search

nyu-mll / spinn

NYU ML² work on sentence encoding with tree structure and dynamic graphs
MIT License
107 stars 38 forks source link

Please put MT architecture behind a flag. #93

Closed woollysocks closed 6 years ago

woollysocks commented 6 years ago

Currently, I can't run classification experiments on master. I get the following error since a target_vocabulary is only defined if you're doing MT.

Traceback (most recent call last): File "/Users/nikita.nangia/anaconda/envs/py36/lib/python3.6/runpy.py", line 193, in _run_module_as_main "main", mod_spec) File "/Users/nikita.nangia/anaconda/envs/py36/lib/python3.6/runpy.py", line 85, in _run_code exec(code, run_globals) File "/Users/nikita.nangia/Documents/General/NYU/ML2/research/spinn/python/spinn/models/rl_classifier.py", line 442, in run(only_forward=FLAGS.expanded_eval_only_mode) File "/Users/nikita.nangia/Documents/General/NYU/ML2/research/spinn/python/spinn/models/rl_classifier.py", line 382, in run FLAGS.training_data_path, FLAGS.eval_data_path) File "/Users/nikita.nangia/Documents/General/NYU/ML2/research/spinn/python/spinn/models/base.py", line 196, in load_data_and_embeddings target_vocabulary=target_vocabulary) if raw_training_data is not None else None UnboundLocalError: local variable 'target_vocabulary' referenced before assignment

woollysocks commented 6 years ago

@anhad13

sleepinyourhat commented 6 years ago

If this isn’t being dealt with now. I think it’s reasonable to roll back all MT related changes for now. AM can fix it on a branch and merge back later. On Mon, Sep 17, 2018 at 6:53 PM Nikita Nangia notifications@github.com wrote:

Currently, I can't run classification experiments on master. I get the following error since a target_vocabulary is only defined if you're doing MT.

Traceback (most recent call last): File "/Users/nikita.nangia/anaconda/envs/py36/lib/python3.6/runpy.py", line 193, in _run_module_as_main "main", mod_spec) File "/Users/nikita.nangia/anaconda/envs/py36/lib/python3.6/runpy.py", line 85, in _run_code exec(code, run_globals) File "/Users/nikita.nangia/Documents/General/NYU/ML2/research/spinn/python/spinn/models/rl_classifier.py", line 442, in run(only_forward=FLAGS.expanded_eval_only_mode) File "/Users/nikita.nangia/Documents/General/NYU/ML2/research/spinn/python/spinn/models/rl_classifier.py", line 382, in run FLAGS.training_data_path, FLAGS.eval_data_path) File "/Users/nikita.nangia/Documents/General/NYU/ML2/research/spinn/python/spinn/models/base.py", line 196, in load_data_and_embeddings target_vocabulary=target_vocabulary) if raw_training_data is not None else None UnboundLocalError: local variable 'target_vocabulary' referenced before assignment

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nyu-mll/spinn/issues/93, or mute the thread https://github.com/notifications/unsubscribe-auth/ABOZWUzwMFAgBnsJCav62-cPhR-qSserks5ucCfJgaJpZM4Ws3-N .

anhad13 commented 6 years ago

I think @woollysocks fixed this. I'll run some NLI experiments just to be sure.

sleepinyourhat commented 6 years ago

Cool! Just chiming in from my phone, so not seeing everything. On Mon, Sep 17, 2018 at 7:54 PM Anhad Mohananey notifications@github.com wrote:

I think @woollysocks https://github.com/woollysocks fixed this. I'll run some NLI experiments just to be sure.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/nyu-mll/spinn/issues/93#issuecomment-422208191, or mute the thread https://github.com/notifications/unsubscribe-auth/ABOZWYrMi3pVFwZU3z0qhTOe5a_UQ5wAks5ucDYbgaJpZM4Ws3-N .

woollysocks commented 6 years ago

Yup! Resolved. Experiments running now.