README.md

metadata

license: apache-2.0
base_model: t5-base
tags:
  - generated_from_trainer
metrics:
  - bleu
  - wer
model-index:
  - name: randomization_model_new
    results: []

randomization_model_new

This model is a fine-tuned version of t5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.5870
• Bleu: 0.0001
• Wer: 0.9532
• Rougel: 0.1256
• Gen Len: 19.0

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: 10
• eval_batch_size: 10
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Wer	Rougel	Gen Len
2.5389	0.1	100	2.1621	0.0	0.9613	0.1062	18.9925
2.509	0.2	200	2.0311	0.0	0.9589	0.1112	19.0
2.4189	0.3	300	1.9480	0.0	0.9577	0.1143	18.9995
2.3174	0.4	400	1.8855	0.0001	0.9569	0.1163	19.0
2.2219	0.5	500	1.8416	0.0001	0.956	0.118	19.0
2.1971	0.6	600	1.8083	0.0001	0.9556	0.1193	19.0
2.1342	0.7	700	1.7758	0.0001	0.9551	0.1207	19.0
2.179	0.8	800	1.7494	0.0001	0.955	0.121	19.0
2.0986	0.9	900	1.7289	0.0001	0.9546	0.1219	19.0
2.0907	1.0	1000	1.7141	0.0001	0.9542	0.1225	19.0
2.0455	1.1	1100	1.6966	0.0001	0.9542	0.1229	19.0
2.033	1.2	1200	1.6802	0.0001	0.954	0.1234	19.0
2.0234	1.3	1300	1.6669	0.0001	0.9538	0.1234	19.0
2.0025	1.4	1400	1.6557	0.0001	0.954	0.1236	19.0
2.0098	1.5	1500	1.6467	0.0001	0.9539	0.1236	19.0
1.9495	1.6	1600	1.6381	0.0001	0.9539	0.1242	19.0
1.964	1.7	1700	1.6310	0.0001	0.9537	0.1244	19.0
1.9228	1.8	1800	1.6233	0.0001	0.9537	0.1245	19.0
1.9449	1.9	1900	1.6184	0.0001	0.9535	0.1248	19.0
1.9273	2.0	2000	1.6118	0.0001	0.9536	0.1245	19.0
1.9191	2.1	2100	1.6066	0.0001	0.9535	0.1248	19.0
1.9337	2.2	2200	1.6031	0.0001	0.9534	0.1251	19.0
1.9273	2.3	2300	1.5989	0.0001	0.9535	0.1253	19.0
1.9076	2.4	2400	1.5945	0.0001	0.9535	0.1251	19.0
1.8714	2.5	2500	1.5939	0.0001	0.9534	0.1253	19.0
1.9247	2.6	2600	1.5915	0.0001	0.9533	0.1253	19.0
1.8908	2.7	2700	1.5884	0.0001	0.9532	0.1256	19.0
1.8858	2.8	2800	1.5875	0.0001	0.9532	0.1254	19.0
1.9158	2.9	2900	1.5872	0.0001	0.9532	0.1256	19.0
1.8725	3.0	3000	1.5870	0.0001	0.9532	0.1256	19.0

Framework versions

• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1