- This model is a fine-tuned version of google/t5-v1_1-base on the
vblagoje/lfqadataset, with training duration of 2 epochs, for a (somewhat) apples-to-apples comparison with t5-base on the standard eli5 dataset.
- This checkpoint does seem to be more coherent than t5-base on the original dataset.
- Compared to bart on lfqa, it seems to be able to respond to some questions independently of retrieval.
NOTE: the inference API is limited to generating approx. 64 chars for runtime reasons, for longer outputs try using it in python as a transformers pipeline object.
Intended uses & limitations
- Q&A, information retrieval
- it is probably better to use it with a retrieval pipeline than alone
Training and evaluation data
- see linked dataset. the dataset was filtered to only included the
asksciencesubreddit in an attempt to focus on academic/technical queries.
The following hyperparameters were used during training:
- learning_rate: 4e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 2
- Transformers 4.16.2
- Pytorch 1.10.0+cu113
- Datasets 1.18.3
- Tokenizers 0.11.0
- Downloads last month
This model can be loaded on the Inference API on-demand.