boyuanzheng010 / mmc Goto Github PK
View Code? Open in Web Editor NEWThis repo contains data and experiment code for the paper "Multilingual Coreference Resolution in Multiparty Dialogue" TACL2023
This repo contains data and experiment code for the paper "Multilingual Coreference Resolution in Multiparty Dialogue" TACL2023
I've found three sentences in conll_format that I believe have malformed parenthesis in the final coreference column (below). Do you know how you handled these cases for the evaluations? I would use the .jsonlines files directly; however, I am planning to use a different tokenizer. Thanks!
The (3 is hanging in:
File: "mmc_fa_corrected/test.farsi.v4_gold_conll"
Doc: "(s09e04c04t); part 004"
s09e04c04t 4 0 من na na na na na Raj na na na (0)
s09e04c04t 4 1 یه na na na na na Raj na na na (1
s09e04c04t 4 2 ایده na na na na na Raj na na na -
s09e04c04t 4 3 برای na na na na na Raj na na na -
s09e04c04t 4 4 آهنگ na na na na na Raj na na na (2
s09e04c04t 4 5 فیلک na na na na na Raj na na na 1)|2)
s09e04c04t 4 6 دارم na na na na na Raj na na na -
s09e04c04t 4 7 یه na na na na na Raj na na na -
s09e04c04t 4 8 چیزی na na na na na Raj na na na (3
s09e04c04t 4 9 تو na na na na na Raj na na na -
s09e04c04t 4 10 مایه na na na na na Raj na na na (4
s09e04c04t 4 11 های na na na na na Raj na na na -
s09e04c04t 4 12 یه na na na na na Raj na na na -
s09e04c04t 4 13 شعر na na na na na Raj na na na -
s09e04c04t 4 14 افسانه na na na na na Raj na na na -
s09e04c04t 4 15 na na na na na Raj na na na -
s09e04c04t 4 16 ای na na na na na Raj na na na -
s09e04c04t 4 17 ابرقهرمانی na na na na na Raj na na na -
s09e04c04t 4 18 در na na na na na Raj na na na -
s09e04c04t 4 19 ابعاد na na na na na Raj na na na (5
s09e04c04t 4 20 حماسی na na na na na Raj na na na 4)|5)
The number of open and closed parentheses doesn't match in:
File: "mmc_en/train.english.v4_gold_conll"
Doc: "(s01e04c12t); part 012"
s01e04c12t 12 0 Mary na na na na na Mrs Cooper na na na (0
s01e04c12t 12 1 Cooper na na na na na Mrs Cooper na na na 0)
s01e04c12t 12 2 , na na na na na Mrs Cooper na na na (0
s01e04c12t 12 3 Sheldon na na na na na Mrs Cooper na na na (0
s01e04c12t 12 4 's na na na na na Mrs Cooper na na na 0)|0)
s01e04c12t 12 5 mom na na na na na Mrs Cooper na na na 0)
s01e04c12t 12 6 . na na na na na Mrs Cooper na na na (0
Unclosed (13 in:
File: "mmc_en/test.english.v4_gold_conll"
Doc: "(s09e09c03t); part 003"
s09e09c03t 3 0 May. na na na na na Penny na na na (13
Hello,
I have a problem. I want to use below command in google colab:
!python run.py train_xlmr_base_ml0_d1_farsi_gold 0 0
But I receive :
File "/content/drive/MyDrive/mmc/experiments/run.py", line 1156, in
runner.train(model)
File "/content/drive/MyDrive/mmc/experiments/run.py", line 151, in train
_, loss = model(*example_gpu)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/content/drive/MyDrive/mmc/experiments/model.py", line 109, in forward
return self.get_predictions_and_loss(*input)
File "/content/drive/MyDrive/mmc/experiments/model.py", line 125, in get_predictions_and_loss
mention_doc, _ = self.bert(input_ids, attention_mask=input_mask) # [num seg, num max tokens, emb size]
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/transformers/models/bert/modeling_bert.py", line 1006, in forward
embedding_output = self.embeddings(
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/transformers/models/bert/modeling_bert.py", line 239, in forward
embeddings = self.LayerNorm(embeddings)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/normalization.py", line 196, in forward
return F.layer_norm(
File "/usr/local/lib/python3.10/dist-packages/torch/nn/functional.py", line 2543, in layer_norm
return torch.layer_norm(input, normalized_shape, weight, bias, eps, torch.backends.cudnn.enabled)
RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA
to enable device-side assertions.
I change self.device in run.py and model.py to "cuda:0"
but I receive same error again.
Can you guide me?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.