README.md

metadata

base_model: yihongLiu/furina
tags:
  - generated_from_trainer
model-index:
  - name: furina_seed42_eng_amh_esp_roman
    results: []

furina_seed42_eng_amh_esp_roman

This model is a fine-tuned version of yihongLiu/furina on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0144
• Spearman Corr: 0.8461

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 32
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Spearman Corr
No log	0.59	200	0.0299	0.6782
No log	1.18	400	0.0251	0.7278
No log	1.76	600	0.0202	0.7493
0.0425	2.35	800	0.0194	0.7584
0.0425	2.94	1000	0.0184	0.7737
0.0425	3.53	1200	0.0189	0.7734
0.0184	4.12	1400	0.0180	0.7906
0.0184	4.71	1600	0.0188	0.7909
0.0184	5.29	1800	0.0171	0.7971
0.0184	5.88	2000	0.0165	0.8055
0.0134	6.47	2200	0.0162	0.8059
0.0134	7.06	2400	0.0164	0.8085
0.0134	7.65	2600	0.0169	0.8131
0.0098	8.24	2800	0.0169	0.8171
0.0098	8.82	3000	0.0158	0.8169
0.0098	9.41	3200	0.0152	0.8201
0.0073	10.0	3400	0.0165	0.8197
0.0073	10.59	3600	0.0150	0.8234
0.0073	11.18	3800	0.0152	0.8284
0.0073	11.76	4000	0.0141	0.8338
0.0059	12.35	4200	0.0144	0.8315
0.0059	12.94	4400	0.0147	0.8348
0.0059	13.53	4600	0.0157	0.8327
0.0049	14.12	4800	0.0147	0.8379
0.0049	14.71	5000	0.0149	0.8365
0.0049	15.29	5200	0.0142	0.8360
0.0049	15.88	5400	0.0140	0.8409
0.0042	16.47	5600	0.0135	0.8414
0.0042	17.06	5800	0.0141	0.8410
0.0042	17.65	6000	0.0144	0.8402
0.0037	18.24	6200	0.0151	0.8435
0.0037	18.82	6400	0.0140	0.8431
0.0037	19.41	6600	0.0140	0.8454
0.0033	20.0	6800	0.0136	0.8453
0.0033	20.59	7000	0.0137	0.8446
0.0033	21.18	7200	0.0144	0.8461

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.1