clipinvariance

Lisadunlap's group workspace

Group: Cub2011-Txt2Img

1-3

of 3

Tags

Notes

Author

lisadunlap

State

Finished

Start time

August 13th, 2023 11:54:18 AM

Runtime

3h 47m 58s

Tracked hours

3h 47m 36s

Run path

clipinvariance/ALIA-Cub2011/8gilzupv

Linux-5.15.0-25-generic-x86_64-with-glibc2.17

Python version

3.8.16

Git repository

git clone git@github.com:lisadunlap/ALIA.git

Git state

git checkout -b "vital-sweep-3" caaddb8bc73c0ea8a0a0b17519a448bf892d25a9

Command

main.py --config=configs/Cub2011/txt2img.yaml --epochs=200 --seed=2

System Hardware

CPU count	54
Logical CPU count	54
GPU count	10
GPU type	NVIDIA GeForce RTX 2080 Ti

W&B CLI Version

0.13.10

Group

Cub2011-Txt2Img

Config parameters are your model's inputs. Learn more

▶
Config parameters:{} 55 keys
- base_config:
  "configs/Cub2011/base.yaml"
- checkpoint_name:
  false
- config:
  "configs/Cub2011/txt2img.yaml"
- cpu_offloading:
  false
- data.augmentation:
  false
- data.base_dataset:
  "Cub2011"
- data.base_root:
  "/shared/lisabdunlap/data"
- data.batch:
  128
- data.class_balance:
  true
- data.embedding_root:
  "/shared/lisabdunlap/dataset_understanding/embeddings"
- data.extra_classes:
  false
- data.extra_dataset:
  "BasicDataset"
- data.extra_root:
  "/shared/lisabdunlap/edited/txt2img/Cub2011"
- data.filter:
  true
- data.num_extra:
  "extra"
- debug:
  false
- device:
  "cuda"
- epochs:
  200
- eval_only:
  false
- filter.checkpoint_name:
  "ckpt-Cub2011-none-filtered-resnet50-0-0.01-0.0001/best.pth"
- filter.filtered_path:
  false
- filter.load:
  true
- filter.model:
  "ViT-L/14"
- ▶
  filter.negative_prompts:[] 8 items
- filter.per_img:
  false
- filter.prompt:
  "a photo of a bird"
- filter.save_dir:
  "filtering_results"
- finetune:
  false
- gptq_act_order:
  false
- gptq_ckpt:
  "None"
- gptq_groupsize:
  -1
- gptq_wbits:
  16
- gpus:
  "None"
- hps.lr:
  0.01
- hps.lr_scheduler:
  "cosine"
- hps.weight_decay:
  0.0001
- load_8bit:
  false
- max_gpu_memory:
  "10GB"
- max_new_tokens:
  512
- message:
  "hello, how are you?"
- model:
  "resnet50"
- model_path:
  "lmsys/vicuna-13b-v1.3"
- name:
  "Cub2011-Txt2Img"
- num_gpus:
  4
- proj:
  "DiffusionTrainingDataCUB-filter"
- repetition_penalty:
  1
- temperature:
  1
- test:
  false
- wandb_silent:
  false
- yaml:
  "configs/Cub2011/txt2img.yaml"

Summary metrics are your model's outputs. Learn more

▶
Summary metrics:{} 22 keys
- best balanced acc:
  70.12235357001789
- ▶
  best class acc:{} 10 keys
- best epoch:
  18
- best group acc:
  69.88
- best val acc:
  69.88263721090783
- epoch:
  199
- lr:
  0.00000061683759169706
- test acc:
  69.88263721090783
- test accuracy:
  69.88263721090784
- test balanced accuracy:
  70.12235357001789
- ▶
  test class accuracy.bins:[] 33 items
- ▶
  test class accuracy.values:[] 32 items
- test photo acc:
  69.88
- train acc:
  100
- train loss:
  0.0018046986276997216
- train_size:
  5,994
- val acc:
  69.15774939592681
- val accuracy:
  69.15774939592683
- val balanced accuracy:
  69.45210191873176
- ▶
  val class accuracy.bins:[] 33 items
- ▶
  val class accuracy.values:[] 32 items
- val photo acc:
  69.16

This run produced these artifacts as outputs. Total: 1. Learn more

wandb-history

run-8gilzupv-history:v0