iva_mt_wslot-m2m100_418M-en-fr
This model is a fine-tuned version of facebook/m2m100_418M on the iva_mt_wslot dataset. It achieves the following results on the evaluation set:
- Loss: 0.0094
- Bleu: 72.5602
- Gen Len: 21.9543
Model description
More information needed
How to use
First please make sure to install pip install transformers
. First download model:
from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
import torch
def translate(input_text, lang):
input_ids = tokenizer(input_text, return_tensors="pt")
generated_tokens = model.generate(**input_ids, forced_bos_token_id=tokenizer.get_lang_id(lang))
return tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
model_name = "cartesinus/iva_mt_wslot-m2m100_418M-0.1.0-en-fr"
tokenizer = M2M100Tokenizer.from_pretrained(model_name, src_lang="en", tgt_lang="fr")
model = M2M100ForConditionalGeneration.from_pretrained(model_name)
Then you can translate either plain text like this:
print(translate("set the temperature on my thermostat", "fr"))
or you can translate with slot annotations that will be restored in tgt language:
print(translate("wake me up at <a>nine am<a> on <b>friday<b>", "fr"))
Limitations of translation with slot transfer:
- Annotated words must be placed between semi-xml tags like this "this is <a>example<a>"
- There is no closing tag for example "<\a>" in the above example - this is done on purpose to omit problems with backslash escape
- If the sentence consists of more than one slot then simply use the next alphabet letter. For example "this is <a>example<a> with more than <b>one<b> slot"
- Please do not add space before the first or last annotated word because this particular model was trained this way and it most probably will lower its results
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 7
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
---|---|---|---|---|---|
0.0132 | 1.0 | 1700 | 0.0110 | 68.7161 | 21.6874 |
0.0083 | 2.0 | 3400 | 0.0093 | 70.3712 | 21.9443 |
0.006 | 3.0 | 5100 | 0.0093 | 71.5485 | 21.995 |
0.0044 | 4.0 | 6800 | 0.0091 | 71.2971 | 21.8371 |
0.0032 | 5.0 | 8500 | 0.0093 | 71.9252 | 21.9268 |
0.0026 | 6.0 | 10200 | 0.0094 | 72.2756 | 21.9543 |
0.002 | 7.0 | 11900 | 0.0094 | 72.5602 | 21.9543 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
Citation
If you use this model, please cite the following:
@article{Sowanski2023SlotLI,
title={Slot Lost in Translation? Not Anymore: A Machine Translation Model for Virtual Assistants with Type-Independent Slot Transfer},
author={Marcin Sowanski and Artur Janicki},
journal={2023 30th International Conference on Systems, Signals and Image Processing (IWSSIP)},
year={2023},
pages={1-5}
}
- Downloads last month
- 16
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Model tree for cartesinus/iva_mt_wslot-m2m100_418M-en-fr
Base model
facebook/m2m100_418MDataset used to train cartesinus/iva_mt_wslot-m2m100_418M-en-fr
Evaluation results
- Bleu on iva_mt_wslotvalidation set self-reported72.560