remove unused memsize_to_str and minor cleanups [pr] (#9211)

* fix edge cases in memsize_to_str()

Inputs <= 1 now return "0.00 B" for 0 and "1.00 B" for 1, avoiding an
IndexError. Also, memsize_to_str(1000) now returns "1.00 KB" instead of
"1000.00 B".

Replaced the list comprehension with a next(...) generator for conciseness
and efficiency.

* simplify code using idiomatic python

- Remove the unused `memsize_to_str()` function in helpers.
- Use a tuple for checking multiple string prefixes/suffixes.
- Avoid unnecessary list construction by using iterables directly.
- Check None in @diskcache to ensure proper caching of falsy values.

* revert generators back to list comprehension

Sometimes building list first could be faster. Keep it as is.

examples/llama3.py

@@ -47,7 +47,7 @@ def concat_weights(models, device=None):
     disk_tensors: List[Tensor] = [model[name] for model in models]
     if len(disk_tensors) == 1 or len(disk_tensors[0].shape) == 1:
       return disk_tensors[0].to(device=device)
-    axis = 1 if name.endswith(".attention.wo.weight") or name.endswith(".feed_forward.w2.weight") else 0
+    axis = 1 if name.endswith((".attention.wo.weight", ".feed_forward.w2.weight")) else 0
     lazy_tensors = [data.to(device=device) for data in disk_tensors]
     return lazy_tensors[0].cat(*lazy_tensors[1:], dim=axis)
   return {name: convert(name) for name in {name: None for model in models for name in model}}