lavis/projects/blip/eval/nocaps_eval.yaml

 # Copyright (c) 2022, salesforce.com, inc.
 # All rights reserved.
 # SPDX-License-Identifier: BSD-3-Clause
 # For full license text, see the LICENSE file in the repo root or https://opensource.org/licenses/BSD-3-Clause

model:
  arch: blip_caption
  model_type: base_coco
  # pretrained: 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/models/model_base_caption_capfilt_large.pth'

datasets:
  nocaps: # name of the dataset builder
    vis_processor:
        eval:
          name: "blip_image_eval"
          image_size: 384
    text_processor:
        eval:
          name: "blip_caption"
          prompt: "a picture of "

run:
  # task: retrieval
  task: captioning
  # optimizer
  batch_size_train: 32
  batch_size_eval: 64
  num_workers: 4

  max_len: 20
  min_len: 5
  num_beams: 3

  seed: 42
  output_dir: "output/BLIP/NoCaps"

  evaluate: True
  test_splits: ["val", "test"]

  device: "cuda"
  world_size: 1
  dist_url: "env://"
  distributed: True

  report_metric: False