README.md

metadata

language:
  - en
library_name: transformers
pipeline_tag: token-classification

Currently Adding Token Classification to mistral model ;

Still need some kinks sorting out -on the way - Testing

! pip install flash_attn
from transformers import AutoModelForTokenClassification, AutoTokenizer
import torch

# Load pre-trained model and tokenizer
model_name = "LeroyDyer/Mixtral_AI_TokenClassification"
tokenizer = AutoTokenizer.from_pretrained("LeroyDyer/Mixtral_AI_PsycoTron")
model = AutoModelForTokenClassification.from_pretrained(model_name, 
                                                        trust_remote_code=True,
                                                        num_labels=3)
model
model.train

# Get loss or logits
# loss = outputs.loss
# logits = outputs.logits