imported ftfy, IPython.

Based on the traceback you provided, the problematic code is located in the file C:\\Users\\Nishant\\OneDrive\\Desktop\\mission darkstar\\emotions_modelling\\ERC\\conv-emotion\\COSMIC\\feature-extraction\\comet\\src\\data\\utils.py, at line 60.

To fix the issue, open the utils.py file, find line 60 where it's using spacy.load('en') and replace it with spacy.load('en_core_web_sm').

My pytorch isnt working with currently installed CUDA 12.2 → 11.7

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. xformers 0.0.19 requires torch==2.0.0, but you have torch 2.0.1+cu117 which is incompatible.

can pose error for xformers

COSMIC, short for "COmmonSense knowledge for eMotion Identification in Conversations," is a framework developed to improve emotion detection in conversations. It uses commonsense knowledge to provide context that can help disambiguate the emotional content of the conversation.

Here are some features that COSMIC uses:

  1. Lexical and Syntactic Features: These include unigrams, bigrams, and part-of-speech tags. They capture the basic language information in the dialogues.
  2. Semantic Features: These include Named Entity Recognition (NER) tags and WordNet Affect terms, which provide information about the specific entities and emotion words used in the conversation.
  3. Sentiment Features: These include polarity scores (positive, negative, and neutral) and sentiment lexicons, which provide overall sentiment information.
  4. Empath Topics and Categories: Empath is a tool that can analyze text and identify a wide range of human sentiments and topics. The categories identified by Empath provide high-level contextual information about the conversation.
  5. Dialogue Act Features: These include categories like statement, emotion, ynQuestion (yes/no question), whQuestion (information-seeking question), etc. These categories provide information about the role of each utterance in the conversation.
  6. Emotion Transitions: These features capture the transition of emotions from one dialogue turn to the next. They help in understanding how the emotion evolves throughout the conversation.
  7. Commonsense-based Features: These features use commonsense knowledge bases like ATOMIC and ConceptNet to provide contextual information that can help disambiguate the emotional content. For example, they can capture cause-and-effect relationships, or the typical emotional reactions to certain events.

By combining all these different types of features, COSMIC aims to create a more comprehensive understanding of the emotional content of conversations.