Overview

what happens if try to load .pt that does not match the model definition? 

Great question. This is exactly the next mechanical thing to understand.
Usually PyTorch throws an error.
Example:
Saved model:
  nn.Linear(2,1)
contains:
  fc.weight shape = [1,2]
  fc.bias shape   = [1]
Now suppose you try to load into:
  nn.Linear(3,1)
PyTorch will complain:
  size mismatch for fc.weight:
  checkpoint shape [1,2]
  current model shape [1,3]
because it doesn't know where to get the extra weight.

Or suppose the layer names differ.
Saved file:
  fc.weight
  fc.bias
New model:
  self.fc2 = nn.Linear(2,1)
Now PyTorch says:
  Missing key(s): fc2.weight, fc2.bias
  Unexpected key(s): fc.weight, fc.bias
because the names don't match.

1 Output

python m05_load_pt_model_mismatch.py 
BEFORE LOAD
Parameter containing:
tensor([[-0.5612, -0.3073,  0.2916]], requires_grad=True)
Parameter containing:
tensor([-0.3175], requires_grad=True)
Traceback (most recent call last):
  File "/mnt/c/Users/terry/Downloads/607_predictive/m05_load_pt_model_mismatch.py", line 32, in <module>
    model.load_state_dict(torch.load("m01_tiny_model.pt"))
  File "/mnt/c/Users/terry/Downloads/607_predictive/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 2638, in load_state_dict
    raise RuntimeError(
RuntimeError: Error(s) in loading state_dict for TinyModel:
        size mismatch for fc.weight: copying a param with shape torch.Size([1, 2]) from checkpoint, the shape in current model is torch.Size([1, 3]).
(venv) terry@LAPTOP-HKPDHF7M:/mnt/c/Users/terry/Downloads/607_predictive$

2 PY scripts

# m05_load_pt_model_mismatch.py

import torch
import torch.nn as nn

# -----------------------
# Model definition
# -----------------------

class TinyModel(nn.Module):
    def __init__(self):
        super().__init__()
#        self.fc = nn.Linear(2, 1)
        self.fc = nn.Linear(3, 1)
    def forward(self, x):
        return self.fc(x)

# -----------------------
# New model object
# -----------------------

model = TinyModel()

print("BEFORE LOAD")
print(model.fc.weight)
print(model.fc.bias)

# -----------------------
# Load weights from disk
# -----------------------

model.load_state_dict(torch.load("m01_tiny_model.pt"))

print()
print("AFTER LOAD")
print(model.fc.weight)
print(model.fc.bias)

4 Code with detailed comments

26.0616 (v1 26.0616)