thatdramebaazguy's picture
Update README.md
a1d2f3a
|
raw
history blame
1.59 kB
metadata
datasets:
  - MIT Movie
language:
  - English
thumbnail: null
tags:
  - roberta
  - roberta-base
  - token-classification
  - NER
  - named-entities
  - BIO
  - movies
license: cc-by-4.0

roberta-base + Movies NER Task

Objective: This is Roberta Base trained for the NER task using MIT Movie Dataset

model_name = "thatdramebaazguy/roberta-base-MITmovie"
pipeline(model=model_name, tokenizer=model_name, revision="v1.0", task="ner")

Overview

Language model: roberta-base
Language: English
Downstream-task: NER
Training data: MIT Movie
Eval data: MIT Movie
Infrastructure: 2x Tesla v100
Code: See example

Hyperparameters

Num examples = 6253  
Num Epochs = 5 
Instantaneous batch size per device = 64
Total train batch size (w. parallel, distributed & accumulation) = 128  

Performance

Eval on MIT Movie

  • epoch = 5.0
  • eval_accuracy = 0.9476
  • eval_f1 = 0.8853
  • eval_loss = 0.2208
  • eval_mem_cpu_alloc_delta = 17MB
  • eval_mem_cpu_peaked_delta = 2MB
  • eval_mem_gpu_alloc_delta = 0MB
  • eval_mem_gpu_peaked_delta = 38MB
  • eval_precision = 0.8833
  • eval_recall = 0.8874
  • eval_runtime = 0:00:03.62
  • eval_samples = 1955

Github Repo: