def train_batches(model, X_train, ids, run_options=None, run_metadata=None): for batch in gen.generate_batch_pairs(model.params, ids, generate_triplets=False): feed_dict = { model.x: X_train[batch['batch_samples']], model.pos_comps: batch['pos_comps'], model.neg_comps: batch['neg_comps'] } yield batch['batch_idx'], feed_dict
def train_batches(model, X_train, ids, run_options=None, run_metadata=None): batch_idx = 0 losses = [model.optimizer, model.cost] for batch in gen.generate_batch_pairs(model.params, ids, generate_triplets=False): feed_dict = { model.x: X_train[batch['batch_samples']], model.pos_comps: batch['pos_comps'], model.neg_comps: batch['neg_comps'], model.n_pos_comps: len(batch['pos_comps']), model.n_neg_comps: len(batch['neg_comps']) } yield batch['batch_idx'], feed_dict
def train_batches(model,X_train, ids, run_options=None, run_metadata=None): for batch in gen.generate_batch_pairs(model.params,ids, generate_triplets=False): feed_dict = {model.x:X_train[batch['batch_samples']], model.batch_ids:batch['batch_ids'], model.n_classes:len(np.unique(batch['batch_ids']))} yield batch['batch_idx'], feed_dict