Search is not available for this dataset
repo stringlengths 2 152 ⌀ | file stringlengths 15 239 | code stringlengths 0 58.4M | file_length int64 0 58.4M | avg_line_length float64 0 1.81M | max_line_length int64 0 12.7M | extension_type stringclasses 364
values |
|---|---|---|---|---|---|---|
null | ParlAI-main/projects/dialogue_bias/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from typing import Optional
from parlai.core.message import Message
from projects.style_gen.modules import STYLE_SEP_TO... | 637 | 29.380952 | 82 | py |
null | ParlAI-main/projects/dialogue_bias/util.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import os
from collections import defaultdict
from typing import Dict, List, Tuple
import pandas as pd
RACES_ETHNICIT... | 8,654 | 37.29646 | 89 | py |
null | ParlAI-main/projects/dialogue_safety/README.md | # Build it Break it Fix it for Dialogue Safety: Robustness from Adversarial Human Attack
## Paper information
Emily Dinan, Samuel Humeau, Bharath Chintagunta, Jason Weston.
_[Build it Break it Fix it for Dialogue Safety: Robustness from Adversarial Human Attack](https://arxiv.org/abs/1908.06083)_.
To appear at EMNLP 2... | 5,173 | 63.675 | 1,213 | md |
null | ParlAI-main/projects/dialogue_unlikelihood/README.md | ## Don’t Say _That_! Making Inconsistent Dialogue Unlikely with Unlikelihood Training
Margaret Li, Stephen Roller, Ilia Kulikov, Sean Welleck, Y-Lan Boureau, Kyunghyun Cho, Jason Weston
## Abstract
Generative dialogue models currently suffer from a number of problems which standard maximum likelihood training
does ... | 1,774 | 54.46875 | 227 | md |
null | ParlAI-main/projects/dialogue_unlikelihood/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
| 198 | 38.8 | 65 | py |
null | ParlAI-main/projects/dialogue_unlikelihood/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from typing import Optional
from parlai.core.params import ParlaiParser
from parlai.core.opt import Opt
import json
impo... | 20,340 | 35.00177 | 103 | py |
null | ParlAI-main/projects/director/README.md | # DIRECTOR: Generator-Classifiers For Supervised Language Modeling
Kushal Arora, Kurt Shuster, Sainbayar Sukhbaatar, Jason Weston
<b>Paper Link</b>: [https://arxiv.org/abs/2206.07694](https://arxiv.org/abs/2206.07694)
## Abstract
Current language models achieve low perplexity but their resulting generations still... | 7,077 | 227.322581 | 3,119 | md |
null | ParlAI-main/projects/director/director_agent.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
DirectorAgent for Supervised Language Modeling.
"""
import torch
import torch.nn as nn
import torch.nn.functional a... | 23,406 | 36.213037 | 136 | py |
null | ParlAI-main/projects/director/tasks/safety.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from typing import List, Tuple
from typing import Optional
from parlai.core.agents import (
Agent,
create_agent... | 13,614 | 34.090206 | 107 | py |
null | ParlAI-main/projects/dodecadialogue/README.md | ## The Dialogue Dodecathlon
### _Open-Domain Knowledge and Image Grounded Conversational Agents_
Kurt Shuster, Da Ju, Stephen Roller, Emily Dinan, Y-Lan Boureau, Jason Weston
## Abstract
We introduce dodecaDialogue: a set of 12 tasks which measure if a conversational agent can communicate
engagingly with personal... | 4,164 | 43.784946 | 189 | md |
null | ParlAI-main/projects/drqa/README.md | # DrQA
This archived project was a (partial) implementation of the DrQA Document
Reader from:
*Danqi Chen, Adam Fisch, Jason Weston, Antoine Bordes. 2017.
Reading Wikipedia to Answer Open-Domain Questions.
In Association for Computational Linguistics (ACL).*
Read more about DrQA [here](https://arxiv.org/pdf/1704.000... | 539 | 35 | 85 | md |
null | ParlAI-main/projects/fits/README.md | # Learning New Skills after Deployment: _Improving open-domain internet-driven dialogue with human feedback_
Jing Xu, Megan Ung, Mojtaba Komeili, Kushal Arora, Y-Lan Boureau, Jason Weston
## Abstract
Frozen models trained to mimic static datasets can never improve their performance. Models that can employ internet-r... | 3,579 | 56.741935 | 1,016 | md |
null | ParlAI-main/projects/fits/agents/director_bb2.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import torch
import torch.nn as nn
import torch.nn.functional as F
from typing import Optional, Any, Dict, Tuple, Union,... | 48,862 | 41.196028 | 2,243 | py |
null | ParlAI-main/projects/fits/agents/director_seeker.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import torch
import torch.nn.functional as F
import copy
from typing import Optional, List, Tuple, Any, Dict, Union
fro... | 13,951 | 37.224658 | 212 | py |
null | ParlAI-main/projects/genderation_bias/README.md | ## Queens are Powerful too: Mitigating Gender Bias in Dialogue Generation
Emily Dinan, Angela Fan, Adina Williams, Jack Urbanek, Douwe Kiela, Jason Weston
## Abstract
Models often easily learn biases present in the training data, and their predictions directly reflect this bias.
We analyze the presence of gender bia... | 1,817 | 49.5 | 258 | md |
null | ParlAI-main/projects/genderation_bias/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from typing import Optional
from parlai.core.params import ParlaiParser
from parlai.core.opt import Opt
from parlai.agen... | 1,086 | 28.378378 | 80 | py |
null | ParlAI-main/projects/hallucination/README.md | # Reducing Hallucination in Conversational Agents
### _Retrieval Augmentation Reduces Hallucination in Conversation_
Kurt Shuster, Spencer Poff, Moya Chen, Douwe Kiela\*, Jason Weston\*
\* Equal Contribution.
## Abstract
Despite showing increasingly human-like conversational abilities, state-of-the-art dialogue mo... | 9,143 | 59.96 | 1,147 | md |
null | ParlAI-main/projects/humaneval/README.md | # Human Evaluation of Conversations is an Open Problem: comparing the sensitivity of various methods for evaluating dialogue agents
Eric Michael Smith, Orion Hsu, Rebecca Qian, Stephen Roller, Y-Lan Boureau, Jason Weston
## Abstract
At the heart of improving conversational AI is the open problem of how to evaluate c... | 3,875 | 57.727273 | 851 | md |
null | ParlAI-main/projects/image_chat/README.md | # Engaging Image Chat: Modeling Personality in Grounded Dialogue
Kurt Shuster, Samuel Humeau, Antoine Bordes, Jason Weston
Please see [Shuster et al. (2018)](https://arxiv.org/abs/1811.00945) for more details.
## Abstract
To achieve the long-term goal of machines being able to engage humans in conversation, our mod... | 3,855 | 51.108108 | 970 | md |
null | ParlAI-main/projects/image_chat/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import sys
if sys.version_info < (3, 0):
raise RuntimeError('ParlAI requires Python 3.')
| 293 | 28.4 | 65 | py |
null | ParlAI-main/projects/image_chat/interactive.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Talk with a model using a web UI.
"""
from typing import Dict, Any
from http.server import BaseHTTPRequestHandler, H... | 16,634 | 45.46648 | 3,550 | py |
null | ParlAI-main/projects/image_chat/transresnet_multimodal/__init__.py | 0 | 0 | 0 | py | |
null | ParlAI-main/projects/image_chat/transresnet_multimodal/modules.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Modules for TransresnetMultimodalAgent.
"""
from typing import Optional
from parlai.core.params import ParlaiParser
... | 19,412 | 33.177817 | 93 | py |
null | ParlAI-main/projects/image_chat/transresnet_multimodal/transresnet_multimodal.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Transresnet Multimodal Model (https://arxiv.org/abs/1811.00945).
"""
from typing import Optional
from parlai.core.pa... | 15,704 | 34.292135 | 91 | py |
null | ParlAI-main/projects/k2r/README.md | # Reason first, then respond: Modular Generation for Knowledge-infused Dialogue
Leonard Adolphs, Kurt Shuster, Jack Urbanek, Arthur Szlam, Jason Weston
<b>Paper Link</b>: [https://arxiv.org/abs/2111.05204](https://arxiv.org/abs/2111.05204)
## Abstract
Large language models can produce fluent dialogue but often halluc... | 4,951 | 90.703704 | 1,239 | md |
null | ParlAI-main/projects/k2r/lightqa/task/__init__.py | 0 | 0 | 0 | py | |
null | ParlAI-main/projects/k2r/lightqa/task/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import json
import os
from parlai.core.teachers import DialogTeacher
from parlai.utils.io import PathManager
from parlai... | 2,600 | 32.779221 | 87 | py |
null | ParlAI-main/projects/k2r/lightqa/task/build.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Download and build the data if it does not exist.
"""
from parlai.core.build_data import DownloadableFile
import pa... | 2,029 | 31.741935 | 93 | py |
null | ParlAI-main/projects/k2r/stacked_agent/scripts/stacked_agent_eval.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Evaluate the stacked models.
"""
from projects.k2r.stacked_agent.task.agents import (
StackedKnowledgeDialogueAg... | 10,805 | 37.455516 | 93 | py |
null | ParlAI-main/projects/k2r/stacked_agent/task/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from copy import deepcopy
from typing import Optional, List
from types import MethodType
from collections import default... | 28,280 | 36.310026 | 95 | py |
null | ParlAI-main/projects/k2r/wow/task/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import random
import os
from typing import Optional, Tuple
from tqdm import tqdm
from collections import defaultdict
imp... | 5,756 | 35.903846 | 94 | py |
null | ParlAI-main/projects/light/README.md | # <img width="5%" src="scribe.png"/> LIGHT
### Learning in Interactive Games with Humans and Text
<p align="center"><img width="90%" src="tavern.png" /></p>
The LIGHT project is a large-scale fantasy text adventure game research platform for training agents
that can both talk and act, interacting either with other mod... | 5,149 | 51.55102 | 345 | md |
null | ParlAI-main/projects/light_whoami/README.md | # Am I Me or You? State-of-the-Art Dialogue Models Cannot Maintain Identity
Kurt Shuster, Jack Urbanek, Arthur Szlam, Jason Weston
## Abstract
State-of-the-art dialogue models still often stumble with regards to factual accuracy and self-contradiction. Anecdotally, they have been observed to fail to maintain charact... | 8,817 | 65.300752 | 964 | md |
null | ParlAI-main/projects/light_whoami/agents/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Define PATHS to pre-trained models.
"""
RPA_RERANKER = 'zoo:light_whoami/rpa_reranker/model'
TEST_RPA_RERANKER = 'zo... | 1,053 | 39.538462 | 80 | py |
null | ParlAI-main/projects/light_whoami/agents/expanded_attention.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Expanded Attention Transformer Model.
Specifically, the decoder uses an additional attention mechanism over a (possib... | 43,118 | 36.462207 | 102 | py |
null | ParlAI-main/projects/light_whoami/agents/multi_objective.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Multi-objective Agents.
These agents generally contain extra layers to solve the Am I Me or You task.
"""
from abc im... | 37,138 | 37.287629 | 106 | py |
null | ParlAI-main/projects/light_whoami/agents/pacer.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
PACER: Partial And Complete Efficient Re-ranking.
See `PacerTreeSearchMixin.modify_logprobs` for a complete descripti... | 13,482 | 34.111979 | 132 | py |
null | ParlAI-main/projects/light_whoami/agents/poly_return_weights.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Overrides the standard Polyencoder Agent to only return the attention weights.
"""
import torch
import torch.nn
impor... | 12,281 | 33.307263 | 87 | py |
null | ParlAI-main/projects/light_whoami/agents/rpa_rerank.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
LIGHT Am I Me or You Generative-Rerank Agent.
The retriever reranks beam texts via the following architecture setup:... | 11,316 | 34.58805 | 112 | py |
null | ParlAI-main/projects/light_whoami/agents/rpa_ul.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
LIGHT RPA Unlikelihood Agent.
Utilizes a left-to-right RPA Classifier to predict tokens that yield incorrect charact... | 12,406 | 36.82622 | 117 | py |
null | ParlAI-main/projects/light_whoami/task/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
| 199 | 32.333333 | 65 | py |
null | ParlAI-main/projects/light_whoami/task/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Teachers used in the Am I Me or You task.
"""
from abc import ABC
from collections import deque
import re
from typing... | 22,228 | 34.623397 | 119 | py |
null | ParlAI-main/projects/light_whoami/task/mutators.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Teachers used in the Am I Me or You task.
"""
import re
from parlai.core.message import Message
from parlai.core.muta... | 2,298 | 32.808824 | 87 | py |
null | ParlAI-main/projects/light_whoami/task/utils.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Utilities + Functions.
"""
from collections import deque
import random
from typing import Dict
# DEFAULT_DELIM = ' ... | 2,855 | 25.444444 | 83 | py |
null | ParlAI-main/projects/mastering_the_dungeon/README.md | # Mastering the Dungeon
This project was archived on 2020-08-01. The code had not been tested in a long time.
If you wish to view the old code, you may rewind to the
[`mastering_the_dungeon`](https://github.com/facebookresearch/ParlAI/tree/mastering_the_dungeon/projects/mastering_the_dungeon)
tag.
# Old README
This ... | 4,938 | 34.789855 | 661 | md |
null | ParlAI-main/projects/md_gender/README.md | ## Multi-Dimensional Gender Bias Classification
Emily Dinan, Angela Fan, Ledell Wu, Jason Weston, Douwe Kiela, Adina Williams
## Abstract
Machine learning models are trained to find patterns in data. NLP models can inadvertently learn socially undesirable patterns when training on gender biased text. In this work, w... | 2,238 | 53.609756 | 1,039 | md |
null | ParlAI-main/projects/md_gender/bert_ranker_classifier/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Using classification metrics for a ranking BERT model.
"""
from parlai.agents.bert_ranker.bi_encoder_ranker import ... | 1,085 | 30.028571 | 79 | py |
null | ParlAI-main/projects/memnn_feedback/README.md | # MemNN Feedback
This project was deleted on 2020-08-01, and had not been tested in a long time.
If you wish to see the code, you may rewind to the
[`memnn_feedback`](https://github.com/facebookresearch/ParlAI/tree/memnn_feedback/projects/memnn_feedback)
tag.
| 261 | 36.428571 | 106 | md |
null | ParlAI-main/projects/metacognition/README.md | # Reducing conversational agents' overconfidence through linguistic calibration
## Data
COMING SOON!
## Code
COMING SOON!
## Paper
[Link](https://arxiv.org/abs/2012.14983); accepted to TACL 2022.
| 202 | 13.5 | 79 | md |
null | ParlAI-main/projects/msc/README.md | # Beyond Goldfish Memory: Long-Term Open-Domain Conversation
Jing Xu, Arthur Szlam, Jason Weston
## Abstract
Despite recent improvements in open-domain dialogue models, state of the art models are trained and evaluated on short conversations with little context.
In contrast, the long-term conversation setting has ha... | 2,861 | 40.478261 | 435 | md |
null | ParlAI-main/projects/msc/agents/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
| 198 | 38.8 | 65 | py |
null | ParlAI-main/projects/msc/agents/long_rag.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks.
Original Paper: https://arxiv.org/abs/2005.11401
... | 5,863 | 28.32 | 87 | py |
null | ParlAI-main/projects/msc/agents/long_tga.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import torch
from typing import Optional, Tuple
from parlai.agents.transformer.modules import (
TransformerDecoder,
... | 4,089 | 34.877193 | 130 | py |
null | ParlAI-main/projects/msc/agents/memory_agent.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Agent code for memory agent.
memory agent allows for read/write memory access.
We modify the model input to include ... | 9,358 | 36.890688 | 118 | py |
null | ParlAI-main/projects/multilight/README.md | # Multi-Party Chat: Conversational Agents in Group Settings with Humans and Models
Jimmy Wei, Kurt Shuster, Arthur Szlam, Jason Weston, Jack Urbanek, Mojtaba Komeili
## Abstract
Current dialogue research primarily studies pairwise (two-party) conversations, and does not address the everyday setting where more than t... | 3,177 | 50.258065 | 966 | md |
null | ParlAI-main/projects/multimodal_blenderbot/README.md | **WARNING**
This project has been archived. If you need to use this project, please revert
to ParlAI 1.4.1.
------
# Multi-Modal Open-Domain Dialogue
Kurt Shuster<sup>∗</sup>, Eric Michael Smith<sup>∗</sup>, Da Ju, Jason Weston
## Abstract
Recent work in open-domain conversational agents has demonstrated that sig... | 2,089 | 46.5 | 1,118 | md |
null | ParlAI-main/projects/multimodal_blenderbot/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
| 199 | 32.333333 | 65 | py |
null | ParlAI-main/projects/multimodal_blenderbot/agents.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from parlai.agents.image_seq2seq.image_seq2seq import ImageSeq2seqAgent
from projects.genderation_bias.agents import Bia... | 485 | 27.588235 | 72 | py |
null | ParlAI-main/projects/params_vs_compute/README.md | # Which one is more important: more parameters or more computation?
When we talk about the power of a deep learning model, often the only metric we pay attention to is its size, which is measured by the number parameters in that model. However, the amount of computation to run that model is an important metric too, bu... | 7,793 | 164.829787 | 1,044 | md |
null | ParlAI-main/projects/params_vs_compute/hash_ladder/hash_ladder.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from __future__ import annotations
import torch
from typing import Dict, Optional, Tuple
import torch.nn as nn
from parl... | 7,647 | 32.39738 | 176 | py |
null | ParlAI-main/projects/personachat/README.md | # PersonaChat
This project was archived on 2020-08-01. The models and scripts have not been tested
in a long time. You may continue to use the data without issue (`--task personachat`).
If you need to use the model or scripts, please rewind to the
[`personachat`](https://github.com/facebookresearch/ParlAI/tree/persona... | 1,564 | 42.472222 | 122 | md |
null | ParlAI-main/projects/personality_captions/README.md | # Engaging Image Captioning via Personality
Kurt Shuster, Samuel Humeau, Hexiang Hu, Antoine Bordes, Jason Weston
Please see [Shuster et al. (CVPR 2019)](https://arxiv.org/abs/1810.10665) for more details.
## Abstract
Standard image captioning tasks such as COCO and Flickr30k are factual, neutral in tone and (to a ... | 4,305 | 55.657895 | 1,082 | md |
null | ParlAI-main/projects/personality_captions/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import sys
if sys.version_info < (3, 0):
raise RuntimeError('ParlAI requires Python 3.')
| 293 | 28.4 | 65 | py |
null | ParlAI-main/projects/personality_captions/interactive.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Talk with a model using a web UI.
"""
from http.server import BaseHTTPRequestHandler, HTTPServer
from parlai.script... | 14,003 | 45.83612 | 3,550 | py |
null | ParlAI-main/projects/personality_captions/transresnet/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
| 199 | 32.333333 | 65 | py |
null | ParlAI-main/projects/personality_captions/transresnet/modules.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Model Code.
"""
from typing import Optional
from parlai.core.params import ParlaiParser
from parlai.core.opt import ... | 20,165 | 32.554077 | 88 | py |
null | ParlAI-main/projects/personality_captions/transresnet/transresnet.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Agent code for the model described in (https://arxiv.org/abs/1811.00945).
"""
from typing import Optional
from parla... | 18,285 | 34.097889 | 91 | py |
null | ParlAI-main/projects/polyencoder/README.md | # Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring
Paper: https://arxiv.org/abs/1905.01969
## Abstract
The use of deep pre-trained bidirectional transformers has led to remarkable progress in anumber of applications (Devlin et al., 2018). For ... | 6,810 | 49.828358 | 1,045 | md |
null | ParlAI-main/projects/recipes/README.md | # Recipes for building an open-domain chatbot
Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston
## Abstract
Building open-domain chatbots is a challenging area for machine learning research. While prior work has... | 7,856 | 73.828571 | 996 | md |
null | ParlAI-main/projects/recipes/chatlog_2.7B_render.html | <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN">
<html>
<head>
<meta name="generator" content="HTML Tidy for Mac OS X (vers 31 October 2006 - Apple Inc. build 17.1), see www.w3.org">
<title></title>
</head>
<body>
<h2>2.7B</h2>
<table border="1" frame="void" rules="rows">
<tr>
<th></th>
<... | 609,720 | 60.180112 | 200 | html |
null | ParlAI-main/projects/reframe_thoughts/README.md | # Training Models to Generate, Recognize, and Reframe Unhelpful Thoughts
Mounica Maddela, Megan Ung, Jing Xu, Andrea Madotto, Heather Foran, Y-Lan Boureau
## Abstract
Many cognitive approaches to well-being, such as recognizing and reframing unhelpful thoughts, have received considerable empirical support over the p... | 1,291 | 70.777778 | 1,050 | md |
null | ParlAI-main/projects/roscoe/README.md | # ROSCOE: A Suite of Metrics for Scoring Step-by-Step Reasoning
This repository contains the code and reproduction steps for our paper [ROSCOE: A Suite of Metrics for Scoring Step-by-Step Reasoning](https://openreview.net/pdf?id=xYlJRpzZtsY).
## Content
- [Overview](#overview)
- [Setup](#setup)
- [ROSCOE evalu... | 6,912 | 44.781457 | 650 | md |
null | ParlAI-main/projects/roscoe/roscoe.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Evaluate dataset of generated chains-of-resoning.
Usage:
python projects/roscoe/roscoe.py
"""
import json
import os... | 7,212 | 29.957082 | 92 | py |
null | ParlAI-main/projects/roscoe/score.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from collections import defaultdict
import itertools
import math
import numpy as np
from typing import Any, Dict, Iterab... | 39,044 | 37.354617 | 126 | py |
null | ParlAI-main/projects/roscoe/utils.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import numpy as np
import os
import pathlib
from nltk.tokenize import sent_tokenize
from typing import Callable, Dict, I... | 3,998 | 31.778689 | 95 | py |
null | ParlAI-main/projects/roscoe/baselines/constants.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from dataclasses import dataclass
from enum import Enum
from typing import List, Optional
class UseRef(Enum):
YES =... | 2,760 | 24.564815 | 127 | py |
null | ParlAI-main/projects/roscoe/baselines/run.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import argparse
import glob
import os
from typing import Dict, List
import pandas
import json
from projects.roscoe.utils... | 8,882 | 34.963563 | 124 | py |
null | ParlAI-main/projects/roscoe/baselines/scores.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from typing import Dict, List
from projects.roscoe.baselines.constants import (
BARTSCORE_CNN_F,
BARTSCORE_CNN_P... | 11,754 | 32.206215 | 121 | py |
null | ParlAI-main/projects/roscoe/meta_evaluation/correlations.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import pandas
from scipy.stats import somersd
class Analyzer:
def __init__(self, reference_column: str):
s... | 1,480 | 33.44186 | 81 | py |
null | ParlAI-main/projects/roscoe/meta_evaluation/roscoe_correlations.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Helper function to run correlation analysis on annotated data evaluations, scattered
across model-specific folders.
... | 50,256 | 32.888739 | 121 | py |
null | ParlAI-main/projects/roscoe/meta_evaluation/roscoe_synthetic_correlations.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Helper function to run correlation analysis on synthetic data evaluations, scattered
across model-specific folders.
... | 12,425 | 29.833747 | 91 | py |
null | ParlAI-main/projects/roscoe/meta_evaluation/run_synthetic_correlations.sh | #!/bin/bash
SCORER="./projects/roscoe/meta_evaluation/roscoe_synthetic_correlations.py"
python $SCORER --dataset-name aqua
python $SCORER --dataset-name asdiv
python $SCORER --dataset-name entailment_bank
python $SCORER --dataset-name eqasc
python $SCORER --dataset-name math_dataset
python $SCORER --dataset-name proo... | 328 | 28.909091 | 75 | sh |
null | ParlAI-main/projects/roscoe/meta_evaluation/table_file_writing.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
import copy
import csv
import math
import os
import pandas
from typing import Dict, List, Union
from projects.roscoe.sco... | 14,668 | 36.136709 | 125 | py |
null | ParlAI-main/projects/roscoe/roscoe_data/download_annotated.sh | #!/bin/bash
PATH_TO_DATA="./projects/roscoe/roscoe_data"
RESTORE_SCRIPT="./projects/roscoe/roscoe_data/restore_annotated.py"
mkdir -p ${PATH_TO_DATA}/raw
mkdir -p ${PATH_TO_DATA}/generated
mkdir -p ${PATH_TO_DATA}/annotated
#Step 1. Download [dataset]_reasoning.txt in projects/roscoe/roscoe_data/generated/,
# and an... | 1,655 | 41.461538 | 112 | sh |
null | ParlAI-main/projects/roscoe/roscoe_data/generate_perturbed_data.sh | #!/bin/bash
# Usage:
# sh roscoe_data/generate_perturbed_data.sh
PATH_TO_DATA="./projects/roscoe/roscoe_data/"
RESTORE_SCRIPT="./projects/roscoe/roscoe_data/restore_data.py"
mkdir -p ${PATH_TO_DATA}/synthetic
mkdir -p ${PATH_TO_DATA}/synthetic_sentinel
echo "Generating perturbations ..."
# ==== data generation for e... | 7,278 | 60.168067 | 255 | sh |
null | ParlAI-main/projects/roscoe/roscoe_data/restore_annotated.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Restore data with reasoning chains.
"""
import json
import csv
import os
import re
import xmltodict
from parlai.core... | 11,686 | 29.594241 | 109 | py |
null | ParlAI-main/projects/roscoe/roscoe_data/restore_data.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Restore perturbed synthetic data to maintain certain percent of perturbations.
Usage:
with perturbation IDs file:
py... | 6,071 | 32 | 124 | py |
null | ParlAI-main/projects/roscoe/synthetic_evaluation/score_all.sh | #!/bin/bash
# Example:
# sh projects/roscoe/synthetic_evaluation/score_all.sh sim_sce facebook/roscoe-512-roberta-base
print_usage() {
printf "Usage: score_all.sh model_type model_name"
}
if [ $# -ne 2 ] ; then
echo "Require 2 arguments: model type, and model name"
print_usage
exit 1
fi
SCORER="./proj... | 1,032 | 42.041667 | 115 | sh |
null | ParlAI-main/projects/roscoe/synthetic_evaluation/synthetic_roscoe.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Evaluate perturbed synthetic dataset.
Usage:
python projects/roscoe/synthetic_evaluation/synthetic_roscoe.py
"""
im... | 8,049 | 29.149813 | 92 | py |
null | ParlAI-main/projects/saferdialogues/README.md | # SaFeRDialogues: Taking Feedback Gracefully after Conversational Safety Failures
Megan Ung, Jing Xu, Y-Lan Boureau
## Abstract
Current open-domain conversational models can easily be made to talk in inadequate ways. Online learning from conversational feedback given by the conversation partner is a promising avenu... | 1,882 | 46.075 | 622 | md |
null | ParlAI-main/projects/safety_bench/README.md | # Safety Bench: Checks for Anticipating Safety Issues with E2E Conversational AI Models
A suite of dialogue safety unit tests and integration tests, in correspondence with the paper [*Anticipating Safety Issues in E2E Conversational AI: Framework and Tooling*](https://arxiv.org/abs/2107.03451).
**Abstract:** Over the... | 3,908 | 56.485294 | 956 | md |
null | ParlAI-main/projects/safety_bench/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
| 199 | 32.333333 | 65 | py |
null | ParlAI-main/projects/safety_bench/prepare_integration_tests.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Prepare data for integration tests. User will need to subsequently follow instructions
in the `Human Evalulations` se... | 6,505 | 34.747253 | 216 | py |
null | ParlAI-main/projects/safety_bench/run_unit_tests.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Run all safety unit tests for a model.
For example, to run all unit tests for the wrapper for the 90M parameter
Blen... | 7,164 | 35.186869 | 143 | py |
null | ParlAI-main/projects/safety_bench/model_wrappers/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
| 199 | 32.333333 | 65 | py |
null | ParlAI-main/projects/safety_bench/model_wrappers/example_wrapper.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Example wrapper which replies `hello` to every text.
"""
from projects.safety_bench.utils.wrapper_loading import regi... | 1,091 | 32.090909 | 92 | py |
null | ParlAI-main/projects/safety_bench/model_wrappers/gpt_wrappers.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Wrappers for GPT models from HF (in ParlAI).
Available models include:
- GPT2 large
- DialoGPT medium
"""
from abc i... | 3,384 | 26.745902 | 79 | py |
null | ParlAI-main/projects/safety_bench/model_wrappers/parlai_model_zoo_wrappers.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Wrappers for ParlAI models in the model zoo.
Available models include:
- blenderbot_90M
- blenderbot_400Mdistill
- b... | 4,901 | 35.857143 | 82 | py |
null | ParlAI-main/projects/safety_bench/unit_tests/__init__.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
| 199 | 32.333333 | 65 | py |
null | ParlAI-main/projects/safety_bench/unit_tests/generate_offensive_language_test.py | #!/usr/bin/env python3
# Copyright (c) Facebook, Inc. and its affiliates.
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
"""
Test the model's propensity to generate offensive language in 4 different settings:
- Safe (using task `blended_skil... | 6,036 | 30.442708 | 87 | py |