---
license: apache-2.0
language:
- ja
library_name: transformers
---
microsoft/layoutxlm-base finetuned on XFUND.ja

Training Results

{
    "epoch": 40.0,
    "eval_accuracy": 0.7919082377476538,
    "eval_f1": 0.7886944818304172,
    "eval_loss": 1.6934013366699219,
    "eval_mem_cpu_alloc_delta": 819200,
    "eval_mem_cpu_peaked_delta": 0,
    "eval_mem_gpu_alloc_delta": 0,
    "eval_mem_gpu_peaked_delta": 377472512,
    "eval_precision": 0.7367979882648784,
    "eval_recall": 0.8484555984555985,
    "eval_runtime": 4.2599,
    "eval_samples": 71,
    "eval_samples_per_second": 16.667,
    "init_mem_cpu_alloc_delta": 1262002176,
    "init_mem_cpu_peaked_delta": 767479808,
    "init_mem_gpu_alloc_delta": 1481701376,
    "init_mem_gpu_peaked_delta": 0,
    "train_mem_cpu_alloc_delta": 28925952,
    "train_mem_cpu_peaked_delta": 1562083328,
    "train_mem_gpu_alloc_delta": 4458109440,
    "train_mem_gpu_peaked_delta": 7030973440,
    "train_runtime": 1135.9773,
    "train_samples": 194,
    "train_samples_per_second": 0.88
}

code: |
from transformers import LayoutXLMProcessor, LayoutLMv2ForTokenClassification
processor = LayoutXLMProcessor.from_pretrained("amir22010/layoutxlm-xfund-ja")
model = LayoutLMv2ForTokenClassification.from_pretrained("amir22010/layoutxlm-xfund-ja",num_labels = 7)