Text Classification
Transformers
Safetensors
English
ultrabert
finance
sentiment-analysis
financial-sentiment-analysis
custom_code
fp16
Eval Results (legacy)
Instructions to use beethogedeon/UltraFinBERT-150M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use beethogedeon/UltraFinBERT-150M with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="beethogedeon/UltraFinBERT-150M", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("beethogedeon/UltraFinBERT-150M", trust_remote_code=True) model = AutoModelForSequenceClassification.from_pretrained("beethogedeon/UltraFinBERT-150M", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download test_comparison.json from beethogedeon/UltraFinBERT-150M: direct link, hf CLI and curl.
- Browser
- Download file 2.15 kB
-
https://huggingface.co/beethogedeon/UltraFinBERT-150M/resolve/main/test_comparison.json
- Command line
-
hf download hf://beethogedeon/UltraFinBERT-150M/test_comparison.json
-
curl -L -o test_comparison.json https://huggingface.co/beethogedeon/UltraFinBERT-150M/resolve/main/test_comparison.json
2.15 kB
| [ | |
| { | |
| "model": "beethogedeon/UltraFinBERT-150M", | |
| "accuracy": 0.9646017699115044, | |
| "balanced_accuracy": 0.9542543773362188, | |
| "macro_f1": 0.9615904929520475, | |
| "weighted_f1": 0.9641735110013125, | |
| "macro_precision": 0.9702531262693043, | |
| "macro_recall": 0.9542543773362188, | |
| "log_loss": 0.24856460459181326, | |
| "macro_f1_ci95_low": 0.9373423664558937, | |
| "macro_f1_ci95_high": 0.9814966497430953, | |
| "test_examples": 339, | |
| "truncation_rate": 0.0, | |
| "parameters": 151034883, | |
| "evaluation_dtype": "torch.float16", | |
| "revision": "local FP16 export", | |
| "training_overlap_note": "Held out from this fine-tuning run; upstream pretraining exposure not audited." | |
| }, | |
| { | |
| "model": "ProsusAI/finbert", | |
| "accuracy": 0.9823008849557522, | |
| "balanced_accuracy": 0.9847408026755854, | |
| "macro_f1": 0.9775293265383634, | |
| "weighted_f1": 0.9823943729157764, | |
| "macro_precision": 0.9708343294286398, | |
| "macro_recall": 0.9847408026755854, | |
| "log_loss": 0.11372488239044834, | |
| "macro_f1_ci95_low": 0.9590552079097949, | |
| "macro_f1_ci95_high": 0.9945137250496199, | |
| "test_examples": 339, | |
| "truncation_rate": 0.0, | |
| "parameters": 109484547, | |
| "evaluation_dtype": "torch.float16", | |
| "revision": "0bc2ac10dda6929bcb1f77ccc43c7ec5c424f031", | |
| "training_overlap_note": "Previously fine-tuned on FinancialPhraseBank; overlap with this test is unknown and possible." | |
| }, | |
| { | |
| "model": "beethogedeon/Modern-FinBERT-large", | |
| "accuracy": 0.9823008849557522, | |
| "balanced_accuracy": 0.9764705882352942, | |
| "macro_f1": 0.9748919274144381, | |
| "weighted_f1": 0.9821485315052908, | |
| "macro_precision": 0.9748525002418029, | |
| "macro_recall": 0.9764705882352942, | |
| "log_loss": 0.09822984893543157, | |
| "macro_f1_ci95_low": 0.9525722119021349, | |
| "macro_f1_ci95_high": 0.9916481561294743, | |
| "test_examples": 339, | |
| "truncation_rate": 0.0, | |
| "parameters": 395834371, | |
| "evaluation_dtype": "torch.float16", | |
| "revision": "a5dd04dae984a46e729698ca39f398181c9aaeea", | |
| "training_overlap_note": "Previously fine-tuned on FinancialPhraseBank; overlap with this test is unknown and possible." | |
| } | |
| ] |