Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

New version crashes all the time #3525

Open
cranyy opened this issue Mar 14, 2025 · 0 comments
Open

New version crashes all the time #3525

cranyy opened this issue Mar 14, 2025 · 0 comments

Comments

@cranyy
Copy link

cranyy commented Mar 14, 2025

Issue

``

Aider v0.77.1.dev+import
Main model: anthropic/claude-3-7-sonnet-20250219 with diff edit format, 31.2k think tokens, infinite output
Weak model: anthropic/claude-3-5-haiku-20241022
Git repo: ..\PianoTest\Piano\.git with 43 files
Repo-map: using 4096 tokens, auto refresh
Added main.py to the chat.
Note: in-chat filenames are always relative to the git working dir, not the current working dir.
Cur working dir: E:\aider
Git working dir: `E:\PianoTest\Piano`

The below error happens constantly.

I see the problem. The error is happening because the model architecture being created during prediction doesn't match
the saved model architecture. Let me fix
Traceback (most recent call last):
  File "E:\aider\aider\coders\base_coder.py", line 1339, in send_message
    yield from self.send(messages, functions=self.functions)
  File "E:\aider\aider\coders\base_coder.py", line 1678, in send
    yield from self.show_send_output_stream(completion)
  File "E:\aider\aider\coders\base_coder.py", line 1826, in show_send_output_stream
    self.live_incremental_response(False)
  File "E:\aider\aider\coders\base_coder.py", line 1848, in live_incremental_response
    self.mdstream.update(show_resp, final=final)
  File "E:\aider\aider\mdstream.py", line 131, in update
    lines = self._render_markdown_to_lines(text)
  File "E:\aider\aider\mdstream.py", line 92, in _render_markdown_to_lines
    console.print(markdown)
  File "E:\aider\env\lib\site-packages\rich\console.py", line 1705, in print
    extend(render(renderable, render_options))
  File "E:\aider\env\lib\site-packages\rich\console.py", line 1326, in render
    for render_output in iter_render:
  File "E:\aider\env\lib\site-packages\rich\markdown.py", line 682, in __rich_console__
    yield from console.render(element, context.options)
  File "E:\aider\env\lib\site-packages\rich\console.py", line 1330, in render
    yield from self.render(render_output, _options)
  File "E:\aider\env\lib\site-packages\rich\console.py", line 1326, in render
    for render_output in iter_render:
  File "E:\aider\env\lib\site-packages\rich\syntax.py", line 628, in __rich_console__
    segments = Segments(self._get_syntax(console, options))
  File "E:\aider\env\lib\site-packages\rich\segment.py", line 681, in __init__
    self.segments = list(segments)
  File "E:\aider\env\lib\site-packages\rich\syntax.py", line 722, in _get_syntax
    ) = self._get_number_styles(console)
  File "E:\aider\env\lib\site-packages\rich\syntax.py", line 586, in _get_number_styles
    background_style = self._get_base_style()
  File "E:\aider\env\lib\site-packages\rich\syntax.py", line 408, in _get_base_style
    default_style = self._theme.get_background_style() + self.background_style
  File "E:\aider\env\lib\site-packages\rich\style.py", line 758, in __add__
    combined_style = self._add(style)
  File "E:\aider\env\lib\site-packages\rich\style.py", line 425, in __eq__
    return self.__hash__() == other.__hash__()

────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
► ANSWER

I see the problem. The error is happening because the model architecture being created during prediction doesn't match
the saved model architecture. Let me fix

or even worse when it just gives up out of nowhere, when it codes for a while ---

         except Exception as pickle_err:
             print(f"Error loading with pickle: {pickle_err}")
             raise
 >>>>>>> REPLACE


Let me know if you have any issues with this revised SEARCH/REPLACE block. I've matched it exactly with the content in
your main.py file.


Tokens: 70k sent, 1.9k received. Cost: $0.24 message, $0.49 session.

(env) E:\aider>
                     

Version and model info

No response

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant