problem when i'm trying to use the --unequal=1 case

[PeriklisTheodoropoulos]

raise ValueError("batch_size should be a positive integer value, "
ValueError: batch_size should be a positive integer value, but got batch_size=0

[GAmyin]

Hello! The problem with src/update.py is that line 50 testloader = DataLoader(DatasetSplit(dataset, idxs_test), batch_size=int(len(idxs_test)/10), shuffle=False), in line 44 idxs_test = idxs[int(0.9*len(idxs)):] means that 10% of the data in the client data set is divided into the test set again. If len(idxs_test) is less than 10, batch_size=int(len(idxs_test)/10) will be equal to 0, so in this setting, the data size of the client dataset cannot be less than 100; Analyzing the mnist_noniid_unequal(dataset, num_users) function at line 55 of src/sampling.py shows that this allocation strategy may result in clients being allocated only one shard or none. Each shard has 50 pieces of data, so at present, some clients are likely to have less than 100 pieces of data, which is the problem. Hope the problem can be solved smoothly.

[GAmyin]

Lines 88 to 140 of src/ samply.py can be modified as follows for reference only.
`

Assign the shards randomly to each client

for i in range(num_users):
    # First assign each client 2 shards to ensure every client has at least 2 shards of data
    rand_set = set(np.random.choice(idx_shard, 2, replace=False))
    idx_shard = list(set(idx_shard) - rand_set)
    for rand in rand_set:
        dict_users[i] = np.concatenate((dict_users[i], idxs[rand * num_imgs:(rand + 1) * num_imgs]), axis=0)

random_shard_size = random_shard_size - 2

# Next, randomly assign the remaining shards
for i in range(num_users):
    if len(idx_shard) == 0:
        break
    shard_size = random_shard_size[i]
    if shard_size > 0:
        if shard_size > len(idx_shard):
            shard_size = len(idx_shard)
        rand_set = set(np.random.choice(idx_shard, shard_size, replace=False))
        idx_shard = list(set(idx_shard) - rand_set)
        for rand in rand_set:
            dict_users[i] = np.concatenate((dict_users[i], idxs[rand * num_imgs:(rand + 1) * num_imgs]), axis=0)

if len(idx_shard) > 0:
    # Add the leftover shards to the client with minimum images:
    shard_size = len(idx_shard)
    # Add the remaining shard to the client with lowest data
    k = min(dict_users, key=lambda x: len(dict_users.get(x)))
    rand_set = set(np.random.choice(idx_shard, shard_size, replace=False))
    idx_shard = list(set(idx_shard) - rand_set)
    for rand in rand_set:
        dict_users[k] = np.concatenate((dict_users[k], idxs[rand * num_imgs:(rand + 1) * num_imgs]), axis=0)

`