| --- |
| language: |
| - afr |
| - als |
| - amh |
| - arb |
| - ars |
| - ary |
| - arz |
| - asm |
| - azj |
| - bel |
| - ben |
| - bew |
| - bos |
| - bul |
| - cat |
| - ces |
| - ckb |
| - cmn |
| - cym |
| - dan |
| - deu |
| - div |
| - ekk |
| - ell |
| - eng |
| - epo |
| - eus |
| - fao |
| - fas |
| - fil |
| - fin |
| - fra |
| - fry |
| - gle |
| - glg |
| - guj |
| - hau |
| - heb |
| - hin |
| - hrv |
| - hun |
| - hye |
| - ind |
| - isl |
| - ita |
| - jpn |
| - kan |
| - kat |
| - kaz |
| - khk |
| - khm |
| - kin |
| - kir |
| - kmr |
| - kor |
| - lao |
| - lat |
| - lit |
| - ltz |
| - lvs |
| - mal |
| - mar |
| - mkd |
| - mlt |
| - mya |
| - nld |
| - nno |
| - nob |
| - npi |
| - nrm |
| - ory |
| - pan |
| - pbt |
| - plt |
| - pol |
| - por |
| - ron |
| - rus |
| - sin |
| - slk |
| - slv |
| - snd |
| - som |
| - spa |
| - srp |
| - swe |
| - swh |
| - tam |
| - tel |
| - tgk |
| - tha |
| - tur |
| - ukr |
| - urd |
| - uzn |
| - vie |
| - xho |
| - yue |
| - zsm |
| license: mit |
| base_model: |
| - intfloat/multilingual-e5-small |
| datasets: |
| - agentlans/multilingual-document-classification |
| metrics: |
| - f1 |
| - loss |
| model-index: |
| - name: multilingual-e5-small-doc-type-v1-classifier |
| results: |
| - task: |
| type: text-classification |
| name: Text Classification |
| metrics: |
| - type: f1 |
| value: 0.8794 |
| name: Evaluation F1 |
| - type: loss |
| value: 0.6096 |
| name: Evaluation Loss |
| --- |
| # multilingual-e5-small Document Type V1 Classifier |
|
|
| A fine-tuned version of the **bert** architecture (`BertForSequenceClassification`) optimized for the `text-classification` task. |
|
|
| - **Model type:** bert |
| - **Problem Type:** single_label_classification |
| - **Number of Labels:** 17 |
| - **Vocabulary Size:** 250037 |
| - **License:** MIT |
|
|
| ## Use |
|
|
| To get started with this model in Python using the Hugging Face Transformers library, run the following code: |
|
|
| ```python |
| from transformers import AutoTokenizer, AutoModelForSequenceClassification |
| import torch |
| |
| model_id = "agentlans/multilingual-e5-small-doc-type-v1-classifier" |
| tokenizer = AutoTokenizer.from_pretrained(model_id) |
| model = AutoModelForSequenceClassification.from_pretrained(model_id) |
| |
| text = "Replace this with your input text." |
| inputs = tokenizer(text, return_tensors="pt") |
| |
| with torch.no_grad(): |
| logits = model(**inputs).logits |
| |
| predicted_class_id = logits.argmax().item() |
| predicted_class_name = model.config.id2label[predicted_class_id] |
| |
| print(f"Predicted Class ID: {predicted_class_id}") |
| print(f"Predicted Class Name: {predicted_class_name}") |
| ``` |
|
|
| ## Intended Uses & Limitations |
|
|
| ### Intended Use |
| This model is designed for sequence classification tasks. Below are the specific class labels mapped to their corresponding IDs: |
|
|
| | Label ID | Label Name | |
| |---|---| |
| | 0 | Academic/Research | |
| | 1 | Adult | |
| | 2 | Code | |
| | 3 | E-Commerce | |
| | 4 | Government | |
| | 5 | Legal | |
| | 6 | Literary | |
| | 7 | Machine-Generated | |
| | 8 | Media | |
| | 9 | News/Editorial | |
| | 10 | Other | |
| | 11 | Personal | |
| | 12 | Promotional | |
| | 13 | Reference | |
| | 14 | Reviews | |
| | 15 | Search | |
| | 16 | Social | |
|
|
| ## Training Details |
|
|
| ### Hyperparameters |
| The following hyperparameters were used during fine-tuning: |
| - **Learning Rate:** 5e-05 |
| - **Train Batch Size:** 8 |
| - **Eval Batch Size:** 8 |
| - **Optimizer:** OptimizerNames.ADAMW_TORCH_FUSED |
| - **Number of Epochs:** 3.0 |
| - **Mixed Precision:** BF16 |
|
|
| <details> |
| <summary><b>Show Advanced Training Configuration</b></summary> |
|
|
| #### Optimization & Regularization |
| - **Gradient Accumulation Steps:** 1 |
| - **Learning Rate Scheduler:** SchedulerType.LINEAR |
| - **Warmup Steps:** 0 |
| - **Warmup Ratio:** None |
| - **Weight Decay:** 0.0 |
| - **Max Gradient Norm:** 1.0 |
|
|
| #### Hardware & Reproducibility |
| - **Number of GPUs:** 1 |
| - **Seed:** 42 |
|
|
| </details> |
|
|
| ## Training Results & Evaluation |
|
|
| During fine-tuning, the model achieved the following results on the evaluation set: |
|
|
| | Metric | Value | |
| |---|---| |
| | **Train Loss** | 0.3726 | |
| | **Validation Loss** | 0.6096 | |
| | **Validation F1 Score** | 0.8794 | |
| | **Total FLOPs** | 7.9063e+15 | |
|
|
| ### Speed Performance |
| - **Training Runtime:** 1624.1756 seconds |
| - **Train Samples per Second:** 295.512 |
| - **Evaluation Runtime:** 10.6093 seconds |
| - **Eval Samples per Second:** 1886.082 |
|
|
|
|
| <details> |
| <summary><b>Show Detailed Training Logs</b></summary> |
|
|
| ### Training Logs History |
|
|
| | Step | Epoch | Learning Rate | Training Loss | Validation Loss | Validation F1 | |
| |---|---|---|---|---|---| |
| | 500 | 0.025 | 4.9584e-05 | 1.2379 | N/A | N/A | |
| | 1000 | 0.05 | 4.9167e-05 | 0.8651 | N/A | N/A | |
| | 1500 | 0.075 | 4.8751e-05 | 0.7379 | N/A | N/A | |
| | 2000 | 0.1 | 4.8334e-05 | 0.7292 | N/A | N/A | |
| | 2500 | 0.125 | 4.7917e-05 | 0.696 | N/A | N/A | |
| | 3000 | 0.15 | 4.7501e-05 | 0.711 | N/A | N/A | |
| | 3500 | 0.175 | 4.7084e-05 | 0.6598 | N/A | N/A | |
| | 4000 | 0.2 | 4.6667e-05 | 0.6057 | N/A | N/A | |
| | 4500 | 0.225 | 4.6251e-05 | 0.585 | N/A | N/A | |
| | 5000 | 0.25 | 4.5834e-05 | 0.5894 | N/A | N/A | |
| | 5500 | 0.275 | 4.5417e-05 | 0.5759 | N/A | N/A | |
| | 6000 | 0.3 | 4.5001e-05 | 0.5605 | N/A | N/A | |
| | 6500 | 0.325 | 4.4584e-05 | 0.5548 | N/A | N/A | |
| | 7000 | 0.35 | 4.4167e-05 | 0.5508 | N/A | N/A | |
| | 7500 | 0.375 | 4.3751e-05 | 0.5182 | N/A | N/A | |
| | 8000 | 0.4 | 4.3334e-05 | 0.5597 | N/A | N/A | |
| | 8500 | 0.425 | 4.2917e-05 | 0.5342 | N/A | N/A | |
| | 9000 | 0.45 | 4.2500e-05 | 0.5154 | N/A | N/A | |
| | 9500 | 0.475 | 4.2084e-05 | 0.5101 | N/A | N/A | |
| | 10000 | 0.5 | 4.1667e-05 | 0.5153 | N/A | N/A | |
| | 10500 | 0.525 | 4.1250e-05 | 0.4962 | N/A | N/A | |
| | 11000 | 0.55 | 4.0834e-05 | 0.5055 | N/A | N/A | |
| | 11500 | 0.575 | 4.0417e-05 | 0.5289 | N/A | N/A | |
| | 12000 | 0.6 | 4.0000e-05 | 0.5024 | N/A | N/A | |
| | 12500 | 0.625 | 3.9584e-05 | 0.481 | N/A | N/A | |
| | 13000 | 0.65 | 3.9167e-05 | 0.4843 | N/A | N/A | |
| | 13500 | 0.675 | 3.8750e-05 | 0.4519 | N/A | N/A | |
| | 14000 | 0.7 | 3.8334e-05 | 0.4829 | N/A | N/A | |
| | 14500 | 0.725 | 3.7917e-05 | 0.4746 | N/A | N/A | |
| | 15000 | 0.75 | 3.7500e-05 | 0.5123 | N/A | N/A | |
| | 15500 | 0.775 | 3.7084e-05 | 0.5058 | N/A | N/A | |
| | 16000 | 0.8 | 3.6667e-05 | 0.453 | N/A | N/A | |
| | 16500 | 0.825 | 3.6250e-05 | 0.4604 | N/A | N/A | |
| | 17000 | 0.85 | 3.5833e-05 | 0.4689 | N/A | N/A | |
| | 17500 | 0.875 | 3.5417e-05 | 0.4689 | N/A | N/A | |
| | 18000 | 0.9 | 3.5000e-05 | 0.4704 | N/A | N/A | |
| | 18500 | 0.925 | 3.4583e-05 | 0.4367 | N/A | N/A | |
| | 19000 | 0.95 | 3.4167e-05 | 0.451 | N/A | N/A | |
| | 19500 | 0.975 | 3.3750e-05 | 0.4538 | N/A | N/A | |
| | 19999 | 1.0 | N/A | N/A | 0.4387 | 0.8656 | |
| | 20000 | 1.0 | 3.3333e-05 | 0.4367 | N/A | N/A | |
| | 20500 | 1.025 | 3.2917e-05 | 0.3614 | N/A | N/A | |
| | 21000 | 1.05 | 3.2500e-05 | 0.3757 | N/A | N/A | |
| | 21500 | 1.075 | 3.2083e-05 | 0.3197 | N/A | N/A | |
| | 22000 | 1.1 | 3.1667e-05 | 0.3649 | N/A | N/A | |
| | 22500 | 1.125 | 3.1250e-05 | 0.3736 | N/A | N/A | |
| | 23000 | 1.15 | 3.0833e-05 | 0.3325 | N/A | N/A | |
| | 23500 | 1.175 | 3.0417e-05 | 0.3472 | N/A | N/A | |
| | 24000 | 1.2 | 3.0000e-05 | 0.3513 | N/A | N/A | |
| | 24500 | 1.225 | 2.9583e-05 | 0.3699 | N/A | N/A | |
| | 25000 | 1.25 | 2.9166e-05 | 0.3847 | N/A | N/A | |
| | 25500 | 1.275 | 2.8750e-05 | 0.3252 | N/A | N/A | |
| | 26000 | 1.3 | 2.8333e-05 | 0.3573 | N/A | N/A | |
| | 26500 | 1.325 | 2.7916e-05 | 0.3704 | N/A | N/A | |
| | 27000 | 1.35 | 2.7500e-05 | 0.3269 | N/A | N/A | |
| | 27500 | 1.375 | 2.7083e-05 | 0.3637 | N/A | N/A | |
| | 28000 | 1.4 | 2.6666e-05 | 0.3503 | N/A | N/A | |
| | 28500 | 1.425 | 2.6250e-05 | 0.3503 | N/A | N/A | |
| | 29000 | 1.45 | 2.5833e-05 | 0.3246 | N/A | N/A | |
| | 29500 | 1.475 | 2.5416e-05 | 0.3507 | N/A | N/A | |
| | 30000 | 1.5 | 2.5000e-05 | 0.3274 | N/A | N/A | |
| | 30500 | 1.525 | 2.4583e-05 | 0.3926 | N/A | N/A | |
| | 31000 | 1.55 | 2.4166e-05 | 0.3445 | N/A | N/A | |
| | 31500 | 1.575 | 2.3750e-05 | 0.3397 | N/A | N/A | |
| | 32000 | 1.6 | 2.3333e-05 | 0.3337 | N/A | N/A | |
| | 32500 | 1.625 | 2.2916e-05 | 0.3398 | N/A | N/A | |
| | 33000 | 1.65 | 2.2499e-05 | 0.3457 | N/A | N/A | |
| | 33500 | 1.675 | 2.2083e-05 | 0.3252 | N/A | N/A | |
| | 34000 | 1.7 | 2.1666e-05 | 0.3691 | N/A | N/A | |
| | 34500 | 1.725 | 2.1249e-05 | 0.3334 | N/A | N/A | |
| | 35000 | 1.75 | 2.0833e-05 | 0.3363 | N/A | N/A | |
| | 35500 | 1.775 | 2.0416e-05 | 0.3454 | N/A | N/A | |
| | 36000 | 1.8 | 1.9999e-05 | 0.3189 | N/A | N/A | |
| | 36500 | 1.825 | 1.9583e-05 | 0.3422 | N/A | N/A | |
| | 37000 | 1.85 | 1.9166e-05 | 0.3355 | N/A | N/A | |
| | 37500 | 1.875 | 1.8749e-05 | 0.3195 | N/A | N/A | |
| | 38000 | 1.9 | 1.8333e-05 | 0.2937 | N/A | N/A | |
| | 38500 | 1.925 | 1.7916e-05 | 0.3382 | N/A | N/A | |
| | 39000 | 1.95 | 1.7499e-05 | 0.3509 | N/A | N/A | |
| | 39500 | 1.975 | 1.7083e-05 | 0.3244 | N/A | N/A | |
| | 39998 | 2.0 | N/A | N/A | 0.515 | 0.8739 | |
| | 40000 | 2.0 | 1.6666e-05 | 0.3325 | N/A | N/A | |
| | 40500 | 2.025 | 1.6249e-05 | 0.2202 | N/A | N/A | |
| | 41000 | 2.05 | 1.5832e-05 | 0.2126 | N/A | N/A | |
| | 41500 | 2.075 | 1.5416e-05 | 0.1978 | N/A | N/A | |
| | 42000 | 2.1 | 1.4999e-05 | 0.2235 | N/A | N/A | |
| | 42500 | 2.125 | 1.4582e-05 | 0.2285 | N/A | N/A | |
| | 43000 | 2.15 | 1.4166e-05 | 0.2114 | N/A | N/A | |
| | 43500 | 2.175 | 1.3749e-05 | 0.2401 | N/A | N/A | |
| | 44000 | 2.2 | 1.3332e-05 | 0.2316 | N/A | N/A | |
| | 44500 | 2.225 | 1.2916e-05 | 0.2356 | N/A | N/A | |
| | 45000 | 2.25 | 1.2499e-05 | 0.2265 | N/A | N/A | |
| | 45500 | 2.275 | 1.2082e-05 | 0.2156 | N/A | N/A | |
| | 46000 | 2.3 | 1.1666e-05 | 0.1985 | N/A | N/A | |
| | 46500 | 2.325 | 1.1249e-05 | 0.2341 | N/A | N/A | |
| | 47000 | 2.35 | 1.0832e-05 | 0.2253 | N/A | N/A | |
| | 47500 | 2.375 | 1.0416e-05 | 0.2155 | N/A | N/A | |
| | 48000 | 2.4 | 9.9988e-06 | 0.1964 | N/A | N/A | |
| | 48500 | 2.425 | 9.5821e-06 | 0.2406 | N/A | N/A | |
| | 49000 | 2.45 | 9.1655e-06 | 0.2345 | N/A | N/A | |
| | 49500 | 2.475 | 8.7488e-06 | 0.2179 | N/A | N/A | |
| | 50000 | 2.5 | 8.3321e-06 | 0.2076 | N/A | N/A | |
| | 50500 | 2.525 | 7.9154e-06 | 0.2387 | N/A | N/A | |
| | 51000 | 2.55 | 7.4987e-06 | 0.2114 | N/A | N/A | |
| | 51500 | 2.575 | 7.0820e-06 | 0.1916 | N/A | N/A | |
| | 52000 | 2.6 | 6.6653e-06 | 0.2074 | N/A | N/A | |
| | 52500 | 2.625 | 6.2486e-06 | 0.2133 | N/A | N/A | |
| | 53000 | 2.65 | 5.8320e-06 | 0.2301 | N/A | N/A | |
| | 53500 | 2.675 | 5.4153e-06 | 0.2216 | N/A | N/A | |
| | 54000 | 2.7 | 4.9986e-06 | 0.2313 | N/A | N/A | |
| | 54500 | 2.725 | 4.5819e-06 | 0.1916 | N/A | N/A | |
| | 55000 | 2.75 | 4.1652e-06 | 0.2055 | N/A | N/A | |
| | 55500 | 2.775 | 3.7485e-06 | 0.2059 | N/A | N/A | |
| | 56000 | 2.8 | 3.3318e-06 | 0.2021 | N/A | N/A | |
| | 56500 | 2.825 | 2.9151e-06 | 0.2075 | N/A | N/A | |
| | 57000 | 2.85 | 2.4985e-06 | 0.1644 | N/A | N/A | |
| | 57500 | 2.875 | 2.0818e-06 | 0.2023 | N/A | N/A | |
| | 58000 | 2.9 | 1.6651e-06 | 0.2175 | N/A | N/A | |
| | 58500 | 2.925 | 1.2484e-06 | 0.2073 | N/A | N/A | |
| | 59000 | 2.95 | 8.3171e-07 | 0.2154 | N/A | N/A | |
| | 59500 | 2.975 | 4.1502e-07 | 0.2132 | N/A | N/A | |
| | 59997 | 3.0 | N/A | N/A | 0.6096 | 0.8794 | |
|
|
| </details> |
|
|
|
|
| ## Framework Versions |
|
|
| - **Transformers:** 5.0.0.dev0 |
| - **PyTorch:** 2.9.1+cu128 |
|
|
