running_log.txt

[INFO|2025-03-21 09:41:46] configuration_utils.py:699 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/config.json

[INFO|2025-03-21 09:41:46] configuration_utils.py:771 >> Model config Qwen2VLConfig {
  "_name_or_path": "Qwen/Qwen2-VL-7B-Instruct",
  "architectures": [
    "Qwen2VLForConditionalGeneration"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 3584,
  "image_token_id": 151655,
  "initializer_range": 0.02,
  "intermediate_size": 18944,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "qwen2_vl",
  "num_attention_heads": 28,
  "num_hidden_layers": 28,
  "num_key_value_heads": 4,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000.0,
  "sliding_window": 32768,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.49.0",
  "use_cache": true,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "in_chans": 3,
    "model_type": "qwen2_vl",
    "spatial_patch_size": 14
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 152064
}


[INFO|2025-03-21 09:41:49] tokenization_utils_base.py:2050 >> loading file vocab.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/vocab.json

[INFO|2025-03-21 09:41:49] tokenization_utils_base.py:2050 >> loading file merges.txt from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/merges.txt

[INFO|2025-03-21 09:41:49] tokenization_utils_base.py:2050 >> loading file tokenizer.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/tokenizer.json

[INFO|2025-03-21 09:41:49] tokenization_utils_base.py:2050 >> loading file added_tokens.json from cache at None

[INFO|2025-03-21 09:41:49] tokenization_utils_base.py:2050 >> loading file special_tokens_map.json from cache at None

[INFO|2025-03-21 09:41:49] tokenization_utils_base.py:2050 >> loading file tokenizer_config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/tokenizer_config.json

[INFO|2025-03-21 09:41:49] tokenization_utils_base.py:2050 >> loading file chat_template.jinja from cache at None

[INFO|2025-03-21 09:41:49] tokenization_utils_base.py:2313 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.

[INFO|2025-03-21 09:41:49] image_processing_base.py:381 >> loading configuration file preprocessor_config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/preprocessor_config.json

[INFO|2025-03-21 09:41:49] image_processing_base.py:381 >> loading configuration file preprocessor_config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/preprocessor_config.json

[WARNING|2025-03-21 09:41:49] logging.py:329 >> Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.48, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.

[INFO|2025-03-21 09:41:49] image_processing_base.py:434 >> Image processor Qwen2VLImageProcessor {
  "do_convert_rgb": true,
  "do_normalize": true,
  "do_rescale": true,
  "do_resize": true,
  "image_mean": [
    0.48145466,
    0.4578275,
    0.40821073
  ],
  "image_processor_type": "Qwen2VLImageProcessor",
  "image_std": [
    0.26862954,
    0.26130258,
    0.27577711
  ],
  "max_pixels": 12845056,
  "merge_size": 2,
  "min_pixels": 3136,
  "patch_size": 14,
  "processor_class": "Qwen2VLProcessor",
  "resample": 3,
  "rescale_factor": 0.00392156862745098,
  "size": {
    "longest_edge": 12845056,
    "shortest_edge": 3136
  },
  "temporal_patch_size": 2
}


[INFO|2025-03-21 09:41:50] tokenization_utils_base.py:2050 >> loading file vocab.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/vocab.json

[INFO|2025-03-21 09:41:50] tokenization_utils_base.py:2050 >> loading file merges.txt from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/merges.txt

[INFO|2025-03-21 09:41:50] tokenization_utils_base.py:2050 >> loading file tokenizer.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/tokenizer.json

[INFO|2025-03-21 09:41:50] tokenization_utils_base.py:2050 >> loading file added_tokens.json from cache at None

[INFO|2025-03-21 09:41:50] tokenization_utils_base.py:2050 >> loading file special_tokens_map.json from cache at None

[INFO|2025-03-21 09:41:50] tokenization_utils_base.py:2050 >> loading file tokenizer_config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/tokenizer_config.json

[INFO|2025-03-21 09:41:50] tokenization_utils_base.py:2050 >> loading file chat_template.jinja from cache at None

[INFO|2025-03-21 09:41:50] tokenization_utils_base.py:2313 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.

[INFO|2025-03-21 09:41:51] processing_utils.py:876 >> Processor Qwen2VLProcessor:
- image_processor: Qwen2VLImageProcessor {
  "do_convert_rgb": true,
  "do_normalize": true,
  "do_rescale": true,
  "do_resize": true,
  "image_mean": [
    0.48145466,
    0.4578275,
    0.40821073
  ],
  "image_processor_type": "Qwen2VLImageProcessor",
  "image_std": [
    0.26862954,
    0.26130258,
    0.27577711
  ],
  "max_pixels": 12845056,
  "merge_size": 2,
  "min_pixels": 3136,
  "patch_size": 14,
  "processor_class": "Qwen2VLProcessor",
  "resample": 3,
  "rescale_factor": 0.00392156862745098,
  "size": {
    "longest_edge": 12845056,
    "shortest_edge": 3136
  },
  "temporal_patch_size": 2
}

- tokenizer: Qwen2TokenizerFast(name_or_path='Qwen/Qwen2-VL-7B-Instruct', vocab_size=151643, model_max_length=32768, is_fast=True, padding_side='left', truncation_side='right', special_tokens={'eos_token': '<|im_end|>', 'pad_token': '<|endoftext|>', 'additional_special_tokens': ['<|im_start|>', '<|im_end|>', '<|object_ref_start|>', '<|object_ref_end|>', '<|box_start|>', '<|box_end|>', '<|quad_start|>', '<|quad_end|>', '<|vision_start|>', '<|vision_end|>', '<|vision_pad|>', '<|image_pad|>', '<|video_pad|>']}, clean_up_tokenization_spaces=False, added_tokens_decoder={
	151643: AddedToken("<|endoftext|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151644: AddedToken("<|im_start|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151645: AddedToken("<|im_end|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151646: AddedToken("<|object_ref_start|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151647: AddedToken("<|object_ref_end|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151648: AddedToken("<|box_start|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151649: AddedToken("<|box_end|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151650: AddedToken("<|quad_start|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151651: AddedToken("<|quad_end|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151652: AddedToken("<|vision_start|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151653: AddedToken("<|vision_end|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151654: AddedToken("<|vision_pad|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151655: AddedToken("<|image_pad|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
	151656: AddedToken("<|video_pad|>", rstrip=False, lstrip=False, single_word=False, normalized=False, special=True),
}
)

{
  "processor_class": "Qwen2VLProcessor"
}


[INFO|2025-03-21 09:41:51] logging.py:157 >> Add <|im_end|> to stop words.

[INFO|2025-03-21 09:41:51] logging.py:157 >> Loading dataset BsKIE3.json...

[INFO|2025-03-21 09:42:04] configuration_utils.py:699 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/config.json

[INFO|2025-03-21 09:42:04] configuration_utils.py:771 >> Model config Qwen2VLConfig {
  "_name_or_path": "Qwen/Qwen2-VL-7B-Instruct",
  "architectures": [
    "Qwen2VLForConditionalGeneration"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 3584,
  "image_token_id": 151655,
  "initializer_range": 0.02,
  "intermediate_size": 18944,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "qwen2_vl",
  "num_attention_heads": 28,
  "num_hidden_layers": 28,
  "num_key_value_heads": 4,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000.0,
  "sliding_window": 32768,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.49.0",
  "use_cache": true,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "in_chans": 3,
    "model_type": "qwen2_vl",
    "spatial_patch_size": 14
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 152064
}


[INFO|2025-03-21 09:42:04] modeling_utils.py:3982 >> loading weights file model.safetensors from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/model.safetensors.index.json

[INFO|2025-03-21 09:43:21] modeling_utils.py:1633 >> Instantiating Qwen2VLForConditionalGeneration model under default dtype torch.bfloat16.

[INFO|2025-03-21 09:43:21] configuration_utils.py:1140 >> Generate config GenerationConfig {
  "bos_token_id": 151643,
  "eos_token_id": 151645
}


[INFO|2025-03-21 09:43:21] modeling_utils.py:1633 >> Instantiating Qwen2VisionTransformerPretrainedModel model under default dtype torch.bfloat16.

[INFO|2025-03-21 09:43:24] modeling_utils.py:4970 >> All model checkpoint weights were used when initializing Qwen2VLForConditionalGeneration.


[INFO|2025-03-21 09:43:24] modeling_utils.py:4978 >> All the weights of Qwen2VLForConditionalGeneration were initialized from the model checkpoint at Qwen/Qwen2-VL-7B-Instruct.
If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2VLForConditionalGeneration for predictions without further training.

[INFO|2025-03-21 09:43:24] configuration_utils.py:1095 >> loading configuration file generation_config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/generation_config.json

[INFO|2025-03-21 09:43:24] configuration_utils.py:1140 >> Generate config GenerationConfig {
  "bos_token_id": 151643,
  "do_sample": true,
  "eos_token_id": [
    151645,
    151643
  ],
  "pad_token_id": 151643,
  "temperature": 0.01,
  "top_k": 1,
  "top_p": 0.001
}


[INFO|2025-03-21 09:43:24] logging.py:157 >> Gradient checkpointing enabled.

[INFO|2025-03-21 09:43:24] logging.py:157 >> Using torch SDPA for faster training and inference.

[INFO|2025-03-21 09:43:24] logging.py:157 >> Upcasting trainable params to float32.

[INFO|2025-03-21 09:43:24] logging.py:157 >> Fine-tuning method: LoRA

[INFO|2025-03-21 09:43:24] logging.py:157 >> Found linear modules: k_proj,gate_proj,v_proj,o_proj,q_proj,up_proj,down_proj

[INFO|2025-03-21 09:43:24] logging.py:157 >> Set vision model not trainable: ['visual.patch_embed', 'visual.blocks'].

[INFO|2025-03-21 09:43:24] logging.py:157 >> Set multi model projector not trainable: visual.merger.

[INFO|2025-03-21 09:43:25] logging.py:157 >> trainable params: 20,185,088 || all params: 8,311,560,704 || trainable%: 0.2429

[INFO|2025-03-21 09:43:25] trainer.py:746 >> Using auto half precision backend

[WARNING|2025-03-21 09:43:25] trainer.py:781 >> No label_names provided for model class `PeftModelForCausalLM`. Since `PeftModel` hides base models input arguments, if label_names is not given, label_names can't be set automatically within `Trainer`. Note that empty label_names list will be used instead.

[INFO|2025-03-21 09:43:25] trainer.py:2405 >> ***** Running training *****

[INFO|2025-03-21 09:43:25] trainer.py:2406 >>   Num examples = 1,050

[INFO|2025-03-21 09:43:25] trainer.py:2407 >>   Num Epochs = 3

[INFO|2025-03-21 09:43:25] trainer.py:2408 >>   Instantaneous batch size per device = 2

[INFO|2025-03-21 09:43:25] trainer.py:2411 >>   Total train batch size (w. parallel, distributed & accumulation) = 16

[INFO|2025-03-21 09:43:25] trainer.py:2412 >>   Gradient Accumulation steps = 8

[INFO|2025-03-21 09:43:25] trainer.py:2413 >>   Total optimization steps = 195

[INFO|2025-03-21 09:43:25] trainer.py:2414 >>   Number of trainable parameters = 20,185,088

[INFO|2025-03-21 09:45:00] logging.py:157 >> {'loss': 0.2783, 'learning_rate': 4.9919e-05, 'epoch': 0.08, 'throughput': 2222.37}

[INFO|2025-03-21 09:46:36] logging.py:157 >> {'loss': 0.1989, 'learning_rate': 4.9676e-05, 'epoch': 0.15, 'throughput': 2194.13}

[INFO|2025-03-21 09:48:17] logging.py:157 >> {'loss': 0.1331, 'learning_rate': 4.9274e-05, 'epoch': 0.23, 'throughput': 2167.74}

[INFO|2025-03-21 09:49:53] logging.py:157 >> {'loss': 0.1029, 'learning_rate': 4.8713e-05, 'epoch': 0.30, 'throughput': 2159.97}

[INFO|2025-03-21 09:51:29] logging.py:157 >> {'loss': 0.0800, 'learning_rate': 4.7999e-05, 'epoch': 0.38, 'throughput': 2155.84}

[INFO|2025-03-21 09:53:04] logging.py:157 >> {'loss': 0.0767, 'learning_rate': 4.7136e-05, 'epoch': 0.46, 'throughput': 2151.49}

[INFO|2025-03-21 09:54:36] logging.py:157 >> {'loss': 0.0596, 'learning_rate': 4.6130e-05, 'epoch': 0.53, 'throughput': 2154.20}

[INFO|2025-03-21 09:56:13] logging.py:157 >> {'loss': 0.0619, 'learning_rate': 4.4986e-05, 'epoch': 0.61, 'throughput': 2151.53}

[INFO|2025-03-21 09:57:44] logging.py:157 >> {'loss': 0.0585, 'learning_rate': 4.3713e-05, 'epoch': 0.69, 'throughput': 2154.83}

[INFO|2025-03-21 09:59:21] logging.py:157 >> {'loss': 0.0642, 'learning_rate': 4.2318e-05, 'epoch': 0.76, 'throughput': 2151.56}

[INFO|2025-03-21 10:00:58] logging.py:157 >> {'loss': 0.0530, 'learning_rate': 4.0811e-05, 'epoch': 0.84, 'throughput': 2151.91}

[INFO|2025-03-21 10:02:34] logging.py:157 >> {'loss': 0.0566, 'learning_rate': 3.9202e-05, 'epoch': 0.91, 'throughput': 2151.74}

[INFO|2025-03-21 10:04:09] logging.py:157 >> {'loss': 0.0496, 'learning_rate': 3.7500e-05, 'epoch': 0.99, 'throughput': 2152.63}

[INFO|2025-03-21 10:05:52] logging.py:157 >> {'loss': 0.0577, 'learning_rate': 3.5717e-05, 'epoch': 1.08, 'throughput': 2151.34}

[INFO|2025-03-21 10:07:18] logging.py:157 >> {'loss': 0.0442, 'learning_rate': 3.3865e-05, 'epoch': 1.15, 'throughput': 2155.99}

[INFO|2025-03-21 10:08:57] logging.py:157 >> {'loss': 0.0469, 'learning_rate': 3.1955e-05, 'epoch': 1.23, 'throughput': 2152.47}

[INFO|2025-03-21 10:10:36] logging.py:157 >> {'loss': 0.0417, 'learning_rate': 3.0001e-05, 'epoch': 1.30, 'throughput': 2151.23}

[INFO|2025-03-21 10:12:08] logging.py:157 >> {'loss': 0.0451, 'learning_rate': 2.8013e-05, 'epoch': 1.38, 'throughput': 2151.68}

[INFO|2025-03-21 10:13:45] logging.py:157 >> {'loss': 0.0438, 'learning_rate': 2.6007e-05, 'epoch': 1.46, 'throughput': 2150.08}

[INFO|2025-03-21 10:15:23] logging.py:157 >> {'loss': 0.0380, 'learning_rate': 2.3993e-05, 'epoch': 1.53, 'throughput': 2149.01}

[INFO|2025-03-21 10:15:23] trainer.py:3942 >> Saving model checkpoint to saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-100

[INFO|2025-03-21 10:15:27] configuration_utils.py:699 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/config.json

[INFO|2025-03-21 10:15:27] configuration_utils.py:771 >> Model config Qwen2VLConfig {
  "architectures": [
    "Qwen2VLForConditionalGeneration"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 3584,
  "image_token_id": 151655,
  "initializer_range": 0.02,
  "intermediate_size": 18944,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "qwen2_vl",
  "num_attention_heads": 28,
  "num_hidden_layers": 28,
  "num_key_value_heads": 4,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000.0,
  "sliding_window": 32768,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.49.0",
  "use_cache": true,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "in_chans": 3,
    "model_type": "qwen2_vl",
    "spatial_patch_size": 14
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 152064
}


[INFO|2025-03-21 10:15:27] tokenization_utils_base.py:2500 >> tokenizer config file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-100/tokenizer_config.json

[INFO|2025-03-21 10:15:27] tokenization_utils_base.py:2509 >> Special tokens file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-100/special_tokens_map.json

[INFO|2025-03-21 10:15:27] image_processing_base.py:261 >> Image processor saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-100/preprocessor_config.json

[INFO|2025-03-21 10:15:27] tokenization_utils_base.py:2500 >> tokenizer config file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-100/tokenizer_config.json

[INFO|2025-03-21 10:15:27] tokenization_utils_base.py:2509 >> Special tokens file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-100/special_tokens_map.json

[INFO|2025-03-21 10:15:27] processing_utils.py:638 >> chat template saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-100/chat_template.json

[INFO|2025-03-21 10:17:03] logging.py:157 >> {'loss': 0.0380, 'learning_rate': 2.1987e-05, 'epoch': 1.61, 'throughput': 2144.53}

[INFO|2025-03-21 10:18:43] logging.py:157 >> {'loss': 0.0397, 'learning_rate': 1.9999e-05, 'epoch': 1.69, 'throughput': 2142.88}

[INFO|2025-03-21 10:20:19] logging.py:157 >> {'loss': 0.0404, 'learning_rate': 1.8045e-05, 'epoch': 1.76, 'throughput': 2142.72}

[INFO|2025-03-21 10:21:58] logging.py:157 >> {'loss': 0.0362, 'learning_rate': 1.6135e-05, 'epoch': 1.84, 'throughput': 2142.84}

[INFO|2025-03-21 10:23:28] logging.py:157 >> {'loss': 0.0397, 'learning_rate': 1.4283e-05, 'epoch': 1.91, 'throughput': 2145.15}

[INFO|2025-03-21 10:25:02] logging.py:157 >> {'loss': 0.0343, 'learning_rate': 1.2500e-05, 'epoch': 1.99, 'throughput': 2146.04}

[INFO|2025-03-21 10:26:45] logging.py:157 >> {'loss': 0.0396, 'learning_rate': 1.0798e-05, 'epoch': 2.08, 'throughput': 2145.80}

[INFO|2025-03-21 10:28:14] logging.py:157 >> {'loss': 0.0394, 'learning_rate': 9.1889e-06, 'epoch': 2.15, 'throughput': 2147.67}

[INFO|2025-03-21 10:29:54] logging.py:157 >> {'loss': 0.0311, 'learning_rate': 7.6819e-06, 'epoch': 2.23, 'throughput': 2146.29}

[INFO|2025-03-21 10:31:26] logging.py:157 >> {'loss': 0.0480, 'learning_rate': 6.2872e-06, 'epoch': 2.30, 'throughput': 2147.33}

[INFO|2025-03-21 10:33:00] logging.py:157 >> {'loss': 0.0374, 'learning_rate': 5.0139e-06, 'epoch': 2.38, 'throughput': 2148.07}

[INFO|2025-03-21 10:34:34] logging.py:157 >> {'loss': 0.0334, 'learning_rate': 3.8702e-06, 'epoch': 2.46, 'throughput': 2148.53}

[INFO|2025-03-21 10:36:10] logging.py:157 >> {'loss': 0.0324, 'learning_rate': 2.8636e-06, 'epoch': 2.53, 'throughput': 2148.97}

[INFO|2025-03-21 10:37:44] logging.py:157 >> {'loss': 0.0342, 'learning_rate': 2.0005e-06, 'epoch': 2.61, 'throughput': 2149.31}

[INFO|2025-03-21 10:39:19] logging.py:157 >> {'loss': 0.0384, 'learning_rate': 1.2866e-06, 'epoch': 2.69, 'throughput': 2149.34}

[INFO|2025-03-21 10:40:58] logging.py:157 >> {'loss': 0.0357, 'learning_rate': 7.2645e-07, 'epoch': 2.76, 'throughput': 2148.19}

[INFO|2025-03-21 10:42:43] logging.py:157 >> {'loss': 0.0351, 'learning_rate': 3.2374e-07, 'epoch': 2.84, 'throughput': 2145.47}

[INFO|2025-03-21 10:44:18] logging.py:157 >> {'loss': 0.0324, 'learning_rate': 8.1067e-08, 'epoch': 2.91, 'throughput': 2146.04}

[INFO|2025-03-21 10:45:51] logging.py:157 >> {'loss': 0.0357, 'learning_rate': 0.0000e+00, 'epoch': 2.99, 'throughput': 2146.55}

[INFO|2025-03-21 10:45:51] trainer.py:3942 >> Saving model checkpoint to saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-195

[INFO|2025-03-21 10:45:51] configuration_utils.py:699 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/config.json

[INFO|2025-03-21 10:45:51] configuration_utils.py:771 >> Model config Qwen2VLConfig {
  "architectures": [
    "Qwen2VLForConditionalGeneration"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 3584,
  "image_token_id": 151655,
  "initializer_range": 0.02,
  "intermediate_size": 18944,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "qwen2_vl",
  "num_attention_heads": 28,
  "num_hidden_layers": 28,
  "num_key_value_heads": 4,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000.0,
  "sliding_window": 32768,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.49.0",
  "use_cache": true,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "in_chans": 3,
    "model_type": "qwen2_vl",
    "spatial_patch_size": 14
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 152064
}


[INFO|2025-03-21 10:45:51] tokenization_utils_base.py:2500 >> tokenizer config file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-195/tokenizer_config.json

[INFO|2025-03-21 10:45:51] tokenization_utils_base.py:2509 >> Special tokens file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-195/special_tokens_map.json

[INFO|2025-03-21 10:45:51] image_processing_base.py:261 >> Image processor saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-195/preprocessor_config.json

[INFO|2025-03-21 10:45:51] tokenization_utils_base.py:2500 >> tokenizer config file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-195/tokenizer_config.json

[INFO|2025-03-21 10:45:51] tokenization_utils_base.py:2509 >> Special tokens file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-195/special_tokens_map.json

[INFO|2025-03-21 10:45:52] processing_utils.py:638 >> chat template saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/checkpoint-195/chat_template.json

[INFO|2025-03-21 10:45:52] trainer.py:2657 >> 

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



[INFO|2025-03-21 10:45:52] image_processing_base.py:261 >> Image processor saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/preprocessor_config.json

[INFO|2025-03-21 10:45:52] tokenization_utils_base.py:2500 >> tokenizer config file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/tokenizer_config.json

[INFO|2025-03-21 10:45:52] tokenization_utils_base.py:2509 >> Special tokens file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/special_tokens_map.json

[INFO|2025-03-21 10:45:52] processing_utils.py:638 >> chat template saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/chat_template.json

[INFO|2025-03-21 10:45:52] trainer.py:3942 >> Saving model checkpoint to saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora

[INFO|2025-03-21 10:45:52] configuration_utils.py:699 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--Qwen--Qwen2-VL-7B-Instruct/snapshots/eed13092ef92e448dd6875b2a00151bd3f7db0ac/config.json

[INFO|2025-03-21 10:45:52] configuration_utils.py:771 >> Model config Qwen2VLConfig {
  "architectures": [
    "Qwen2VLForConditionalGeneration"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 3584,
  "image_token_id": 151655,
  "initializer_range": 0.02,
  "intermediate_size": 18944,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "qwen2_vl",
  "num_attention_heads": 28,
  "num_hidden_layers": 28,
  "num_key_value_heads": 4,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000.0,
  "sliding_window": 32768,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.49.0",
  "use_cache": true,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "in_chans": 3,
    "model_type": "qwen2_vl",
    "spatial_patch_size": 14
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 152064
}


[INFO|2025-03-21 10:45:52] tokenization_utils_base.py:2500 >> tokenizer config file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/tokenizer_config.json

[INFO|2025-03-21 10:45:52] tokenization_utils_base.py:2509 >> Special tokens file saved in saves/Qwen2-VL-7B-Instruct/lora/BS_riche_lora/special_tokens_map.json

[WARNING|2025-03-21 10:45:52] logging.py:162 >> No metric eval_loss to plot.

[WARNING|2025-03-21 10:45:52] logging.py:162 >> No metric eval_accuracy to plot.

[INFO|2025-03-21 10:45:52] modelcard.py:449 >> Dropping the following result as it does not have all the necessary fields:
{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}