T5 (small) fine-tuned on Text2Log
This model is a fine-tuned version of t5-small on an Text2Log dataset. It achieves the following results on the evaluation set:
- Loss: 0.0313
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: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 6
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
0.0749 | 1.0 | 21661 | 0.0509 |
0.0564 | 2.0 | 43322 | 0.0396 |
0.0494 | 3.0 | 64983 | 0.0353 |
0.0425 | 4.0 | 86644 | 0.0332 |
0.04 | 5.0 | 108305 | 0.0320 |
0.0381 | 6.0 | 129966 | 0.0313 |
Usage:
from transformers import AutoTokenizer, T5ForConditionalGeneration
MODEL_CKPT = "mrm8488/t5-small-finetuned-text2log"
model = T5ForConditionalGeneration.from_pretrained(MODEL_CKPT).to(device)
tokenizer = AutoTokenizer.from_pretrained(MODEL_CKPT)
def translate(text):
inputs = tokenizer(text, padding="longest", max_length=64, return_tensors="pt")
input_ids = inputs.input_ids.to(device)
attention_mask = inputs.attention_mask.to(device)
output = model.generate(input_ids, attention_mask=attention_mask, early_stopping=False, max_length=64)
return tokenizer.decode(output[0], skip_special_tokens=True)
prompt_nl_to_fol = "translate to fol: "
prompt_fol_to_nl = "translate to nl: "
example_1 = "Every killer leaves something."
example_2 = "all x1.(_woman(x1) -> exists x2.(_emotion(x2) & _experience(x1,x2)))"
print(translate(prompt_nl_to_fol + example_1)) # all x1.(_killer(x1) -> exists x2._leave(x1,x2))
print(translate(prompt_fol_to_nl + example_2)) # Every woman experiences emotions.
Framework versions
- Transformers 4.17.0.dev0
- Pytorch 1.10.0+cu111
- Datasets 1.18.3
- Tokenizers 0.11.0
- Downloads last month
- 21
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.