Downloads · 30 days
11
38% of all-time downloads
s-haturusinghe/mrp_0.5_seed42_b16_e5_radam_s42_msk0.5_ep2
mrp_0.5_seed42_b16_e5_radam_s42_msk0.5_ep2 is a machine learning model from s-haturusinghe. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Downloads · 30 days
11
38% of all-time downloads
All-time downloads
29
Public
Parameters
278M
2.2 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors1.1 GB · 100%
From the Hugging Face model README
Base Model: xlm-roberta-base
Intermediate Task: mrp-
Pre-Finetuned Model: None
Wandb Run URL: https://wandb.ai/s_haturusinghe-university-of-kelaniya/xlmr-base-stage1-42-b16-e5-radam-s42-mrp-msk0.5/runs/j19zo0fd
Epochs: 2
Steps: 410
Validation Time: 0.01s
Validation Loss: 0.177587
Validation Accuracy: 0.939160
Validation F1: 0.484313
{'0': {'precision': 0.9391596856098586, 'recall': 1.0, 'f1-score': 0.9686254232482111, 'support': 29036.0}, '1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 1881.0}, 'accuracy': 0.9391596856098586, 'macro avg': {'precision': 0.4695798428049293, 'recall': 0.5, 'f1-score': 0.48431271162410555, 'support': 30917.0}, 'weighted avg': {'precision': 0.8820209150748085, 'recall': 0.9391596856098586, 'f1-score': 0.9096939479715063, 'support': 30917.0}}
Masked Accuracy: 0.937277
Masked F1: 0.483811
{'0': {'precision': 0.9372767243748282, 'recall': 1.0, 'f1-score': 0.9676229653533813, 'support': 13643.0}, '1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 913.0}, 'accuracy': 0.9372767243748282, 'macro avg': {'precision': 0.4686383621874141, 'recall': 0.5, 'f1-score': 0.48381148267669066, 'support': 14556.0}, 'weighted avg': {'precision': 0.8784876580548077, 'recall': 0.9372767243748282, 'f1-score': 0.9069304833962752, 'support': 14556.0}}
"learning_rate": 2e-05
"epochs": 5
"batch_size": 16
"model": xlm-roberta-base
"intermediate_task": mrp
"n_tk_label": 2
"mask_ratio": 0.5
"seed": 42
"dataset": sold
"finetuning_stage": pre
"val_int": 10000
"patience": 3
"skip_empty_rat": True