You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug
When Replicate isn't "warmed up", there are strange exceptions.
For example, examples/callback/README.md tells me to python repair_main.py, but this produces a confusing stack trace.
Perhaps the code that invokes Replicate or LiteLLM needs to retry on certain failures?
It would be easier to fix this if I could reproduce this. Once the failure happens then I can't produce the failure again.
To Reproduce python repair_main.py when Replicate isn't warmed up.
Expected behavior
Either correct behavior or an error message that tells me to run the program again would be preferrable.
Additional context
The confusing stack trace was
Traceback (most recent call last):
File "/Users/snible/src/prompt-declaration-language/examples/callback/repair_main.py", line 30, in <module>
pdl_output: PDLResult = pdl.pdl.exec_file(
~~~~~~~~~~~~~~~~~^
"./repair_prompt.pdl", scope=cast(pdl.pdl.ScopeType, pdl_input)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl.py", line 151, in exec_file
result = exec_program(program, config, scope, loc, output)
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl.py", line 73, in exec_program
result, _, scope, trace = process_prog(state, scope, prog, loc)
~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl_interpreter.py", line 241, in process_prog
for result, document, final_scope, trace in schedule([doc_generator]):
~~~~~~~~^^^^^^^^^^^^^^^^^
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl_scheduler.py", line 107, in schedule
msg = gen.send(v)
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl_interpreter.py", line 299, in step_block
result, background, scope, trace = yield from step_advanced_block(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
state, scope, block, loc
^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl_interpreter.py", line 350, in step_advanced_block
raise exc from exc
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl_interpreter.py", line 332, in step_advanced_block
result, background, scope, trace = yield from step_block_body(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
state, scope, block, loc
^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl_interpreter.py", line 449, in step_block_body
result, background, scope, trace = yield from step_blocks_of(
^^^^^^^^^^^^^^^^^^^^^^^^^^
...<6 lines>...
)
^
File "/Users/snible/.pyenv/versions/pdl/lib/python3.13/site-packages/pdl/pdl_interpreter.py", line 833, in step_blocks_of
raise PDLRuntimeError(
...<3 lines>...
) from exc
pdl.pdl_interpreter.PDLRuntimeError: Error during 'replicate/ibm-granite/granite-3.0-8b-instruct' model call: litellm.APIError: ReplicateException - can only join an iterable
The text was updated successfully, but these errors were encountered:
Describe the bug
When Replicate isn't "warmed up", there are strange exceptions.
For example, examples/callback/README.md tells me to
python repair_main.py
, but this produces a confusing stack trace.Perhaps the code that invokes Replicate or LiteLLM needs to retry on certain failures?
It would be easier to fix this if I could reproduce this. Once the failure happens then I can't produce the failure again.
To Reproduce
python repair_main.py
when Replicate isn't warmed up.Expected behavior
Either correct behavior or an error message that tells me to run the program again would be preferrable.
Additional context
The confusing stack trace was
The text was updated successfully, but these errors were encountered: