Executing Olive for QNN EP for llama3.2 1B instruct, getting Invalid engine (type=value_error)

After executing
olive run --config qnn_config.json

i get
/Olive/olive-ai/lib/python3.10/site-packages/olive/engine/packaging/packaging_generator.py:17: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81.
import pkg_resources
[2025-06-11 02:12:09,619] [WARNING] [config_utils.py:337:validate_config] Keys {'provider_options'} are not part of PythonEnvironmentTargetUserConfig. Ignoring them.
[2025-06-11 02:12:09,619] [WARNING] [config_utils.py:337:validate_config] Keys {'tokenizer'} are not part of DataConfig. Ignoring them.
[2025-06-11 02:12:09,620] [WARNING] [config_utils.py:337:validate_config] Keys {'tokenizer'} are not part of DataConfig. Ignoring them.
Traceback (most recent call last):
File "/Olive/olive-ai/bin/olive", line 8, in
sys.exit(main())
File "/Olive/olive-ai/lib/python3.10/site-packages/olive/cli/launcher.py", line 64, in main
service.run()
File "/Olive/olive-ai/lib/python3.10/site-packages/olive/cli/run.py", line 37, in run
olive_run(**var_args)
File "/Olive/olive-ai/lib/python3.10/site-packages/olive/workflows/run/run.py", line 230, in run
run_config: RunConfig = RunConfig.parse_file_or_obj(run_config)
File "/Olive/olive-ai/lib/python3.10/site-packages/olive/common/config_utils.py", line 126, in parse_file_or_obj
return cls.parse_file(file_path)
File "/Olive/olive-ai/lib/python3.10/site-packages/pydantic/v1/main.py", line 575, in parse_file
return cls.parse_obj(obj)
File "/Olive/olive-ai/lib/python3.10/site-packages/pydantic/v1/main.py", line 532, in parse_obj
return cls(**obj)
File "~/Olive/olive-ai/lib/python3.10/site-packages/pydantic/v1/main.py", line 347, in init
raise validation_error
pydantic.v1.error_wrappers.ValidationError: 12 validation errors for RunConfig
engine -> tokenizer
extra fields not permitted (type=value_error.extra)
passes -> q
Invalid engine (type=value_error)
passes -> g
Invalid engine (type=value_error)
passes -> cs
Invalid engine (type=value_error)
passes -> mb
Invalid engine (type=value_error)
passes -> mq
Invalid engine (type=value_error)
passes -> gs
Invalid engine (type=value_error)
passes -> sq
Invalid engine (type=value_error)
passes -> sp
Invalid engine (type=value_error)
passes -> st
Invalid engine (type=value_error)
passes -> cb
Invalid engine (type=value_error)
passes -> cp
Invalid engine (type=value_error)

I have installed QNN EP by creating an ONNX runtime wheel using QNN, i have installed the wheel file and can now see QNN-EP as a listed provider in onnx.

I am trying to do olive optimization for llama3 1b instruct model for 8 gen 3 and am unable to pinpoint the source of this error. Also What kind of GPU memory should i expect to be used for this process? currently i have am using a 12 GB VRAM

Start with qnn_config.json and the validation trace in olive/common/config_utils.py and olive/workflows/run/run.py. Check how RunConfig parses engine and passes, then compare the supplied configuration with the accepted schema. Done means the configuration validates without the tokenizer and Invalid engine errors and the QNN workflow can begin.