README.md

metadata

license: bigscience-bloom-rail-1.0
tags:
  - generated_from_trainer
datasets:
  - pszemraj/riddlesense_plusplus
metrics:
  - accuracy
model-index:
  - name: bloom-better-riddles
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: pszemraj/riddlesense_plusplus
          type: pszemraj/riddlesense_plusplus
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.6594206731193033
parameters:
  min_length: 16
  max_length: 96
  no_repeat_ngram_size: 1
  do_sample: true
pipeline_tag: text-generation

bloom-better-riddles

This model is a fine-tuned version of bigscience/bloom-560m on the pszemraj/riddlesense_plusplus dataset. It achieves the following results on the evaluation set:

• Loss: 4.8107
• Accuracy: 0.6594

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.01
• num_epochs: 10.0

Training results

Framework versions

• Transformers 4.25.0.dev0
• Pytorch 1.13.0+cu117
• Datasets 2.7.0
• Tokenizers 0.13.2