running_log.txt

[INFO|2025-05-12 09:58:27] tokenization_utils_base.py:2034 >> loading file tokenizer.model from cache at None

[INFO|2025-05-12 09:58:27] tokenization_utils_base.py:2034 >> loading file tokenizer.json from cache at /home/kiho/.cache/huggingface/hub/models--deepseek-ai--deepseek-coder-7b-instruct-v1.5/snapshots/2a050a4c59d687a85324d32e147517992117ed30/tokenizer.json

[INFO|2025-05-12 09:58:27] tokenization_utils_base.py:2034 >> loading file added_tokens.json from cache at None

[INFO|2025-05-12 09:58:27] tokenization_utils_base.py:2034 >> loading file special_tokens_map.json from cache at None

[INFO|2025-05-12 09:58:27] tokenization_utils_base.py:2034 >> loading file tokenizer_config.json from cache at /home/kiho/.cache/huggingface/hub/models--deepseek-ai--deepseek-coder-7b-instruct-v1.5/snapshots/2a050a4c59d687a85324d32e147517992117ed30/tokenizer_config.json

[INFO|2025-05-12 09:58:27] tokenization_utils_base.py:2034 >> loading file chat_template.jinja from cache at None

[INFO|2025-05-12 09:58:28] 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|2025-05-12 09:58:29] configuration_utils.py:696 >> loading configuration file config.json from cache at /home/kiho/.cache/huggingface/hub/models--deepseek-ai--deepseek-coder-7b-instruct-v1.5/snapshots/2a050a4c59d687a85324d32e147517992117ed30/config.json

[INFO|2025-05-12 09:58:29] configuration_utils.py:768 >> Model config LlamaConfig {
  "_name_or_path": "deepseek-ai/deepseek-coder-7b-instruct-v1.5",
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 100000,
  "eos_token_id": 100015,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 11008,
  "max_position_embeddings": 4096,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 30,
  "num_key_value_heads": 32,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.48.2",
  "use_cache": true,
  "vocab_size": 102400
}


[INFO|2025-05-12 09:58:29] tokenization_utils_base.py:2034 >> loading file tokenizer.model from cache at None

[INFO|2025-05-12 09:58:29] tokenization_utils_base.py:2034 >> loading file tokenizer.json from cache at /home/kiho/.cache/huggingface/hub/models--deepseek-ai--deepseek-coder-7b-instruct-v1.5/snapshots/2a050a4c59d687a85324d32e147517992117ed30/tokenizer.json

[INFO|2025-05-12 09:58:29] tokenization_utils_base.py:2034 >> loading file added_tokens.json from cache at None

[INFO|2025-05-12 09:58:29] tokenization_utils_base.py:2034 >> loading file special_tokens_map.json from cache at None

[INFO|2025-05-12 09:58:29] tokenization_utils_base.py:2034 >> loading file tokenizer_config.json from cache at /home/kiho/.cache/huggingface/hub/models--deepseek-ai--deepseek-coder-7b-instruct-v1.5/snapshots/2a050a4c59d687a85324d32e147517992117ed30/tokenizer_config.json

[INFO|2025-05-12 09:58:29] tokenization_utils_base.py:2034 >> loading file chat_template.jinja from cache at None

[INFO|2025-05-12 09:58:30] 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|2025-05-12 09:58:30] logging.py:157 >> Loading dataset Codes3_query_filtered_330k_nlx.json...

[INFO|2025-05-12 09:59:23] configuration_utils.py:696 >> loading configuration file config.json from cache at /home/kiho/.cache/huggingface/hub/models--deepseek-ai--deepseek-coder-7b-instruct-v1.5/snapshots/2a050a4c59d687a85324d32e147517992117ed30/config.json

[INFO|2025-05-12 09:59:23] configuration_utils.py:768 >> Model config LlamaConfig {
  "_name_or_path": "deepseek-ai/deepseek-coder-7b-instruct-v1.5",
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 100000,
  "eos_token_id": 100015,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 11008,
  "max_position_embeddings": 4096,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 30,
  "num_key_value_heads": 32,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.48.2",
  "use_cache": true,
  "vocab_size": 102400
}


[WARNING|2025-05-12 09:59:23] logging.py:162 >> Input length is smaller than max length. Consider increase input length.

[INFO|2025-05-12 09:59:23] logging.py:157 >> Using llama3 scaling strategy and setting scaling factor to 1.0.

[INFO|2025-05-12 09:59:23] logging.py:157 >> Using block diagonal attention for sequence packing without cross-attention.

[INFO|2025-05-12 09:59:23] logging.py:157 >> Liger kernel has been applied to the model.

[INFO|2025-05-12 09:59:24] modeling_utils.py:3904 >> loading weights file model.safetensors from cache at /home/kiho/.cache/huggingface/hub/models--deepseek-ai--deepseek-coder-7b-instruct-v1.5/snapshots/2a050a4c59d687a85324d32e147517992117ed30/model.safetensors.index.json

[INFO|2025-05-12 10:01:40] modeling_utils.py:1582 >> Instantiating LlamaForCausalLM model under default dtype torch.bfloat16.

[INFO|2025-05-12 10:01:40] configuration_utils.py:1140 >> Generate config GenerationConfig {
  "bos_token_id": 100000,
  "eos_token_id": 100015
}


[INFO|2025-05-12 10:01:45] modeling_utils.py:4888 >> All model checkpoint weights were used when initializing LlamaForCausalLM.


[INFO|2025-05-12 10:01:45] modeling_utils.py:4896 >> All the weights of LlamaForCausalLM were initialized from the model checkpoint at deepseek-ai/deepseek-coder-7b-instruct-v1.5.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.

[INFO|2025-05-12 10:01:46] configuration_utils.py:1095 >> loading configuration file generation_config.json from cache at /home/kiho/.cache/huggingface/hub/models--deepseek-ai--deepseek-coder-7b-instruct-v1.5/snapshots/2a050a4c59d687a85324d32e147517992117ed30/generation_config.json

[INFO|2025-05-12 10:01:46] configuration_utils.py:1140 >> Generate config GenerationConfig {
  "bos_token_id": 100000,
  "eos_token_id": 100015
}


[INFO|2025-05-12 10:01:46] logging.py:157 >> Gradient checkpointing enabled.

[INFO|2025-05-12 10:01:46] logging.py:157 >> Using torch SDPA for faster training and inference.

[INFO|2025-05-12 10:01:46] logging.py:157 >> Upcasting trainable params to float32.

[INFO|2025-05-12 10:01:46] logging.py:157 >> Fine-tuning method: Freeze

[INFO|2025-05-12 10:01:46] logging.py:157 >> Set trainable layers: .14.,.29.

[INFO|2025-05-12 10:01:46] logging.py:157 >> trainable params: 404,766,720 || all params: 6,910,365,696 || trainable%: 5.8574

[INFO|2025-05-12 10:01:46] trainer.py:741 >> Using auto half precision backend

[INFO|2025-05-12 10:01:47] logging.py:157 >> Found linear modules: k_proj,v_proj,o_proj,down_proj,q_proj,up_proj,gate_proj

[INFO|2025-05-12 10:01:47] 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|2025-05-12 10:01:47] trainer.py:2369 >> ***** Running training *****

[INFO|2025-05-12 10:01:47] trainer.py:2370 >>   Num examples = 35,434

[INFO|2025-05-12 10:01:47] trainer.py:2371 >>   Num Epochs = 1

[INFO|2025-05-12 10:01:47] trainer.py:2372 >>   Instantaneous batch size per device = 16

[INFO|2025-05-12 10:01:47] trainer.py:2375 >>   Total train batch size (w. parallel, distributed & accumulation) = 512

[INFO|2025-05-12 10:01:47] trainer.py:2376 >>   Gradient Accumulation steps = 8

[INFO|2025-05-12 10:01:47] trainer.py:2377 >>   Total optimization steps = 69

[INFO|2025-05-12 10:01:47] trainer.py:2378 >>   Number of trainable parameters = 404,766,720

[INFO|2025-05-12 10:04:36] logging.py:157 >> {'loss': 0.8402, 'learning_rate': 4.9974e-05, 'epoch': 0.01, 'throughput': 12482.16}

[INFO|2025-05-12 10:07:14] logging.py:157 >> {'loss': 0.7686, 'learning_rate': 4.9896e-05, 'epoch': 0.03, 'throughput': 12852.44}

[INFO|2025-05-12 10:09:52] logging.py:157 >> {'loss': 0.7563, 'learning_rate': 4.9767e-05, 'epoch': 0.04, 'throughput': 12981.50}

[INFO|2025-05-12 10:12:30] logging.py:157 >> {'loss': 0.7139, 'learning_rate': 4.9587e-05, 'epoch': 0.06, 'throughput': 13054.16}

[INFO|2025-05-12 10:15:08] logging.py:157 >> {'loss': 0.6793, 'learning_rate': 4.9355e-05, 'epoch': 0.07, 'throughput': 13098.17}

[INFO|2025-05-12 10:17:46] logging.py:157 >> {'loss': 0.6631, 'learning_rate': 4.9073e-05, 'epoch': 0.09, 'throughput': 13124.77}

[INFO|2025-05-12 10:20:25] logging.py:157 >> {'loss': 0.6400, 'learning_rate': 4.8741e-05, 'epoch': 0.10, 'throughput': 13144.27}

[INFO|2025-05-12 10:23:03] logging.py:157 >> {'loss': 0.6145, 'learning_rate': 4.8360e-05, 'epoch': 0.12, 'throughput': 13158.63}

[INFO|2025-05-12 10:25:41] logging.py:157 >> {'loss': 0.5997, 'learning_rate': 4.7930e-05, 'epoch': 0.13, 'throughput': 13173.26}

[INFO|2025-05-12 10:28:19] logging.py:157 >> {'loss': 0.5896, 'learning_rate': 4.7453e-05, 'epoch': 0.14, 'throughput': 13182.64}

[INFO|2025-05-12 10:30:56] logging.py:157 >> {'loss': 0.6025, 'learning_rate': 4.6930e-05, 'epoch': 0.16, 'throughput': 13191.88}

[INFO|2025-05-12 10:33:35] logging.py:157 >> {'loss': 0.5644, 'learning_rate': 4.6360e-05, 'epoch': 0.17, 'throughput': 13197.39}

[INFO|2025-05-12 10:36:12] logging.py:157 >> {'loss': 0.5558, 'learning_rate': 4.5747e-05, 'epoch': 0.19, 'throughput': 13204.31}

[INFO|2025-05-12 10:38:51] logging.py:157 >> {'loss': 0.5705, 'learning_rate': 4.5091e-05, 'epoch': 0.20, 'throughput': 13208.35}

[INFO|2025-05-12 10:41:29] logging.py:157 >> {'loss': 0.5694, 'learning_rate': 4.4393e-05, 'epoch': 0.22, 'throughput': 13212.19}

[INFO|2025-05-12 10:44:06] logging.py:157 >> {'loss': 0.5590, 'learning_rate': 4.3655e-05, 'epoch': 0.23, 'throughput': 13217.11}

[INFO|2025-05-12 10:46:44] logging.py:157 >> {'loss': 0.5628, 'learning_rate': 4.2878e-05, 'epoch': 0.25, 'throughput': 13220.46}

[INFO|2025-05-12 10:49:23] logging.py:157 >> {'loss': 0.5553, 'learning_rate': 4.2064e-05, 'epoch': 0.26, 'throughput': 13223.11}

[INFO|2025-05-12 10:52:01] logging.py:157 >> {'loss': 0.5401, 'learning_rate': 4.1215e-05, 'epoch': 0.27, 'throughput': 13225.48}

[INFO|2025-05-12 10:54:39] logging.py:157 >> {'loss': 0.5750, 'learning_rate': 4.0332e-05, 'epoch': 0.29, 'throughput': 13226.89}

[INFO|2025-05-12 10:57:17] logging.py:157 >> {'loss': 0.5398, 'learning_rate': 3.9417e-05, 'epoch': 0.30, 'throughput': 13228.58}

[INFO|2025-05-12 10:59:55] logging.py:157 >> {'loss': 0.5389, 'learning_rate': 3.8472e-05, 'epoch': 0.32, 'throughput': 13231.10}

[INFO|2025-05-12 11:02:32] logging.py:157 >> {'loss': 0.5451, 'learning_rate': 3.7500e-05, 'epoch': 0.33, 'throughput': 13234.31}

[INFO|2025-05-12 11:05:11] logging.py:157 >> {'loss': 0.5558, 'learning_rate': 3.6502e-05, 'epoch': 0.35, 'throughput': 13235.05}

[INFO|2025-05-12 11:07:48] logging.py:157 >> {'loss': 0.5403, 'learning_rate': 3.5479e-05, 'epoch': 0.36, 'throughput': 13238.00}

[INFO|2025-05-12 11:10:26] logging.py:157 >> {'loss': 0.5405, 'learning_rate': 3.4435e-05, 'epoch': 0.38, 'throughput': 13240.74}

[INFO|2025-05-12 11:13:04] logging.py:157 >> {'loss': 0.5577, 'learning_rate': 3.3372e-05, 'epoch': 0.39, 'throughput': 13242.13}

[INFO|2025-05-12 11:15:42] logging.py:157 >> {'loss': 0.5345, 'learning_rate': 3.2291e-05, 'epoch': 0.40, 'throughput': 13243.46}

[INFO|2025-05-12 11:18:20] logging.py:157 >> {'loss': 0.5484, 'learning_rate': 3.1195e-05, 'epoch': 0.42, 'throughput': 13243.76}

[INFO|2025-05-12 11:20:58] logging.py:157 >> {'loss': 0.5485, 'learning_rate': 3.0086e-05, 'epoch': 0.43, 'throughput': 13245.01}

[INFO|2025-05-12 11:23:36] logging.py:157 >> {'loss': 0.5299, 'learning_rate': 2.8967e-05, 'epoch': 0.45, 'throughput': 13245.72}

[INFO|2025-05-12 11:26:14] logging.py:157 >> {'loss': 0.5390, 'learning_rate': 2.7840e-05, 'epoch': 0.46, 'throughput': 13246.45}

[INFO|2025-05-12 11:28:52] logging.py:157 >> {'loss': 0.5248, 'learning_rate': 2.6706e-05, 'epoch': 0.48, 'throughput': 13247.54}

[INFO|2025-05-12 11:31:30] logging.py:157 >> {'loss': 0.5342, 'learning_rate': 2.5569e-05, 'epoch': 0.49, 'throughput': 13248.46}

[INFO|2025-05-12 11:34:08] logging.py:157 >> {'loss': 0.5431, 'learning_rate': 2.4431e-05, 'epoch': 0.51, 'throughput': 13249.53}

[INFO|2025-05-12 11:36:45] logging.py:157 >> {'loss': 0.5471, 'learning_rate': 2.3294e-05, 'epoch': 0.52, 'throughput': 13250.51}

[INFO|2025-05-12 11:39:23] logging.py:157 >> {'loss': 0.5420, 'learning_rate': 2.2160e-05, 'epoch': 0.53, 'throughput': 13251.85}

[INFO|2025-05-12 11:42:01] logging.py:157 >> {'loss': 0.5420, 'learning_rate': 2.1033e-05, 'epoch': 0.55, 'throughput': 13253.57}

[INFO|2025-05-12 11:44:38] logging.py:157 >> {'loss': 0.5420, 'learning_rate': 1.9914e-05, 'epoch': 0.56, 'throughput': 13255.10}

[INFO|2025-05-12 11:47:16] logging.py:157 >> {'loss': 0.5425, 'learning_rate': 1.8805e-05, 'epoch': 0.58, 'throughput': 13256.36}

[INFO|2025-05-12 11:49:54] logging.py:157 >> {'loss': 0.5482, 'learning_rate': 1.7709e-05, 'epoch': 0.59, 'throughput': 13257.04}

[INFO|2025-05-12 11:52:31] logging.py:157 >> {'loss': 0.5598, 'learning_rate': 1.6628e-05, 'epoch': 0.61, 'throughput': 13258.73}

[INFO|2025-05-12 11:55:08] logging.py:157 >> {'loss': 0.5402, 'learning_rate': 1.5565e-05, 'epoch': 0.62, 'throughput': 13260.29}

[INFO|2025-05-12 11:57:46] logging.py:157 >> {'loss': 0.5413, 'learning_rate': 1.4521e-05, 'epoch': 0.64, 'throughput': 13261.50}

[INFO|2025-05-12 12:00:24] logging.py:157 >> {'loss': 0.5560, 'learning_rate': 1.3498e-05, 'epoch': 0.65, 'throughput': 13262.39}

[INFO|2025-05-12 12:03:01] logging.py:157 >> {'loss': 0.5341, 'learning_rate': 1.2500e-05, 'epoch': 0.66, 'throughput': 13263.23}

[INFO|2025-05-12 12:05:39] logging.py:157 >> {'loss': 0.5436, 'learning_rate': 1.1528e-05, 'epoch': 0.68, 'throughput': 13264.02}

[INFO|2025-05-12 12:08:16] logging.py:157 >> {'loss': 0.5393, 'learning_rate': 1.0583e-05, 'epoch': 0.69, 'throughput': 13264.84}

[INFO|2025-05-12 12:10:54] logging.py:157 >> {'loss': 0.5598, 'learning_rate': 9.6683e-06, 'epoch': 0.71, 'throughput': 13265.65}

[INFO|2025-05-12 12:13:32] logging.py:157 >> {'loss': 0.5329, 'learning_rate': 8.7854e-06, 'epoch': 0.72, 'throughput': 13265.87}

[INFO|2025-05-12 12:16:10] logging.py:157 >> {'loss': 0.5384, 'learning_rate': 7.9362e-06, 'epoch': 0.74, 'throughput': 13266.73}

[INFO|2025-05-12 12:18:47] logging.py:157 >> {'loss': 0.5447, 'learning_rate': 7.1223e-06, 'epoch': 0.75, 'throughput': 13267.67}

[INFO|2025-05-12 12:21:25] logging.py:157 >> {'loss': 0.5291, 'learning_rate': 6.3454e-06, 'epoch': 0.77, 'throughput': 13268.20}

[INFO|2025-05-12 12:24:02] logging.py:157 >> {'loss': 0.5258, 'learning_rate': 5.6072e-06, 'epoch': 0.78, 'throughput': 13269.48}

[INFO|2025-05-12 12:26:40] logging.py:157 >> {'loss': 0.5508, 'learning_rate': 4.9092e-06, 'epoch': 0.79, 'throughput': 13270.12}

[INFO|2025-05-12 12:29:18] logging.py:157 >> {'loss': 0.5439, 'learning_rate': 4.2529e-06, 'epoch': 0.81, 'throughput': 13270.24}

[INFO|2025-05-12 12:31:56] logging.py:157 >> {'loss': 0.5261, 'learning_rate': 3.6395e-06, 'epoch': 0.82, 'throughput': 13270.03}

[INFO|2025-05-12 12:34:34] logging.py:157 >> {'loss': 0.5518, 'learning_rate': 3.0704e-06, 'epoch': 0.84, 'throughput': 13270.36}

[INFO|2025-05-12 12:37:11] logging.py:157 >> {'loss': 0.5444, 'learning_rate': 2.5468e-06, 'epoch': 0.85, 'throughput': 13270.70}

[INFO|2025-05-12 12:39:49] logging.py:157 >> {'loss': 0.5469, 'learning_rate': 2.0697e-06, 'epoch': 0.87, 'throughput': 13270.96}

[INFO|2025-05-12 12:42:27] logging.py:157 >> {'loss': 0.5335, 'learning_rate': 1.6402e-06, 'epoch': 0.88, 'throughput': 13271.26}

[INFO|2025-05-12 12:45:04] logging.py:157 >> {'loss': 0.5367, 'learning_rate': 1.2590e-06, 'epoch': 0.90, 'throughput': 13272.14}

[INFO|2025-05-12 12:47:42] logging.py:157 >> {'loss': 0.5510, 'learning_rate': 9.2707e-07, 'epoch': 0.91, 'throughput': 13272.46}

[INFO|2025-05-12 12:50:20] logging.py:157 >> {'loss': 0.5597, 'learning_rate': 6.4502e-07, 'epoch': 0.92, 'throughput': 13272.79}

[INFO|2025-05-12 12:52:57] logging.py:157 >> {'loss': 0.5369, 'learning_rate': 4.1346e-07, 'epoch': 0.94, 'throughput': 13273.81}

[INFO|2025-05-12 12:55:33] logging.py:157 >> {'loss': 0.5451, 'learning_rate': 2.3285e-07, 'epoch': 0.95, 'throughput': 13276.09}

[INFO|2025-05-12 12:58:09] logging.py:157 >> {'loss': 0.5237, 'learning_rate': 1.0358e-07, 'epoch': 0.97, 'throughput': 13279.42}

[INFO|2025-05-12 13:00:44] logging.py:157 >> {'loss': 0.5604, 'learning_rate': 2.5908e-08, 'epoch': 0.98, 'throughput': 13283.18}

[INFO|2025-05-12 13:03:19] logging.py:157 >> {'loss': 0.5340, 'learning_rate': 0.0000e+00, 'epoch': 1.00, 'throughput': 13286.80}

[INFO|2025-05-12 13:03:19] trainer.py:3910 >> Saving model checkpoint to saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/checkpoint-69

[INFO|2025-05-12 13:03:19] configuration_utils.py:420 >> Configuration saved in saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/checkpoint-69/config.json

[INFO|2025-05-12 13:03:19] configuration_utils.py:909 >> Configuration saved in saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/checkpoint-69/generation_config.json

[INFO|2025-05-12 13:03:40] modeling_utils.py:2996 >> The model is bigger than the maximum size per checkpoint (5GB) and is going to be split in 3 checkpoint shards. You can find where each parameters has been saved in the index located at saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/checkpoint-69/model.safetensors.index.json.

[INFO|2025-05-12 13:03:40] tokenization_utils_base.py:2491 >> tokenizer config file saved in saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/checkpoint-69/tokenizer_config.json

[INFO|2025-05-12 13:03:40] tokenization_utils_base.py:2500 >> Special tokens file saved in saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/checkpoint-69/special_tokens_map.json

[INFO|2025-05-12 13:03:40] trainer.py:2643 >> 

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



[INFO|2025-05-12 13:03:40] trainer.py:3910 >> Saving model checkpoint to saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k

[INFO|2025-05-12 13:03:40] configuration_utils.py:420 >> Configuration saved in saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/config.json

[INFO|2025-05-12 13:03:40] configuration_utils.py:909 >> Configuration saved in saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/generation_config.json

[INFO|2025-05-12 13:04:02] modeling_utils.py:2996 >> The model is bigger than the maximum size per checkpoint (5GB) and is going to be split in 3 checkpoint shards. You can find where each parameters has been saved in the index located at saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/model.safetensors.index.json.

[INFO|2025-05-12 13:04:02] tokenization_utils_base.py:2491 >> tokenizer config file saved in saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/tokenizer_config.json

[INFO|2025-05-12 13:04:02] tokenization_utils_base.py:2500 >> Special tokens file saved in saves/DeepSeek-Coder-7B-Instruct/freeze/deepseek-nlx-330k/special_tokens_map.json

[WARNING|2025-05-12 13:04:02] logging.py:162 >> No metric eval_loss to plot.

[WARNING|2025-05-12 13:04:02] logging.py:162 >> No metric eval_accuracy to plot.

[INFO|2025-05-12 13:04:02] modelcard.py:449 >> Dropping the following result as it does not have all the necessary fields:
{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}