transformers/examples/research_projects/seq2seq-distillation/precomputed_pseudo_labels.md

3.7 KiB

Saved Pseudo-Labels

These are the generations of various large models on various large training sets. All in all they took about 200 GPU hours to produce.

Available Pseudo-labels

Dataset Model Link Rouge Scores Notes
XSUM facebook/bart-large-xsum download 49.8/28.0/42.5
XSUM google/pegasus-xsum download 53.3/32.7/46.5
XSUM facebook/bart-large-xsum download Bart pseudolabels filtered to those with Rouge2 > 10.0 w GT.
CNN/DM sshleifer/pegasus-cnn-ft-v2 download 47.316/26.65/44.56 do not worry about the fact that train.source is one line shorter.
CNN/DM facebook/bart-large-cnn download 5K (2%) are missing, there should be 282173
CNN/DM google/pegasus-xsum download 21.5/6.76/25 extra labels for xsum distillation Used max_source_length=512, (and all other pegasus-xsum configuration).
EN-RO Helsinki-NLP/opus-mt-en-ro download
EN-RO facebook/mbart-large-en-ro download

(EN_RO = WMT 2016 English-Romanian).

Example Download Command:

curl -S https://cdn-datasets.huggingface.co/pseudo/xsum/bart_xsum_pl.tgz | tar -xvz -C .

Generating New Pseudolabels

Here is the command I used to generate the pseudolabels in the second row of the table, after downloading XSUM from here.

python -m torch.distributed.launch --nproc_per_node=8 run_distributed_eval.py \
    --model_name google/pegasus-xsum \ 
    --save_dir pegasus_xsum \ 
    --data_dir xsum \
    --bs 8 --sync_timeout 60000 \
    --max_source_length 512 \
    --type_path train
  • These commands takes a while to run. For example, pegasus_cnn_cnn_pls.tgz took 8 hours on 8 GPUs.
  • Pegasus does not work in fp16 :(, Bart, mBART and Marian do.
  • Even if you have 1 GPU, run_distributed_eval.py is 10-20% faster than run_eval.py because it uses SortishSampler to minimize padding computation.

Contributions

Feel free to contribute your own pseudolabels via PR. Add a row to this table with a new google drive link (or other command line downloadable link).