metadata
license: mit
tags:
- generated_from_trainer
datasets:
- itihasa
metrics:
- bleu
model-index:
- name: M2M101
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: itihasa
type: itihasa
config: Itihasa
split: test
args: Itihasa
metrics:
- name: Bleu
type: bleu
value: 0
M2M101
This model is a fine-tuned version of facebook/m2m100_418M on the itihasa dataset. It achieves the following results on the evaluation set:
- Loss: 7.7213
- Bleu: 0.0
- Gen Len: 2.0
- Meteor: 0.0
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | Meteor |
|---|---|---|---|---|---|---|
| 0.0001 | 1.0 | 4698 | 7.5812 | 0.0 | 2.0 | 0.0 |
| 0.0 | 2.0 | 9396 | 7.0783 | 0.0 | 2.0 | 0.0 |
| 0.0 | 3.0 | 14094 | 8.0417 | 0.0 | 2.0 | 0.0 |
| 0.0 | 4.0 | 18792 | 7.7102 | 0.0 | 2.0 | 0.0 |
| 0.0 | 5.0 | 23490 | 7.7213 | 0.0 | 2.0 | 0.0 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.3.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3