微调之前的那个版本的模型

The steps to mimic the Bengali fine-tuning process are as follows: https://GitHub.com/PaddlePaddle/ERNIE/blob/release/v1.4/docs/paddleocr_vl_sft_zh.md In the tutorial, the command to install the specified version of paddleocr is given as Python -m pip install -U "paddleocr[doc-parser]" and Python -m pip install -U "paddleocr[doc-parser]"==3.3.0. However, when verifying the model, there is an issue where the weight direction does not match the direction expected by the current PaddleOCR-VL inference loader. Here is the error message that was encountered: (.venv_paddleocr_v1) test@DESKTOP-DGS7RL:~$ paddleocr doc_parser -i testImage.png --vl_rec_model_name "PaddleOCR-VL-0.9B" --vl_rec_model_dir "./PaddleOCR-VL-SFT-Bengali" --save_path="./PaddleOCR-VL-SFT-Bengali_response" Checking connectivity to the model hosts, this may take a while. To bypass this check, set `DISABLE_MODEL_SOURCE_CHECK` to `True`. /home/test/.venv_paddleocr_v1/lib/python3.12/site-packages/paddle/utils/cpp_extension/extension_utils.py:718: UserWarning: No ccache found. Please be aware that recompiling all source files may be required. You can download and install ccache from: https://GitHub.com/ccache/ccache/blob/master/doc/INSTALL.md warnings.warn(warning_message) Creating model: ('PP-DocLayoutV2', None) Model files already exist. Using cached files. To redownload, please delete the directory manually: `/home/test/.paddlex/official_models/PP-DocLayoutV2`. Creating model: ('PaddleOCR-VL-0.9B', './PaddleOCR-VL-SFT-Bengali') Loading configuration file PaddleOCR-VL-SFT-Bengali/config.json Loading weights file PaddleOCR-VL-SFT-Bengali/model.safetensors.index.json use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 use GQA - num_heads: 16- num_key_value_heads: 2 /home/test/.venv_paddleocr_v1/lib/python3.12/site-packages/paddle/utils/decorator_utils.py:420: Warning: Non compatible API. Please refer to https://www.paddlepaddle.org.cn/documentation/docs/en/develop/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.split.html first. warnings.warn( /home/test/.venv_paddleocr_v1/bin/paddleocr, line 8, in <module> sys.exit(console_entry()) ^^^^^^^^^^^^^^^ /home/test/.venv_paddleocr_v1/lib/python3.12/site-packages/paddleocr/__main__.py, line 26, in console_entry main() /home/test/.venv_paddleocr_v1/lib/python3.12/site-packages/paddleocr/_cli.py, line 192, in main …