils_base.py:2491 >> tokenizer config file saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/tokenizer_config.json

[INFO|2025-04-28 21:57:52] tokenization_utils_base.py:2500 >> Special tokens file saved in saves/Qwen2.5-Coder-7B-Instruct/freeze/qwen_nsx/special_tokens_map.json

[WARNING|2025-04-28 21:57:52] logging.py:162 >> No metric eval_loss to plot.

[WARNING|2025-04-28 21:57:52] logging.py:162 >> No metric eval_accuracy to plot.

[INFO|2025-04-28 21:57:52] modelcard.py:449 >> Dropping the following result as it does not have all the necessary fields:
{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}