How to iterate over this?

[AntonLobach]

Hi,

I'm trying to translate this Python code, but bot sure how to iterate over test_dataloader:

batch_size = 64

# Create data loaders.
train_dataloader = DataLoader(training_data, batch_size=batch_size)
test_dataloader = DataLoader(test_data, batch_size=batch_size)

for X, y in test_dataloader:
    print(f"Shape of X [N, C, H, W]: {X.shape}")
    print(f"Shape of y: {y.shape} {y.dtype}")
    break

My attempts:

(doseq [[X y] (iterator-seq test-dataloader)]
  (print (format "Shape of X [N, C, H, W]: %s" (.shape X)))
  (print (format "Shape of y: %s" (.shape y)))) ;; Doesn't work
  
(let [it (.__iter__ test-dataloader)
        [x y] (.__next__ it)]
    (print (format "Shape of X [N,C,H,W]: %s" (.shape X)))
    (print (format "Shape of y: %s" (.shape y)))) ;; Doesn't work

[ikappaki]

Hi @AntonLobach,

The following worked for me out of the box. Also note that f-strings can be translated directly in Basilisp with the #f reader macro, thanks to @chrisrink10’s brilliance:

(doseq [[X y] train-dataloader]
  (println #f "Shape of X [N, C, H, W]: {(.-shape X)}")
  (println #f "Shape of Y: {(.-shape y)} {(.-dtype y)}"))
;; Shape of X [N, C, H, W]: #py (64 1 28 28)
;; Shape of Y: #py (64) torch.int64
;; ...

[AntonLobach]

Hi @ikappaki,

Thanks again for your help, it works!

I also found that beside my incorrect code, there was also an issue with the numpy version, but I didn't see the full error in Emacs and only spotted it when using the bare bones repl in the terminal.