run_imagenet_egc_eval_cls.sh

#!/usr/bin/env bash
# install pkg
pip install blobfile

# execute
LOGDIR=$1
GPU=$2
NODE=$3
RANK=$4
MADDR=$5
CKPT_PATH=$6
OPT=${@:7}

OPENAI_LOGDIR=./openai_log/${LOGDIR} python3 -m torch.distributed.run \
--nproc_per_node=$GPU --nnodes=$NODE \
--node_rank=$RANK --master_addr=$MADDR --master_port=29500 \
scripts/image_classification_eval.py --attention_resolutions 32,16,8 --class_cond True --diffusion_steps 1000 --dropout 0.0 \
--image_size 32 --learn_sigma False --noise_schedule linear --linear_start 0.00085 --linear_end 0.012 \
--num_channels 384 --num_head_channels 64 --num_res_blocks 2 --resblock_updown False --use_new_attention_order True \
--use_fp16 False --use_scale_shift_norm False --pool sattn --batch_size 8 --channel_mult 1,2,4 \
--in_channels 4 --use_spatial_transformer True --context_dim 512 --transformer_depth 1 \
--model_path $CKPT_PATH $OPT