Modify README of bert example #174
Conversation
chenmoneygithub
commented
May 9, 2022
chenmoneygithub
commented
- Fix the curl command, and add curl instruction for vocab downloading.
- Add output flag for glue_fine_tuning script.
examples/bert/README.md
Outdated
| [here](https://storage.googleapis.com/tensorflow/keras-nlp/examples/bert/bert_vocab_uncased.txt), | ||
| or in your terminal run: | ||
| ```shell | ||
| curl https://storage.googleapis.com/tensorflow/keras-nlp/examples/bert/bert_vocab_uncased.txt \ |
There was a problem hiding this comment.
same here curl -O url to keep this a one liner.
There was a problem hiding this comment.
I am doing a renaming here - in the README we are mixing vocab.txt with bert_vocab_uncased.txt, so to unify them I am only keeping vocab.txt at every reference.
|
|
||
| ```shell | ||
| python3 examples/bert/run_glue_finetuning.py \ | ||
| --saved_model_input path/to/model/ \ |
There was a problem hiding this comment.
Can probably just leave this as is?
We don't really need to the output model from fine-tuning to be saved anywhere usually. Really what we would want is to save some predictions in the GLUE submission format (tracked here #115).
There was a problem hiding this comment.
Yea, I feel that users may find the need to reevaluate the model? If we do not save the model then people won't have access to the finetuned model except finetuning again.
chenmoneygithub
force-pushed
the
small-fix
branch
from
806eafc to
ae9d0a2
Compare
examples/bert/README.md
Outdated
|
|
||
| ```shell | ||
| curl https://dumps.wikimedia.org/enwiki/latest/enwiki-latest-pages-articles.xml.bz2 | ||
| curl -O https://dumps.wikimedia.org/enwiki/latest/enwiki-latest-pages-articles.xml.bz2 |
There was a problem hiding this comment.
looks like trailing space here.
There was a problem hiding this comment.
good catch!
chenmoneygithub
force-pushed
the
small-fix
branch
from
08379fc to
d836c28
Compare
