Python evaluation.encode_data() Examples
The following are 6
code examples of evaluation.encode_data().
You can vote up the ones you like or vote down the ones you don't like,
and go to the original project or source file by following the links above each example.
You may also want to check out all available functions/classes of the module
evaluation
, or try the search function
.
Example #1
Source File: train.py From VSE-C with MIT License | 6 votes |
def validate_caption_only(opt, val_loader, model): # compute the encoding for all the validation images and captions img_embs, cap_embs = encode_data( model, val_loader, opt.log_step, logging.info) # caption retrieval (r1, r5, r10, medr, meanr) = i2t_text_only(img_embs, cap_embs, measure=opt.measure) logging.info("Image to text: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1, r5, r10, medr, meanr)) # sum of recalls to be used for early stopping currscore = r1 + r5 + r10 # record metrics in tensorboard tb_logger.log_value('r1', r1, step=model.Eiters) tb_logger.log_value('r5', r5, step=model.Eiters) tb_logger.log_value('r10', r10, step=model.Eiters) tb_logger.log_value('medr', medr, step=model.Eiters) tb_logger.log_value('meanr', meanr, step=model.Eiters) tb_logger.log_value('rsum', currscore, step=model.Eiters) return currscore
Example #2
Source File: train.py From VSE-C with MIT License | 5 votes |
def validate(opt, val_loader, model): # compute the encoding for all the validation images and captions img_embs, cap_embs = encode_data( model, val_loader, opt.log_step, logging.info) # caption retrieval (r1, r5, r10, medr, meanr) = i2t(img_embs, cap_embs, measure=opt.measure) logging.info("Image to text: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1, r5, r10, medr, meanr)) # image retrieval (r1i, r5i, r10i, medri, meanr) = t2i( img_embs, cap_embs, measure=opt.measure) logging.info("Text to image: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1i, r5i, r10i, medri, meanr)) # sum of recalls to be used for early stopping currscore = r1 + r5 + r10 + r1i + r5i + r10i # record metrics in tensorboard tb_logger.log_value('r1', r1, step=model.Eiters) tb_logger.log_value('r5', r5, step=model.Eiters) tb_logger.log_value('r10', r10, step=model.Eiters) tb_logger.log_value('medr', medr, step=model.Eiters) tb_logger.log_value('meanr', meanr, step=model.Eiters) tb_logger.log_value('r1i', r1i, step=model.Eiters) tb_logger.log_value('r5i', r5i, step=model.Eiters) tb_logger.log_value('r10i', r10i, step=model.Eiters) tb_logger.log_value('medri', medri, step=model.Eiters) tb_logger.log_value('meanr', meanr, step=model.Eiters) tb_logger.log_value('rsum', currscore, step=model.Eiters) return currscore
Example #3
Source File: train.py From vsepp with Apache License 2.0 | 5 votes |
def validate(opt, val_loader, model): # compute the encoding for all the validation images and captions img_embs, cap_embs = encode_data( model, val_loader, opt.log_step, logging.info) # caption retrieval (r1, r5, r10, medr, meanr) = i2t(img_embs, cap_embs, measure=opt.measure) logging.info("Image to text: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1, r5, r10, medr, meanr)) # image retrieval (r1i, r5i, r10i, medri, meanr) = t2i( img_embs, cap_embs, measure=opt.measure) logging.info("Text to image: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1i, r5i, r10i, medri, meanr)) # sum of recalls to be used for early stopping currscore = r1 + r5 + r10 + r1i + r5i + r10i # record metrics in tensorboard tb_logger.log_value('r1', r1, step=model.Eiters) tb_logger.log_value('r5', r5, step=model.Eiters) tb_logger.log_value('r10', r10, step=model.Eiters) tb_logger.log_value('medr', medr, step=model.Eiters) tb_logger.log_value('meanr', meanr, step=model.Eiters) tb_logger.log_value('r1i', r1i, step=model.Eiters) tb_logger.log_value('r5i', r5i, step=model.Eiters) tb_logger.log_value('r10i', r10i, step=model.Eiters) tb_logger.log_value('medri', medri, step=model.Eiters) tb_logger.log_value('meanr', meanr, step=model.Eiters) tb_logger.log_value('rsum', currscore, step=model.Eiters) return currscore
Example #4
Source File: train.py From CAMP_iccv19 with Apache License 2.0 | 5 votes |
def validate(opt, val_loader, model, tb_logger): # compute the encoding for all the validation images and captions print("start validate") model.val_start() img_embs, cap_embs, cap_masks = encode_data( model, val_loader, opt.log_step, logging.info) # caption retrieval (i2t_r1, i2t_r5, i2t_r10, i2t_medr, i2t_meanr), (t2i_r1, t2i_r5, t2i_r10, t2i_medr, t2i_meanr) = i2t(img_embs, cap_embs, cap_masks, measure=opt.measure, model=model) logging.info("Image to text: %.1f, %.1f, %.1f, %.1f, %.1f" % (i2t_r1, i2t_r5, i2t_r10, i2t_medr, i2t_meanr)) # image retrieval #(r1i, r5i, r10i, medri, meanr) = t2i( # img_embs, cap_embs, measure=opt.measure, model=model) logging.info("Text to image: %.1f, %.1f, %.1f, %.1f, %.1f" % (t2i_r1, t2i_r5, t2i_r10, t2i_medr, t2i_meanr)) # sum of recalls to be used for early stopping currscore = i2t_r1 + i2t_r5 + i2t_r10 + t2i_r1 + t2i_r5 + t2i_r10 # record metrics in tensorboard tb_logger.log_value('i2t_r1', i2t_r1, step=model.Eiters) tb_logger.log_value('i2t_r5', i2t_r5, step=model.Eiters) tb_logger.log_value('i2t_r10', i2t_r10, step=model.Eiters) tb_logger.log_value('i2t_medr', i2t_medr, step=model.Eiters) tb_logger.log_value('i2t_meanr', i2t_meanr, step=model.Eiters) tb_logger.log_value('t2i_r1', t2i_r1, step=model.Eiters) tb_logger.log_value('t2i_r5', t2i_r5, step=model.Eiters) tb_logger.log_value('t2i_r10', t2i_r10, step=model.Eiters) tb_logger.log_value('t2i_medr', t2i_medr, step=model.Eiters) tb_logger.log_value('t2i_meanr', t2i_meanr, step=model.Eiters) tb_logger.log_value('rsum', currscore, step=model.Eiters) return currscore
Example #5
Source File: train.py From SCAN with Apache License 2.0 | 4 votes |
def validate(opt, val_loader, model): # compute the encoding for all the validation images and captions img_embs, cap_embs, cap_lens = encode_data( model, val_loader, opt.log_step, logging.info) img_embs = numpy.array([img_embs[i] for i in range(0, len(img_embs), 5)]) start = time.time() if opt.cross_attn == 't2i': sims = shard_xattn_t2i(img_embs, cap_embs, cap_lens, opt, shard_size=128) elif opt.cross_attn == 'i2t': sims = shard_xattn_i2t(img_embs, cap_embs, cap_lens, opt, shard_size=128) else: raise NotImplementedError end = time.time() print("calculate similarity time:", end-start) # caption retrieval (r1, r5, r10, medr, meanr) = i2t(img_embs, cap_embs, cap_lens, sims) logging.info("Image to text: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1, r5, r10, medr, meanr)) # image retrieval (r1i, r5i, r10i, medri, meanr) = t2i( img_embs, cap_embs, cap_lens, sims) logging.info("Text to image: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1i, r5i, r10i, medri, meanr)) # sum of recalls to be used for early stopping currscore = r1 + r5 + r10 + r1i + r5i + r10i # record metrics in tensorboard tb_logger.log_value('r1', r1, step=model.Eiters) tb_logger.log_value('r5', r5, step=model.Eiters) tb_logger.log_value('r10', r10, step=model.Eiters) tb_logger.log_value('medr', medr, step=model.Eiters) tb_logger.log_value('meanr', meanr, step=model.Eiters) tb_logger.log_value('r1i', r1i, step=model.Eiters) tb_logger.log_value('r5i', r5i, step=model.Eiters) tb_logger.log_value('r10i', r10i, step=model.Eiters) tb_logger.log_value('medri', medri, step=model.Eiters) tb_logger.log_value('meanr', meanr, step=model.Eiters) tb_logger.log_value('rsum', currscore, step=model.Eiters) return currscore
Example #6
Source File: test_modules.py From CAMP_iccv19 with Apache License 2.0 | 4 votes |
def test_CAMP_model(config_path): print("OK!") logging.basicConfig(format='%(asctime)s %(message)s', level=logging.INFO) parser = argparse.ArgumentParser() #config_path = "./experiments/f30k_cross_attention/config_test.yaml" with open(config_path) as f: opt = yaml.load(f) opt = EasyDict(opt['common']) vocab = pickle.load(open(os.path.join(opt.vocab_path, '%s_vocab.pkl' % opt.data_name), 'rb')) opt.vocab_size = len(vocab) train_logger = LogCollector() print("----Start init model----") CAMP = model.CAMP(opt) CAMP.logger = train_logger if opt.resume is not None: ckp = torch.load(opt.resume) CAMP.load_state_dict(ckp["model"]) CAMP.train_start() print("----Model init success----") """ fake_img = torch.randn(16, 36, opt.img_dim) fake_text = torch.ones(16, 32).long() fake_lengths = torch.Tensor([32] * 16) fake_pos = torch.ones(16, 32).long() fake_ids = torch.ones(16).long() CAMP.train_emb(fake_img, fake_text, fake_lengths, instance_ids=fake_ids) print("----Test train_emb success----") """ train_loader, val_loader = data.get_loaders( opt.data_name, vocab, opt.crop_size, 128, 4, opt) test_loader = data.get_test_loader("test", opt.data_name, vocab, opt.crop_size, 128, 4, opt) CAMP.val_start() img_embs, cap_embs, cap_masks = encode_data( CAMP, test_loader, opt.log_step, logging.info) (r1, r5, r10, medr, meanr), (r1i, r5i, r10i, medri, meanri), score_matrix= i2t(img_embs, cap_embs, cap_masks, measure=opt.measure, model=CAMP, return_ranks=True) logging.info("Image to text: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1, r5, r10, medr, meanr)) logging.info("Text to image: %.1f, %.1f, %.1f, %.1f, %.1f" % (r1i, r5i, r10i, medri, meanri))