@@ -33,7 +33,9 @@ This directory is in a bit of messy state and is undergoing some cleaning, pleas
## New script
The new script for fine-tuning a model on a summarization or translation task is `run_seq2seq.py`. It is a lightweight example of how to download and preprocess a dataset from the [🤗 Datasets](https://github.com/huggingface/datasets) library or use your own files (json or csv), then fine-tune one of the architectures above on it.
The new script for fine-tuning a model on a summarization or translation task is `run_seq2seq.py`. It is a lightweight example of how to download and preprocess a dataset from the [🤗 Datasets](https://github.com/huggingface/datasets) library or use your own files (jsonlines or csv), then fine-tune one of the architectures above on it.
For custom datasets in `jsonlines` format please see: https://huggingface.co/docs/datasets/loading_datasets.html#json-files
And here is how you would use it on your own files (replace `path_to_csv_or_json_file`, `text_column_name` and `summary_column_name` by the relevant values):
And here is how you would use it on your own files (replace `path_to_csv_or_jsonlines_file`, `text_column_name` and `summary_column_name` by the relevant values):
```bash
python examples/seq2seq/run_seq2seq.py \
-model_name_or_path t5-small \
--model_name_or_path t5-small \
--do_train\
--do_eval\
--task summarization \
--train_file path_to_csv_or_json_file \
--validation_file path_to_csv_or_json_file \
--train_file path_to_csv_or_jsonlines_file \
--validation_file path_to_csv_or_jsonlines_file \
--output_dir ~/tmp/tst-summarization \
--overwrite_output_dir\
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--predict_with_generate\
--text_column text_column_name \
--summary_column summary_column_name
--summary_column summary_column_name
```
The training and validation files should have a column for the inputs texts and a column for the summaries.
Here the files are expected to be JSON files, with each input being a dictionary with a key `"translation"` containing one key per language (here `"en"` and `"ro"`).
Here the files are expected to be JSONLINES files, with each input being a dictionary with a key `"translation"` containing one key per language (here `"en"` and `"ro"`).
here is a command to run xsum evaluation on 8 GPUS. It is more than linearly faster than run_eval.py in some cases
because it uses SortishSampler to minimize padding. You can also use it on 1 GPU. `data_dir` must have
here is a command to run xsum evaluation on 8 GPUS. It is more than linearly faster than run_eval.py in some cases
because it uses SortishSampler to minimize padding. You can also use it on 1 GPU. `data_dir` must have
`{type_path}.source` and `{type_path}.target`. Run `./run_distributed_eval.py --help` for all clargs.
```bash
...
...
@@ -320,17 +322,17 @@ When using `run_eval.py`, the following features can be useful:
`--info` is an additional argument available for the same purpose of tracking the conditions of the experiment. It's useful to pass things that weren't in the argument list, e.g. a language pair `--info "lang:en-ru"`. But also if you pass `--info` without a value it will fallback to the current date/time string, e.g. `2020-09-13 18:44:43`.
If using `--dump-args --info`, the output will be:
* if you need to perform a parametric search in order to find the best ones that lead to the highest BLEU score, let `run_eval_search.py` to do the searching for you.
...
...
@@ -341,14 +343,14 @@ When using `run_eval.py`, the following features can be useful:
which will generate `12``(2*3*2)` searches for a product of each hparam. For example the example that was just used will invoke `run_eval.py` repeatedly with:
